Is our DNA objected oriented?
LAMP-PU...@worldnet.att.net
I thought this post might be better off in an genome newsgroup, but I doubt
those guys know anything about OO programing. :)
Well anyway, where I work we have been programming in procedural style
type methods for decades (using only assembly language). After reviewing
OO methods and re-evaluating our code, we have come to the conclusion that
much of our existing code is actually objected oriented. How could this be?
We didn't know anything about OO methods when we were coding. Hmmm.
Well, assuming OO code is superior to procedural code, I can only guess
that it was the survival of the fittest code. Any code that was small,
independent, had few interfaces, worked well with closely
related features, simply survived, copied onto the next project.
In addition many of the features of C++, Eiffel and other high level
OO languages such as type checking, variable scoping, have been
incorporated into our assembly language code indirectly via
preprocessors, code reviews, and specification enhancements.
So the question is with our genome, since it is the oldest program
known, is it possible it is objected oriented in design? From
some light reading in genetics it seems that scientists are
having a hard time trying to figure out how that "program" works.
But I would too, if I were only examining the executable code generated
from C++ or Eiffel. Or perhaps our genome structure is far more
advanced that OO. Anybody know of any books on this subject?
thanks
steve
Ell
>So the question is with our genome, since it is the oldest program
>known, is it possible it is objected oriented in design?
I guess it would be good idea to explain what "genome" means for those
who don't know. Genome is the set of chromosomes in the germ cells -
sperm, or ovum - as opposed to the set of chromosomes in the other
parts of the body - somatic cells. Germ cells have half the number of
chromosomes that somatic cells have. On to relating genome with OO.
Being that OO derives from generalizing about the world, I would
expect that the design of the genome does have OO features - entities
with behavior and properties, commonality amongst entities, interfaces
which hide internals, entities with state, unique entities, different
yet related phenomena which can serve as alternatives for each other,
etc.
And significantly the way the genome is created and how it operates in
terms of genetics and evolution generally more than likely has the
same kinds of OO features.
>From some light reading in genetics it seems that scientists are
>having a hard time trying to figure out how that "program" works.
>But I would too, if I were only examining the executable code generated
>from C++ or Eiffel.
OO is certainly larger and greater than what any one OOPL can express.
>Or perhaps our genome structure is far more
>advanced that OO.
Possibly. While OO is extremely powerful as a modelling tool, it is
likely not the be all and end all of modelling paradigms. And in all
probability the genome, genetics, DNA, etc would also have the
features of this new paradigm incorporated into them.
Anybody know of any books on this subject?
I make a brief reference to mitochondria as it relates to private
inheritance in my "Broad New Vistas" paper at my web site. I also
discuss humans relationship to chimpanzees in terms of inheritance in
the same paper. I discuss more about mitochondria and OO features in
the paper "Beauty and Power of C++" which was published in the book
'Wisdom of the OO Gurus' by SIGS in '96. I also discuss polymorphism
and the immune system in that paper. I will be posting the paper at
my web site in 1 or 2 days.
Elliott
--
:=***=: VOTE NO TO MODERATION! :=***=:
CRAFTISM SHOULD NOT USE USENET RESOURCES TO AVOID CRITICISM!
MODERATORS SHOULD NOT HAVE LIFETIME TERMS!
:=***=: Objective * Pre-code Modelling * Holistic :=***=:
Hallmarks of the best SW Engineering
Study Phony Crafite OO vs. Genuine OO: http://www.access.digex.net/~ell
Copyright 1998 Elliott. exclusive of others' writing. may be copied
without permission only in the comp.* usenet and bitnet groups.
Joachim Durchholz
>
> Well anyway, where I work we have been programming in procedural style
> type methods for decades (using only assembly language).
> In addition many of the features of C++, Eiffel and other high level
> OO languages such as type checking, variable scoping, have been
> incorporated into our assembly language code indirectly via
> preprocessors, code reviews, and specification enhancements.
Interesting to hear that there is some sort of convergent evolution at
work here :)
> So the question is with our genome, since it is the oldest program
> known, is it possible it is objected oriented in design?
Probably not.
Anything that comes out of evolution isn't "designed" at all, to begin
with. It just happened to work out.
When looking at the interactions in the human body, I see a wild mixture
of physical, electrical, and neural signal transmission. Nearly none of
these signals has a clearly defined recipient, nor is any one signal a
clear indication of some condition. Everything "talks" to everything -
the exact opposite of good engineering.
Of course, Nature has no business in engineering. There's nobody around
who wants or needs to understand how a human body works. So there's no
point in *complaining* that Nature "does" it the unstructured way - it's
just that humans cannot use this for design. (Note that we can still
learn from Nature's tricks. It's just that I doubt that we can copy any
design principles: our brains aren't geared towards understanding
tightly interwoven, complex interactions, so a system with an
evolution-style design would not be understandable to a human,
regardless of how well it did its job.)
Regards,
Joachim
--
Please don't send unsolicited ads.
Robert C. Martin
LAMP-PU...@worldnet.att.net wrote in message
<710q7e$9...@bgtnsc02.worldnet.att.net>...
>
>Well anyway, where I work we have been programming in procedural style
>OO methods and re-evaluating our code, we have come to the conclusion that
>much of our existing code is actually objected oriented. How could this be?
>We didn't know anything about OO methods when we were coding. Hmmm.
>
>Well, assuming OO code is superior to procedural code, I can only guess
>that it was the survival of the fittest code. Any code that was small,
>independent, had few interfaces, worked well with closely
>related features, simply survived, copied onto the next project.
>OO languages such as type checking, variable scoping, have been
>incorporated into our assembly language code indirectly via
>preprocessors, code reviews, and specification enhancements.
Here are some questions you can apply to your assembly language code to
ascertain whether or not it is truly object oriented.
1. Are the vast majority of subroutines called indirectly through pointers?
2. Are the concepts that your program manipulates represented by data
structures that contain little more than pointers to subroutines. Call
these pointer structures interfaces?
3. Are their data structures that represent the low level implementation of
those concepts, and that are manipulated by functions called through the
interface pointers. Call these low level data structures objects?
4. Are there several different kinds of objects that are controlled by the
exact same interface?
5. Are there subroutines that don't care what kind of object they are
manipulating because they simply use an interface?
>So the question is with our genome, since it is the oldest program
>known, is it possible it is objected oriented in design? From
>some light reading in genetics it seems that scientists are
>having a hard time trying to figure out how that "program" works.
>But I would too, if I were only examining the executable code generated
>from C++ or Eiffel. Or perhaps our genome structure is far more
>advanced that OO. Anybody know of any books on this subject?
The genetic language is exceedingly complex. Its simple beginnings in the
four base nucleosides is misleading. A DNA molecule can be divided into
coding sections and non-coding sections. The coding sections contain the
information needed to build proteins. The non-coding sections are something
of a mystery. (They probably aren't comments...)
But just the coding sections are complex enough. The four necleosides are
Adenine, Cytosine, Guanine, and Thiamine. or ACGT (Which is where the name
of the film Gattaca came from). The coding sections are subdivided into
groups of three bases each. These groups are called codons. There are 64
possible codons. Most of them map to one of the twenty or so amino acids
that are the constituents of proteins. (The mapping is performed by a
molecule known as transfer RNA (tRNA) which is the real keeper of the
genetic code). But a few other codons represent control sequences like
"start here" and "stop here".
A molecule by the name of RNA-transcriptase binds to the DNA molecule at a
"start-here" location. It then walks along the DNA molecule building an
equivalent messenger RNA molecule (mRNA). It stops at the "stop-here"
point. You can think of the mRNA molecule as a short copy of a segment of a
DNA molecule. The mRNA becomes the input to a wonderful little factory
called a ribosome. Lots of ribosomes stud the inner structures of a cell.
The mRNA binds to it and is read by it. The ribosome gathers tRNA from the
intercellular medium and matches it against the mRNA it is reading. The
tRNA is bound to a particular amino acid. The ribosome rips the amino acid
off the tRNA and sticks it on the end of a growing protein molecule.
Thus the growing protein molecule is a translated copy of the DNA molecule,
where each three bases in the DNA correspond to an amino acid in the
protein. The amino acids all have a fundementally similar structure that
forces them to coil up into a spiral like shape called the alpha helix
(dicovered by Linus Pauling after he switched to life sciences having been
disillusioned by the nuclear physics in the 40s and 50s).
The amino acids are all slighly different. Each has a different grouping of
atoms in a key location in its structure. These differences give each amino
acid a different characteristic. Some have a slight negative charge. (Or
at least they show a charge at the part of the molecule that faces the
outside.) Some have a slight positive charge. Some are oily and aren't
bound by water (they don't get wet). Some have an affinity for water (they
get wet), etc, etc.
Because of these differences, one part of the helix will be attracted to, or
repelled by another. These forces (and others that we have trouble
identifying) cause the protein to fold into unique and distinctive shapes.
Each protein has its own specific shape. These shapes can be very
complicated. They can have parts that are structurally stable, and parts
that are structurally flexible. (i.e. they can bend and twist in very
specific ways). And at just the right point on their sufaces they can
express a concentration of charge, or wetness, or in some cases an atom of
iron or some other element with interesting chemistry.
(BTW, the reason egg whites become white and solid when you cook them is
because the heat shakes delicate protein shapes loose. The proteins become
long threads instead of carefully folded structures. Then the threads get
all tangled up in each other, binding to each other in unnatural ways, thus
becoming quasi-solid. This process is called denaturing.)
Thus, these folded chaines of amino acids become little machines with very
specific purposes. These little machines get their power from a whole bunch
of cocked springs that are floating around in the intercellular medium.
These springs are called adenosine-triphosphate (ATP). An ATP molecule has
a shape that can be recognized by the complimentary shapes of many proteins.
When an ATP molecule happens to fall into the right pocket of a protein, it
stresses the ATP molecule in just the right way to make one of the phosphate
groups pop off. This energetically changes the shape of the resulting
adenosine-diphosphate molecule (ADP) which places mechanical pressure on the
protein molecule. The protein molecule changes shape accordingly.
Thus, these little protein machines have functions that are generally
related to their shape, and the way ATP makes then change their shape. (Not
all proteins function this way. Some are static and have no dynamic
function at all. Others derive their energy from other sources. But many,
possibly the majority, derive their energy from ATP.)
So, is the genetic code OO? It certainly does not resemble a set of classes
with polymorphic abstract interfaces if that's what you mean. On the other
hand one could view the proteins as objects. And one could talk about how
encapsulated and decoupled those objects are. However that's probably an
inappropriate comparison.
Each of the protiens fits into a system. The system is a grouping of many
different proteins and structures which perform a function. Each such
system is just a component of higher order systems, which are yet
subcomponents of higher order systems. And what we know about those systems
shows that they are not very nicely decoupled. They work, no doubt about
it, but there are some really icky dependencies in the system that would be
better off without.
One such intriguing dependency is breathing. We breathe to keep the levels
of oxygen in our blood high. But we also breath to keep the acidity of our
blood low. CO2 disolved in water becomes carbonic acid. When blood acid
levels rise, breathing quickens in order to flush the acid forming CO2 from
our blood; irrespective of the need for oxygen. Indeed, our drive to
breathe is more strongly regulated by blood acid than by lack of oxygen.
This odd dependency leads to positive feedback loops in cases of panic, or
high altitude. That is why panic attacks are often treated by forcing a
person to breathe into a paper bag.
This is the kind of bizzare interdependency that a good OO designer would
use the tools of OO to break. So, in the final analysis, I would have to
say that the genetic code is not OO, nor does it lead to a system that is
OO, simply because the system has some rather nasty interdependencies that a
good OO designer would not allow.
Robert C. Martin | Design Consulting | Training courses offered:
Object Mentor | rma...@oma.com | Object Oriented Design
14619 N Somerset Cr | Tel: (800) 338-6716 | C++
Green Oaks IL 60048 | Fax: (847) 918-1023 | http://www.oma.com
Phlip
>
>LAMP-PU...@worldnet.att.net wrote in message
><710q7e$9...@bgtnsc02.worldnet.att.net>...
>>
>>much of our existing code is actually objected oriented...
The heart of this problem is deciphering the OO properties of code that was
not designed in the normal fashion.
>Here are some questions you can apply to your assembly language code to
>ascertain whether or not it is truly object oriented.
>
>1. Are the vast majority of subroutines called indirectly through pointers?
While this matters to OO code, and while in that realm it can be argued that
the majority of calls should be static and that command and control calls
should be virtualized, in DNA code messages are rarely broadcast only to
objects. They are always sent to classes of objects, some of whom pick it up
at random (imagine the flowchart!). But consider a ribosome working on tRNA.
The codon signals in the gene sequence affect one ribosome directly - static
linking. (But at a smaller scale the message is still broadcast to the class
of energy molecules in the ribosome.) At a larger scale, a cell divides
because the signal to do so built up until the entire system underwent a
dynamic cascade event - this represents dynamic linking of command and
control.
>2. Are the concepts that your program manipulates represented by data
>structures that contain little more than pointers to subroutines. Call
>these pointer structures interfaces?
The surfaces of enzymes and their magnetic properties form the interfaces.
But we risk going orthogonal to the original question. It's not enough to
say that DNA has classes and objects that support protein meta-objects; that
does not answer whether DNA has classes or objects!
>3. Are their data structures that represent the low level implementation of
>those concepts, and that are manipulated by functions called through the
>interface pointers. Call these low level data structures objects?
Yes. But "structures" are not automatically "objects".
>4. Are there several different kinds of objects that are controlled by the
>exact same interface?
Yes. Many sequences of DNA code the same sequences over and over again,
sometimes with minor variations. Meta-classes.
>5. Are there subroutines that don't care what kind of object they are
>manipulating because they simply use an interface?
Some molecules chemically match those of a similar shape (THC and dopamine,
for example, or aspirin and acetaminophene).
>>...So the question is with our genome, since it is the oldest program
>>known, is it possible it is objected oriented in design?...
>The genetic language is exceedingly complex. Its simple beginnings in the
>four base nucleosides is misleading. A DNA molecule can be divided into
>coding sections and non-coding sections. The coding sections contain the
>information needed to build proteins. The non-coding sections are
something
>of a mystery. (They probably aren't comments...)
They are commented code. But...
Exons can be like 'if (false)', 'if (never)', 'if (maybe)', 'if
(sometimes)', or 'if (frequently)'. Just because today the ribosome snipped
them out of the tRNA does not mean that tomorrow they will, or in three
centuries they still will. Scientists have speculated we could analyze a
chicken's DNA and find enough fossile sequences to reconstruct a dinosaur -
just don't pen the thing in a security fence controlled by Unix!
>But just the coding sections are complex enough. The four necleosides are
>Adenine, Cytosine, Guanine, and Thiamine. or ACGT (Which is where the
name
>of the film Gattaca came from)...
Is a ribosome a Turing engine?
>...The coding sections are subdivided into
>groups of three bases each. These groups are called codons. There are 64
>possible codons. Most of them map to one of the twenty or so amino acids
>that are the constituents of proteins. (The mapping is performed by a
>molecule known as transfer RNA (tRNA) which is the real keeper of the
>genetic code). But a few other codons represent control sequences like
>"start here" and "stop here".
And some scientists have speculated that because the mapping between codons
and amino acids is arbitrary, one could create a cell that used a completely
different map and it would still work. Like switching between the QWERTY and
Dvorak keyboards. The magnetic properties of the codons would affect this -
they are not perfectly uniform, of course, but many transfers of meaning are
probably possible.
>So, is the genetic code OO? It certainly does not resemble a set of
classes
>with polymorphic abstract interfaces if that's what you mean. On the other
>hand one could view the proteins as objects. And one could talk about how
>encapsulated and decoupled those objects are. However that's probably an
>inappropriate comparison.
The appropriateness of the comparison begins with the fundamental "code" and
"execution" dichotomy - DNA is code and protein is its program running in
memory. A prion is _not_ OO. The comparison lives where we perceive classes
and objects dispatching messages virtually to each other and processing
thru-put. The comparison ends with the concept of "design" - OO _must_ be
designed from without, and Life must _not_. But before we reach this
eschatological limit the analogy is very close.
>[No,] simply because the system has some rather nasty interdependencies
that a
>good OO designer would not allow.
This is an _effect_ of the lack of design from without, but it is just a
symptom of this ultimate reason. Encapsulation still exists, but it simply
holds much more objects than we are familiar with keeping mental track of.
And nature does not need a perfect design, just random mass unit testing.
-- Phlip at politizen dot com (address munged)
======= http://users.deltanet.com/~tegan/home.html =======
-- if you can't dazzle them with brilliance,
then baffle them with PERL --
Darren New
> Here are some questions you can apply to your assembly language code to
> ascertain whether or not it is truly object oriented.
>
> 1. Are the vast majority of subroutines called indirectly through pointers?
Yep.
> 2. Are the concepts that your program manipulates represented by data
> structures that contain little more than pointers to subroutines. Call
> these pointer structures interfaces?
Yep.
> 3. Are their data structures that represent the low level implementation of
> those concepts, and that are manipulated by functions called through the
> interface pointers. Call these low level data structures objects?
Yep.
> 4. Are there several different kinds of objects that are controlled by the
> exact same interface?
Yep.
> 5. Are there subroutines that don't care what kind of object they are
> manipulating because they simply use an interface?
Yep.
Does that mean FORTH is an OO system? Eeeep!
--
Darren New / Senior Software Architect / First Virtual Holdings Inc
http://www.fv.com or in...@fv.com -=|=- PGP Key: ftp://ftp.fv.com/pub/fv
Fingerprint: 61 7D AF 9E 00 CC C2 ED / D8 4C D7 AA E4 C2 A0 73
Robert C. Martin
Phlip wrote in message <712bah$skm$1...@news0-alterdial.uu.net>...
>But consider a ribosome working on tRNA.
>The codon signals in the gene sequence affect one ribosome directly -
static
>linking.
But consider, that the ribosome translates a codon into an amino acid. The
tRNA is a vtbl. The codon is a virtual function call. The Amino acid is
the function implementation.
We could completely alter the genetic code by changing the tRNA molecules.
That's an interesting layer of abstraction....
>Is a ribosome a Turing engine?
It doesn't need to be. DNA by itself is. There was an interesting article
in Scientific American having to do with computation with genetics. The
interesting thing about this is that intractable problems for computers
become tractable for DNA computations because you can have several billion
DNA molecules participating in the computation simultateously. Lots of
parallelism...
Phlip
>To continue this absurd but entertaining analogy...
>
>Phlip wrote in message <712bah$skm$1...@news0-alterdial.uu.net>...
>>The codon signals in the gene sequence affect one ribosome directly -
>static
>>linking.
>
>tRNA is a vtbl. The codon is a virtual function call. The Amino acid is
>the function implementation.
Another point in the Negative category:
My example compared the minute mechanics between two molecules already
plugged into each other as static linking and the broadcast of many
hormone-like chemicals to many different types of receivers as dynamic
dispatch (and the Observer DP ;-). But you confused my example to mean that
the ribosome is statically linked to the entire protein it emits. Enzymes
are hitting that protein as it comes out, so the end-result is as v-tabular
as you want.
But the confusion arose because nuclear material is deeply self-referential
in a way programming languages can't be. Proteins generated by DNA form
enzymes that re-arrange tRNA. >Bzzzt.< No OO in a system that couples in a
way that a programming language _can't_ couple.
>We could completely alter the genetic code by changing the tRNA molecules.
>That's an interesting layer of abstraction....
mRNA - the L-shaped one that holds a proto-amino acid on one arm and its
respective codon on the other. IIRC.
>It doesn't need to be. DNA by itself is. There was an interesting article
>in Scientific American having to do with computation with genetics. The
>interesting thing about this is that intractable problems for computers
>become tractable for DNA computations because you can have several billion
>DNA molecules participating in the computation simultateously. Lots of
>parallelism...
So that's how I politely explain to the students on the C++ newsgroups why
they can't have { void main() | getche() | virtual constructors |
mind-reading typecasts }. "Because the Standard permits that someday C++ be
compiled to DNA."
Tim Ottinger
> Well anyway, where I work we have been programming in procedural style
> type methods for decades (using only assembly language). After reviewing
> OO methods and re-evaluating our code, we have come to the conclusion that
> much of our existing code is actually objected oriented. How could this be?
> We didn't know anything about OO methods when we were coding. Hmmm.
It's kind of cool, since Kay's concept (for which he coined the term
Object-Oriented) was a biological model.
Thanks for the amusing interjection!
Richard Melvin
writes
>And nature does not need a perfect design, just random mass unit testing.
So, how long till we see this technique in commercial use?
Why waste time shipping the same software to all your beta testers?
Give each shipped copy a slight, random variation. Log support calls,
until you find one that works more than half the time.
This is your seed for the next generation...
--
Richard Melvin
Matt Kennel
:LAMP-PU...@worldnet.att.net wrote:
:
:>So the question is with our genome, since it is the oldest program
:>known, is it possible it is objected oriented in design?
:
:I guess it would be good idea to explain what "genome" means for those
:who don't know. Genome is the set of chromosomes in the germ cells -
:sperm, or ovum - as opposed to the set of chromosomes in the other
:parts of the body - somatic cells. Germ cells have half the number of
:chromosomes that somatic cells have. On to relating genome with OO.
:
:Being that OO derives from generalizing about the world, I would
:expect that the design of the genome does have OO features - entities
:with behavior and properties, commonality amongst entities, interfaces
:which hide internals, entities with state, unique entities, different
:yet related phenomena which can serve as alternatives for each other,
:etc.
No, I truly doubt so. Genomic ''computation'' often occurs via subtle
gene-gene interactions mediated quite indirectly via chemistry, kinetics,
and assumptions about external inputs.
There is massive ''global state'' (concentrations of various molecules
in the cell) and ''calls'' are thermodynamic and probabilistic.
I would call much of it Very Bad Programming Practice. But it keeps
molecular biologists quite well employed.
:And significantly the way the genome is created and how it operates in
:terms of genetics and evolution generally more than likely has the
:same kinds of OO features.
Programs 'evolved' via genetic algorithms and competition (e.g. Core
Wars) in simulation (not actual cells) have often turned out to be
exceptionally clever in the canonical Undisciplined Genius Assembly
Hacker way; relying on obscure side-effects or interactions.
Why?
Because in no way are there any evolutionary constraints giving a benefit to
"human understandibility".
--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD
*
* "To chill, or to pop a cap in my dome, whoomp! there it is."
* Hamlet, Fresh Prince of Denmark.
Biju Thomas
>
> The genetic language is exceedingly complex. Its simple beginnings in the
> four base nucleosides is misleading. A DNA molecule can be divided into
> coding sections and non-coding sections. The coding sections contain the
> information needed to build proteins. The non-coding sections are something
> of a mystery. (They probably aren't comments...)
>
They probably are unused virtual functions. God may not had enough time
or patience to find an optimization algorithm to get rid all of them :-)
Concerning the original point, can we write a program to generate
classes from bits? If that is possible, probably, we can apply it to the
genetic code to find the OO program behind genetic code.
Regards,
Biju Thomas
Jason Stokes
Matt Kennel <NOSPAMmbke...@yahoo.com> wrote:
>Programs 'evolved' via genetic algorithms and competition (e.g. Core
>Wars) in simulation (not actual cells) have often turned out to be
>exceptionally clever in the canonical Undisciplined Genius Assembly
>Hacker way; relying on obscure side-effects or interactions.
>
>Why?
>
>Because in no way are there any evolutionary constraints giving a benefit to
>"human understandibility".
Evolution is a rotten software engineer. It just has software life
cycles on the line of millions of years.
--
Jason Stokes: js...@bluedog.apana.org.au
LAMP-PU...@worldnet.att.net
I see my wacky post create a surprising large response, on to
the specific points
"Robert C. Martin" <rma...@oma.com> wrote:
>
>
>Here are some questions you can apply to your assembly language code to
>ascertain whether or not it is truly object oriented.
>
>snip
Our code is certainly not truly object oriented, the point was that
it was evolving into OO without any design intent, maybe its
only 10% OO, but the trend was certainly measureable.
>
>The genetic language is exceedingly complex. Its simple beginnings in the
>four base nucleosides is misleading.
Yes, but so is the simple instruction set of a modern RISC processor, the
ADD instruction decodes into a series of many microcodes that actually
performs the operation of enabling buffers, disabling memory devices, etc
all with precise timing constraints. I think of the analogy as
Biological Computational
nucleotides (base 4) memory bit (base 2)
all 64 codons the set of microcodes
codon a specific microcode
gene a machine instruction
group of genes a subroutine
RNA transcription
and ADT interaction a procedure call, changing and creating the data
protein the data
shape of protein the state of the data (values of each attribute)
skin the complete set of data for a high level class
Obviously this analogy is much different than yours, and probabaly
is absurb, as I changed it 10 times while writing this post and it still has
inconsistences,
but... I believe the analysis of a true object oriented
system would be equally complex if examined at the microcode level.
>One such intriguing dependency is breathing. We breathe to keep the levels
>
>This is the kind of bizzare interdependency that a good OO designer would
>use the tools of OO to break. So, in the final analysis, I would have to
>say that the genetic code is not OO, nor does it lead to a system that is
>OO, simply because the system has some rather nasty interdependencies that a
>good OO designer would not allow.
>
>
We design flight controls for positive feedback fighter aircraft, does that
mean we can't use OO techniques to code the control laws? Nasty
interdependecies are the nature of physical laws and can't be avoided.
The purpose of OO is to reduce interdependencies to a minimum,
but NOT any smaller. Any complex system is going to have bizzare convoluted
strange interdependencies, but if they were reduced to a minimum by OO design
techniques, then OO has done its job to the best of its abilities and the
code can rightly be called OO. Has our DNA reduced this interdependencies
to a minimum? I don't think we know the answer to that question yet.
steve
Robert C. Martin
LAMP-PU...@worldnet.att.net wrote in message
<713n45$4...@bgtnsc02.worldnet.att.net>...
>
>I see my wacky post create a surprising large response, on to
>the specific points
>
>"Robert C. Martin" <rma...@oma.com> wrote:
>>
>>
>>ascertain whether or not it is truly object oriented.
>>
>
>Our code is certainly not truly object oriented, the point was that
>it was evolving into OO without any design intent, maybe its
>only 10% OO, but the trend was certainly measureable.
I understand what you are getting at. And I have seen similar situations
before. In 1977 I was working on a system that was 30K of 8085 assembler
written onto 30 1K PROMs. We were going to ship thousands of these systems
around the world. We needed a way to patch the software when it was in the
field. Shipping 30 new proms to each site was impractical.
So, I divided all the software up between the 30 proms, making sure that
each prom had a couple hundred bytes left. Then I created a jump table for
each prom. All function calls (whether within the prom, or between proms)
took place using these jump tables. The jump tables were kept in RAM and
were built by an initialization process at boot.
This allowed us to add new code to an existing prom, moving things around on
that prom, without changing any of the other proms. In essence I created 30
vtbls and the proms represented 30 classes.
Clearly this was not OO, The functions on the proms were not related
conceptually. Indeed, in order to get the best fit onto each prom, I had to
shuffle the functions around based only on their size. So each prom was
composed of the functions that fit best, not those that were conceptually
related. Thus, the logical to physical mapping was horrendous.
Still, the use of the jump tables to achieve the open closed principle was a
precursor to an OO technique.
Necessity is the mother of invention. We needed a solution, and we solved
it. We did not, however, identify this as a new paradigm, or push forward
to the next step of creating true classes and objects. As prom technology
improved, the number of proms in the system shrank from 30 down to 1. The
jump table scheme was no longer needed. Nobody enhanced it, and it was
removed.
Robert C. Martin | Design Consulting | Training courses offered:
Object Mentor | rma...@oma.com | Object Oriented Design
14619 N Somerset Cr | Tel: (800) 338-6716 | C++
Green Oaks IL 60048 | Fax: (847) 918-1023 | http://www.oma.com
>
>
>>
>>The genetic language is exceedingly complex. Its simple beginnings in the
>>four base nucleosides is misleading.
>
>Yes, but so is the simple instruction set of a modern RISC processor, the
>ADD instruction decodes into a series of many microcodes that actually
>performs the operation of enabling buffers, disabling memory devices, etc
>all with precise timing constraints. I think of the analogy as
>
> Biological Computational
>
> nucleotides (base 4) memory bit (base 2)
> all 64 codons the set of microcodes
> codon a specific microcode
> gene a machine instruction
> group of genes a subroutine
> RNA transcription
> and ADT interaction a procedure call, changing and creating the data
> protein the data
> shape of protein the state of the data (values of each attribute)
> skin the complete set of data for a high level class
>
>Obviously this analogy is much different than yours, and probabaly
>is absurb, as I changed it 10 times while writing this post and it still
has
>inconsistences,
>but... I believe the analysis of a true object oriented
>system would be equally complex if examined at the microcode level.
>
>
>>One such intriguing dependency is breathing. We breathe to keep the
levels
>
>>
>>This is the kind of bizzare interdependency that a good OO designer would
>>use the tools of OO to break. So, in the final analysis, I would have to
>>say that the genetic code is not OO, nor does it lead to a system that is
>>OO, simply because the system has some rather nasty interdependencies that
a
>>good OO designer would not allow.
>>
>>
>
Robert C. Martin
>We design flight controls for positive feedback fighter aircraft, does that
>mean we can't use OO techniques to code the control laws? Nasty
>interdependecies are the nature of physical laws and can't be avoided.
They can be avoided in the code.
>The purpose of OO is to reduce interdependencies to a minimum,
>but NOT any smaller. Any complex system is going to have bizzare convoluted
>strange interdependencies, but if they were reduced to a minimum by OO
design
>techniques, then OO has done its job to the best of its abilities and the
>code can rightly be called OO. Has our DNA reduced this interdependencies
>to a minimum? I don't think we know the answer to that question yet.
In a computer system, the necessary dependencies exist at runtime. However,
at compile time, they can all be broken. There is no limit to the
decoupling one can exert upon the source code.
LAMP-PU...@worldnet.att.net
>
>In a computer system, the necessary dependencies exist at runtime. However,
>at compile time, they can all be broken. There is no limit to the
>decoupling one can exert upon the source code.
>
Minimum coupling is a constant, you can attempt to hide in high level abstraction
but it is still there in the source code. Its just seems to disappear under
the standard measurement systems you use.
steve
Graham Perkins
> Well anyway, where I work we have been programming in procedural style
> type methods for decades (using only assembly language). After reviewing
> OO methods and re-evaluating our code, we have come to the conclusion that
> much of our existing code is actually objected oriented. How could this be?
> We didn't know anything about OO methods when we were coding. Hmmm.
Most of the techniques used in OO programming to increase
engineered quality of software can be done without OO languages.
After all, how did OO language features come to be implemented
and widely accepted? People were trying to manage without them,
and demanding them.
> Well, assuming OO code is superior to procedural code,
Contentious!
First off, find some non-procedural OO code to show us!
> In addition many of the features of C++, Eiffel and other high level
> OO languages such as type checking, variable scoping, have been
> incorporated into our assembly language code indirectly via
> preprocessors, code reviews, and specification enhancements.
Type checking and variable scoping sounds a bit thin.
Unless you used pre-proc support to make safe programming
with type unions possible .. which would enable polymorphism.
> So the question is with our genome, since it is the oldest
> program known, is it possible it is objected oriented in design?
Sorry, that is a non-question. Our genetic code is not designed.
Nor is it a program. That's just a shorthand analogy useful in
certain cases.
> some light reading in genetics it seems that scientists are
> having a hard time trying to figure out how that "program" works.
> But I would too, if I were only examining the executable code
> generated from C++ or Eiffel.
I suspect that a bit of reverse engineering, a bit of AI,
a bit of encoded domain knowledge, and a few decades of
investment comparable to that in genetic research and engineering,
could produce excellent de-compilers and reverse specification
generators. :-)
> Or perhaps our genome structure is far more
> advanced than OO.
Probably .. it's been "brewing" for four billion years.
> Anybody know of any books on this subject?
Check out the evolutionary biology popularisation texts by Stephen J
Gould.
Then read some Richard Dawkin (The Selfish Gene) to ensure your
remaining
optimism is completely stamped out.
----------------------------------------------------------------
Graham Perkins, De Montfort University, Milton Keynes
http://www.mk.dmu.ac.uk/~gperkins/
Paul E. Bennett
> Does that mean FORTH is an OO system? Eeeep!
Having asked that question before and answered it to my satisfaction the
conclusion I came to was:-
Forth is not an OOP of itself, but exhibits some of the facets of an OOPL
and through its extensibility can be moulded into a full OOPL. Take a look
at the work of those who have been very heavily involved in merging the
OO wordset specifications to create a workable basis for adding the OO
features.
--
Paul E. Bennett ................... <p...@tcontec.demon.co.uk>
Transport Control Technology Ltd. <http://www.tcontec.demon.co.uk/>
+44 (0)117-9499861 <enq...@tcontec.demon.co.uk>
Going Forth Safely
Joachim Durchholz
>
> >4. Are there several different kinds of objects that are controlled
> >by the exact same interface?
>
> Yes. Many sequences of DNA code the same sequences over and over
> again, sometimes with minor variations. Meta-classes.
Sounds rather like "reuse by cut&paste" to me, which isn't even
structured programming...
> >5. Are there subroutines that don't care what kind of object they are
> >manipulating because they simply use an interface?
>
> Some molecules chemically match those of a similar shape (THC and
> dopamine, for example, or aspirin and acetaminophene).
Well, look at the system crashes that this causes :)
(A more appropriate analogy might be interpreting the contents of a
screen bitmap as executable code ...)
> Is a ribosome a Turing engine?
No. It cannot write the tRNA it's decoding.
It's more like an old Jaquard weaving machine (read it off a punched
card and produce patterns as indicated by the holes).
Ell
>LAMP-PU...@worldnet.att.net wrote:
>>
>> So the question is with our genome, since it is the oldest
>> program known, is it possible it is objected oriented in design?
>Sorry, that is a non-question. Our genetic code is not designed.
It doesn't have a design - a shape, structure, architecture - after
millions of years of evolution?
Wouldn't it incorporate OO features being that OO features are fairly
common throughout the real word?
Juergen Schlegelmilch
>[ DNA ]
>Wouldn't it incorporate OO features being that OO features are fairly
>common throughout the real word?
_Your view_ of the real world incorporates OO features, the world
does not necessarily. There are many valid views on what the real world
is, among them the OO viewpoint, but it is not the only one, nor the
only correct one, nor necessarily the best one under all circumstances.
Please do not mistake your viewpoint/opinion/mental model of the world
for reality.
As for DNA, it can certainly be modelled using UML, and its behaviour can
be modelled using OO methods. If it is beneficial for your application,
do it; if not, why should you try to do it?
Regards,
Jürgen
--
/-----------------------------------------------------------------------------\
Juergen Schlegelmilch http://www.informatik.uni-rostock.de/~schlegel
Database Research Group mailto:schl...@Informatik.Uni-Rostock.de
University of Rostock, Computer Science Department, 18051 Rostock, Germany
Ell
wrote:
>On Wed, 28 Oct 1998 04:26:22 GMT, Ell <e...@access.digex.net> wrote:
>>[ DNA ]
>>Wouldn't it incorporate OO features being that OO features are fairly
>>common throughout the real word?
>_Your view_ of the real world incorporates OO features, the world
>does not necessarily.
"the world does not necessarily"? Either it does or it doesn't. I
see where the real world does in fact have OO features.
>There are many valid views on what the real world
>is, among them the OO viewpoint, but it is not the only one, nor the
>only correct one, nor necessarily the best one under all circumstances.
While the world has OO features, OO is not always the best way to
model the world in all contexts. I know that. That is why I also use
various other forms of modelling: algebra, structured programming,
statistics, ER diagrams, etc.
>Please do not mistake your viewpoint/opinion/mental model of the world
>for reality.
It's not a mistake, but fact. I would ask you try to see those facts.
The facts of all the major OO features being in the real world. If
you want examples please see the paper "Broad New Vistas" at my web
site.
>As for DNA, it can certainly be modelled using UML, and its behaviour can
>be modelled using OO methods. If it is beneficial for your application, do it;
>if not, why should you try to do it?
I agree. For whatever you are doing use the best modelling paradigm
for the context. In most programming cases, I find OO to be the best.
Getting back to the history of the thread, the writer asked if anyone
saw OO features in DNA and genomes. I answered yes. That does not
mean that I would always use OO to model every situation involving DNA
and genomes.
But I make no bones about stating that OO features are objectively in
the world, in DNA and in genomes, and that I see OO as the best
modelling paradigm for many if not most programming cases and
contexts.
Please don't rain on my parade because you don't see the bright sunny
day of OO. :-}
Juergen Schlegelmilch
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>On Wed, 28 Oct 1998 04:26:22 GMT, Ell <e...@access.digex.net> wrote:
>>>[ DNA ]
>>>Wouldn't it incorporate OO features being that OO features are fairly
>>>common throughout the real word?
>
>>_Your view_ of the real world incorporates OO features, the world
>>does not necessarily.
>
>"the world does not necessarily"? Either it does or it doesn't. I
>see where the real world does in fact have OO features.
Two points, closely related:
1. Whether the world does or does not incorporate OO features is mostly
irrelevant since we always handle models of the world. Whether we build
these models using OO features or not depends on whether it fits our
needs for a specific application. Everyone has a mental model of the
world, but because everyone has differing experiences, all those models
differ from each other. For specific areas, we can usually agree on a
very large set of commonalities, so one can get the impression that this
set is _the_ model of the real world. I doubt that this is possible in
general, because forming a model relies on judgement which is subjective.
If two models are equally powerful (i.e. represent the same information)
and consistent (with themselves), it boils down to personal preferences.
2. You wrote "I see where the real world does in fact have OO features".
This is almost literally what I meant: from your point of view, the world
incorporates OO features. That's ok, but does not imply that anyone else
must also see this. There may be other mental models, as consistent as the
OO model that you have built using your experience and perception. We
should pick the model based on our needs, not on believes.
>>Please do not mistake your viewpoint/opinion/mental model of the world
>>for reality.
>
>It's not a mistake, but fact. I would ask you try to see those facts.
Even if I saw this, it would just be another view on the world
(literally!), another mental model (most probably mostly agreeing
with yours), not the world itself.
Any "fact" can only be proven within a model that provides the axioms
and inference rules; if one assumes a specific model, then one can
prove that the world conforms to that model (i.e. there is no
experiment that contradicts the theorems built using the model).
That does not prove that there are no other models possible in
which the same experiment also verifies the theorems of that
model. As long as all models are consistent in themselves and
with our perception, they are equally valid; there is then no one
model that "better" describes the world (except where tracability
and understandability, i.e. processing by humans, comes into
play; some models are more easily understood than others, depending
on the background of the reader). The world only gives us results
(of experiments), we draw conclusions based on models.
>Please don't rain on my parade because you don't see the bright sunny
>day of OO. :-}
But I think I do see the sunshine, since I too think that the OO
paradigm is well-suited for many if not most programming cases
(that I have encountered; there may be whole domains where it is
not appropriate). However, this is not because I think the world
is object-oriented but because my mindset makes OO specifications
easier to understand for me, and the OO techniques help building
robust yet flexible software systems.
Regards,
Jürgen Schlegelmilch
Ell
>Elll <e...@access.digex.net> wrote:
>>
>>Please don't rain on my parade because you don't see the bright sunny
>>day of OO. :-}
>But I think I do see the sunshine, since I too think that the OO
>paradigm is well-suited for many if not most programming cases
>(that I have encountered; there may be whole domains where it is
>not appropriate). However, this is not because I think the world
>is object-oriented but because my mindset makes OO specifications
>easier to understand for me, and the OO techniques help building
>robust yet flexible software systems.
We'll have to agree to disagree. My own and others experience has
demonstrated to us that OO features, some major ones being:
a.) entities with behavior and properties
b.) encapsulation
c.) polymorphism
d.) identity
e.) state
f.) commonality
exist objectively in the world.
Our experience has shown not that OO are the only features of the
world, not that we can't and at times shouldn't model in other ways,
just that OO features do exist objectively.
Juergen Schlegelmilch
>We'll have to agree to disagree. My own and others experience has
>demonstrated to us that OO features, some major ones being:
>exist objectively in the world.
Ok, so we disagree. One point where we obviously disagree,
is your interpretation of `objectively'. You and others may
see the world from an OO point of view, but that is _your_
view on the world -- just as I have a certain view on the
world, and neither is identical to the world itself. Any
view is subjective, by the very definition (the subject is
the one having the view; there is no view without an observer).
Anyway, you already had discussions about this with others
(what `reality' is, what it means `to prove' etc.), and I'll
check dejanews for that, to spare us both the repetition.
BTW, all this does not invalidate any advantages that the
OO paradigm offers for software development, be it analysis
or design.
Regards,
Jürgen
Ell
wrote:
>On Wed, 28 Oct 1998 14:35:39 GMT, Ell <e...@access.digex.net> wrote:
>>We'll have to agree to disagree. My own and others experience has
>>demonstrated to us that OO features, some major ones being:
>[ list of OO features snipped ]
>>exist objectively in the world.
>Ok, so we disagree. One point where we obviously disagree,
>is your interpretation of `objectively'.
No, it seems to me that our understanding of the word is the same -
something that exists independently of the mind. It's just that you
don't think OO features exist objectively whereas I do.
>You and others may see the world from an OO point of view, but
>that is _your_ view on the world
Again I disagree, just as I did earlier. I understood what you meant
earlier and I understand what you mean now.
I'm telling you what myself and others have found to be true - that OO
features exist objectively, not just as a view.
Tim Ottinger
>
> On Wed, 28 Oct 1998 14:35:39 GMT, Ell <e...@access.digex.net> wrote:
> >We'll have to agree to disagree. My own and others experience has
> >demonstrated to us that OO features, some major ones being:
> [ list of OO features snipped ]
> >exist objectively in the world.
>
> Ok, so we disagree. One point where we obviously disagree,
> see the world from an OO point of view, but that is _your_
> world, and neither is identical to the world itself. Any
> view is subjective, by the very definition (the subject is
> the one having the view; there is no view without an observer).
Like Ralph Johnson likes to say (or did) "The metaphor is not
the thing". Certainly there are real-world(tm) analogies and
metaphors to the OO features. I'd never deny that, especially
considering the roots that OO had in the biological model and
in simulation. But two people meeting and conversing is not
*exactly* OO double-dispatch. Playing a role in a play is
not exactly the same thing as implementing an interface, and
swapping drivers at the end of a bus shift is not exactly
OO polymorphism. However, they're good analogies.
I have a like problem with the use of "objectively", but
I do concur that features in OO can be described via analogy
to the real world. I also have a chicken-and-egg problem,
here. The real world isn't OO, OO is the analogy to situations
in the real world. Even simulation is only able to create
a thin shadow of some aspects of the real world. The world is
far more wonderful than any OO.
Tim
Phlip
>Like Ralph Johnson likes to say (or did) "The metaphor is not
>the thing". Certainly there are real-world(tm) analogies and
>metaphors to the OO features. I'd never deny that, especially
>considering the roots that OO had in the biological model and
>in simulation. But two people meeting and conversing is not
>*exactly* OO double-dispatch. Playing a role in a play is
>not exactly the same thing as implementing an interface, and
>swapping drivers at the end of a bus shift is not exactly
>OO polymorphism. However, they're good analogies.
>
>I have a like problem with the use of "objectively", but
>I do concur that features in OO can be described via analogy
>to the real world. I also have a chicken-and-egg problem,
>here. The real world isn't OO, OO is the analogy to situations
>in the real world. Even simulation is only able to create
>a thin shadow of some aspects of the real world. The world is
>far more wonderful than any OO.
But the problem with this question is it's _not_ a metaphor. DNA _is_ a
programming language - it just may or may not be OO! We have a
"chicken-and-egg problem" in more ways than one.
Others have noted that DNA objects are absurdly and self-referentially
coupled, and that DNA programmers work in project life cycles of millions of
years. Neither of these matter to DNA like they do to us. Encapsulation and
RAD are common design requirements of software projects, but we should not
let our learned sensitivities trigger a determination that DNA is not OO
because it does not follow our parochial coding standards.
-- Phlip at politizen dot com (address munged)
======= http://users.deltanet.com/~tegan/home.html =======
-- Linux. We "get" it. --
mmo...@wiretel.com
Joachim Durchholz <joachim....@munich.netsurf.de> wrote:
> LAMP-PU...@worldnet.att.net wrote:
> >
> > Well anyway, where I work we have been programming in procedural style
> > type methods for decades (using only assembly language).
> > In addition many of the features of C++, Eiffel and other high level
> > OO languages such as type checking, variable scoping, have been
> > incorporated into our assembly language code indirectly via
> > preprocessors, code reviews, and specification enhancements.
>
> work here :)
>
> > So the question is with our genome, since it is the oldest program
> > known, is it possible it is objected oriented in design?
>
> Anything that comes out of evolution isn't "designed" at all, to begin
> with. It just happened to work out.
I'm not sure whether the original question asked if nature used an OO model
to create a human body. I think the original author is asking if his program
models human genetics in an OO fashion. Certainly nature didn't use OO to
design and extend the human body, but that doesn't mean we can't attempt to
use OO to allow our software to simulate bodily functions.
> When looking at the interactions in the human body, I see a wild mixture
> of physical, electrical, and neural signal transmission. Nearly none of
> these signals has a clearly defined recipient, nor is any one signal a
> clear indication of some condition. Everything "talks" to everything -
> the exact opposite of good engineering.
> Of course, Nature has no business in engineering. There's nobody around
> who wants or needs to understand how a human body works. So there's no
> point in *complaining* that Nature "does" it the unstructured way - it's
> just that humans cannot use this for design. (Note that we can still
> learn from Nature's tricks. It's just that I doubt that we can copy any
> design principles: our brains aren't geared towards understanding
> tightly interwoven, complex interactions, so a system with an
> evolution-style design would not be understandable to a human,
> regardless of how well it did its job.)
I sincerely believe that if we have a good understanding of any domain,
including interactions within a human, we can model that domain with OO. If
we are lacking knowledge of what body part interacts with which other part or
how interaction is accomplished, we would be hard pressed to model that
domain with any tool. In the case of the overall human body and brain
interactions, its the domain knowledge that is lacking.
Mike Moore
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Gerry Quinn
>Sounds rather like "reuse by cut&paste" to me, which isn't even
>structured programming...
>
Worse than that - the cut and paste is random, and if the system
crashes, then maybe a slightly different cut and paste will work!
The products are surprisingly good, but the development time has been
many trillions of creature-years.
- Gerry
----------------------------------------------------------
ger...@indigo.ie (Gerry Quinn)
----------------------------------------------------------
Mark Addleman
Gerry Quinn wrote in message <71898b$q88$1...@news.indigo.ie>...
>The products are surprisingly good, but the development time has been
>many trillions of creature-years.
I never want to hear anyone complain about a schedule slip again!
Ell
>Ell <e...@access.digex.net> wrote:
>>
>>We'll have to agree to disagree. My own and others experience has
>>demonstrated to us that OO features, some major ones being:
>> a.) entities with behavior and properties
>> b.) encapsulation
>> c.) polymorphism
>> d.) identity
>> e.) state
>> f.) commonality
>>exist objectively in the world.
>You and others may
>see the world from an OO point of view, but that is _your_
>view on the world
You mean to tell me that you don't see these features existing
objectively in the world?
I can look around right now at the wastebasket, the chair, my pants,
the phone, the keyboard and they all exhibit OO features. I can test
to see if the OO features are objective by formulating my OO feature
observation as hypothesis which can be tested. When I test those
hypothesis, the hypothesis are proven as true. Hence I know that OO
features exist objectively. My view of OO features in the things
around me is not just subjective gut feeling.
Juergen Schlegelmilch
>Juergen Schlegelmilch wrote:
>>Ell <e...@access.digex.net> wrote:
>>>We'll have to agree to disagree. My own and others experience has
>>>demonstrated to us that OO features, some major ones being:
>
>>You and others may see the world from an OO point of view, but that
>>is _your_ view on the world
>
>You mean to tell me that you don't see these features existing
>objectively in the world?
I see them in the world or, more precisely, I see aspects of the
world that I explain to myself using the terms that you listed
(not all of them but that is just a matter of degree). But that
is only a view with a mental model behind it to explain features
that are not directly visible. Whether the world really is built
that way, is an open question, and I have no way to answer it
definitely since _all_ my experiences are indirect, via my senses.
So, there is no way to determine whether any of these features
exists `objectively' in the world.
>I can look around right now at the wastebasket, the chair, my pants,
>the phone, the keyboard and they all exhibit OO features. I can test
>to see if the OO features are objective by formulating my OO feature
>observation as hypothesis which can be tested. When I test those
>hypothesis, the hypothesis are proven as true. Hence I know that OO
>features exist objectively. My view of OO features in the things
>around me is not just subjective gut feeling.
"I look" is equivalent to "I have a view", which is subjective
(you being the subject that looks around). Looking around is an
experiment, and you then test whether the result fits into your
mental model. If it does, then this proves that your model
conforms to the observations. This does not say anything about
objectivity or subjectivity. I'd call something `objective' if
everyone agrees on it; otherwise it is subjective. If there is
one that does not agree, then one has to conclude that the
result is subjective.
You also have to be careful about direct results and indirect
conclusions. For example, everyone will agree that there are
apples, and if you put an apple 1m above ground and open your
fingers, it will fall down; these are all direct results. This
experiment does not prove that gravity exists, because that is
a conclusion, based on a model. The commonly agreed model for this
physicalprocess is the force of gravity with its laws, and all we
can do is test whether our perception conforms to this model. If
the test succeeds, then this does not say anything about the real
nature of the physical process that we observed; it merely shows
that your model is able to explain this aspect of the world.
The history of science is full of models that turned out to be
wrong or inaccurate after they had been accepted for quite a
while. Of course the world has not changed because we withdrew
some of our models; it simply exists, and will do so independently
of all our attempts to describe it properly.
Regards,
Jürgen
PS: All this has been discussed at length in the thread `What does
"Real" mean?' in June of this year. A related thread in July/August
of 1997 (Subject: Axioms etc.) shows that you use a different
interpretation of `proof' than most others do (i.e. something can be
proven temporarily).
PPS: One has to be careful even with direct results because our
senses are not perfect. There are several visual tricks that
show that our eyes betray us in certain circumstances. Even
our nerves are not perfect: neurologists can induce visual
patterns into our conciousness using electric impulses. Taken
to the extreme, one cannot even be sure that there exists
anything except the own conciousness (read Hofstadter and
Dennett "The Mind's I" for a very interesting discussion of
all this).
Ell
wrote:
>On Thu, 29 Oct 1998 03:57:29 GMT, Ell <e...@access.digex.net> wrote:
>>Juergen Schlegelmilch wrote:
>>>Ell <e...@access.digex.net> wrote:
>>>>We'll have to agree to disagree. My own and others experience has
>>>>demonstrated to us that OO features, some major ones being:
>[ list of OO features snipped ]
>>>>exist objectively in the world.
>>
>>>You and others may see the world from an OO point of view, but that
>>>is _your_ view on the world
>>
>>You mean to tell me that you don't see these features existing
>>objectively in the world?
>
>I see them in the world or, more precisely, I see aspects of the
>world that I explain to myself using the terms that you listed
>(not all of them but that is just a matter of degree). But that
>is only a view with a mental model behind it to explain features
>that are not directly visible. Whether the world really is built
>that way, is an open question, and I have no way to answer it
>definitely since _all_ my experiences are indirect, via my senses.
>So, there is no way to determine whether any of these features
>exists `objectively' in the world.
The way to determine what is objective is the scientific method.
>>I can look around right now at the wastebasket, the chair, my pants,
>>the phone, the keyboard and they all exhibit OO features. I can test
>>to see if the OO features are objective by formulating my OO feature
>>observation as hypothesis which can be tested. When I test those
>>hypothesis, the hypothesis are proven as true. Hence I know that OO
>>features exist objectively. My view of OO features in the things
>>around me is not just subjective gut feeling.
>"I look" is equivalent to "I have a view", which is subjective
>(you being the subject that looks around).
OK.
>Looking around is an experiment,and you then test whether
>the result fits into your mental model.
Taking for granted that the "looking around" is controlled experiment.
>If it does, then this proves
>that your model conforms to the observations. This does not say
>anything about objectivity or subjectivity.
If the evidence from "looking around" comes from controlled experiment
why isn't that objective?
"I look" is equivalent to "I have a view", which is subjective
(you being the subject that looks around). Looking around is an
experiment, and you then test whether the result fits into your
mental model. If it does, then this proves that your model
conforms to the observations. This does not say anything about
objectivity or subjectivity. I'd call something `objective' if
everyone agrees on it; otherwise it is subjective. If there is
one that does not agree, then one has to conclude that the
result is subjective.
>I'd call something `objective' if everyone agrees on it;
I disagree. For me what is objective is what has been proved through
controlled experiment.
>You also have to be careful about direct results and indirect
>conclusions. For example, everyone will agree that there are
>apples, and if you put an apple 1m above ground and open your
>fingers, it will fall down; these are all direct results. This
>experiment does not prove that gravity exists, because that is
>a conclusion, based on a model. The commonly agreed model for this
>physicalprocess is the force of gravity with its laws, and all we
>can do is test whether our perception conforms to this model. If
>the test succeeds, then this does not say anything about the real
>nature of the physical process that we observed; it merely shows
>that your model is able to explain this aspect of the world.
>The history of science is full of models that turned out to be
>wrong or inaccurate after they had been accepted for quite a
>while.
Yes. to confirm a theory it typically requires more than one result
confirming the predictions of the theory.
>Of course the world has not changed because we withdrew
>some of our models; it simply exists, and will do so independently
>of all our attempts to describe it properly.
True, the world will exist independently of our theories and
hypotheses, but that doesn't mean that our theories and hypotheses are
never correct either.
>PS: All this has been discussed at length in the thread `What does
> "Real" mean?' in June of this year. A related thread in July/August
> of 1997 (Subject: Axioms etc.) shows that you use a different
> interpretation of `proof' than most others do (i.e. something can be
> proven temporarily).
The point I making is that sometimes the world exists objectively in a
certain way for a limited time period. So that once the conditions
change it may not even be possible to test a theory/hypothesis. But I
want to make clear that as long as specific conditions hold the
predictions of a correct theory/hypothesis should hold.
>PPS: One has to be careful even with direct results because our
> senses are not perfect. There are several visual tricks that
> show that our eyes betray us in certain circumstances. Even
> our nerves are not perfect: neurologists can induce visual
> patterns into our conciousness using electric impulses. Taken
> to the extreme, one cannot even be sure that there exists
> anything except the own conciousness (read Hofstadter and
> Dennett "The Mind's I" for a very interesting discussion of
> all this).
That is why a.) we run controlled experiments, and b.) we run
controlled experiments on controlled experiments {i.e. we test how
experiments were carried out}.
Ell
>Juergen Schlegelmilch wrote:
>>
>>Ell <e...@access.digex.net> wrote:
>>>
>>>Juergen Schlegelmilch wrote:
>>>>
>>>>Ell <e...@access.digex.net> wrote:
>>>>>
>>>>>We'll have to agree to disagree. My own and others experience has
>>>>>demonstrated to us that OO features, some major ones being:
>>>>>a.) entities with behavior and properties
>>>>>b.) encapsulation
>>>>>c.) polymorphism
>>>>>d.) identity
>>>>>e.) state
>>>>>f.) commonality
>OK.
********** THIS WAS J. SCHLEGELMILCH NOT ELLIOTT ******************
>"I look" is equivalent to "I have a view", which is subjective
>(you being the subject that looks around). Looking around is an
>experiment, and you then test whether the result fits into your
>mental model. If it does, then this proves that your model
>conforms to the observations. This does not say anything about
>objectivity or subjectivity. I'd call something `objective' if
>everyone agrees on it; otherwise it is subjective. If there is
>one that does not agree, then one has to conclude that the
>result is subjective.
*********************************************************************
J. Schlegelmilch continues:
Graham Perkins
> I sincerely believe that if we have a good understanding of any domain,
> including interactions within a human, we can model that domain with OO.
well you need more, since OO is largely about how encapsulated
objects interact. you'd also need a load of stats, maths, chemistry,
and maybe even quantum mechanics .. whether or not you bundle your
model in OO fashion is a minor issue!
however, you are argument falls down for other reasons. Most
biological systems are extremely complex. Axiomatic models
of them would undoubtedly be more powerful than Peano Natural
Number axioms. Probably they would need framework as powerful
as 2nd order logic, and probably you could describe systems
as complex as Turing machines. So you'd be stuck with
inconsistency, incolpleteness, intractability, undecidability,
NP-completeness....
Tim Ottinger
> I'm not sure whether the original question asked if nature used an OO model
> design and extend the human body, but that doesn't mean we can't attempt to
> use OO to allow our software to simulate bodily functions.
>
> including interactions within a human, we can model that domain with OO.
Thank you for being a kindred spirit. This is what I've been trying to
say, and lacking eloquence for, for years. Everything is not an
object,
but you can make an object for anything.
Richard MacDonald
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch)
>wrote:
>
>The way to determine what is objective is the scientific method.
But you cannot escape the fact that what the scientific method gives you is
a *view* on the real world, not the real world itself. This is the
fundamental issue that you keep refusing to deal with. It cannot be made
much clearer than Juergen Schlegelmilch's post and many others before, so
I'm going to limit my repetition here.
>[ ] For me what is objective is what has been proved through
>controlled experiment.
But the scientific world can only prove the double negative or null
hypothesis. It cannot say something is "true". It can only say that the
model fits the facts *so far*. That is a very shaky objectiveness. Was the
Newton model objective? If you thought so, you would have been wrong when
Einstein came along.
>
>Yes. to confirm a theory it typically requires more than one result
>confirming the predictions of the theory.
To confirm a theory only means that you have one more piece of evidence that
does not prove it incorrect.
>>Of course the world has not changed because we withdrew
>>some of our models; it simply exists, and will do so independently
>>of all our attempts to describe it properly.
>
>True, the world will exist independently of our theories and
>hypotheses, but that doesn't mean that our theories and hypotheses are
>never correct either.
Depends how you define "correct". The fact that the world is independent of
our theories means that our theories are incomplete. I think you are arguing
for correctness as an absolute, which cannot be correct. Care to clarify how
something can be correct but not complete?
>>PS: All this has been discussed at length in the thread `What does
>> "Real" mean?' in June of this year.
And Ell battled bravely in that thread as well.
>
>>PPS: One has to be careful even with direct results because our
>> senses are not perfect. There are several visual tricks that
>> show that our eyes betray us in certain circumstances. Even
>> our nerves are not perfect: neurologists can induce visual
>> patterns into our conciousness using electric impulses. Taken
>> to the extreme, one cannot even be sure that there exists
>> anything except the own conciousness (read Hofstadter and
>> Dennett "The Mind's I" for a very interesting discussion of
>> all this).
>
>That is why a.) we run controlled experiments, and b.) we run
>controlled experiments on controlled experiments {i.e. we test how
>experiments were carried out}.
>
And we run controlled experiments on our controlled experiments on our
controlled experiments.
And we run controlled experiments on our controlled experiments on our
controlled experiments on our controlled experiments.
And we run controlled experiments on our controlled experiments on our
controlled experiments on our controlled experiments on our controlled
experiments.
And we run controlled experiments on our controlled experiments on our
controlled experiments on our controlled experiments on our controlled
experiments on our controlled experiments.
Ell, would you please finish this thread for me? Note: "etc" is not the
correct answer.
Juergen Schlegelmilch
>>>>>OO features [ list snipped ] exist objectively in the world.
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>I see them in the world or, more precisely, I see aspects of the
>>world that I explain to myself using the terms that you listed
>>(not all of them but that is just a matter of degree). But that
>>is only a view with a mental model behind it to explain features
>>that are not directly visible. Whether the world really is built
>>that way, is an open question, and I have no way to answer it
>>definitely since _all_ my experiences are indirect, via my senses.
>>So, there is no way to determine whether any of these features
>>exists `objectively' in the world.
Ell <e...@access.digex.net> wrote:
>The way to determine what is objective is the scientific method.
There seem to be two points of disagreement:
- What the scientific method achieves
- What "objective" means
The scientific works like this: You have a model (a set axioms and
inference rules) which you want to test. You then choose one of the
theorems, design an experiment for it, and perform that experiment.
There are only two possible outcomes: either the results of the
experiment match the predictions of the theorem, or they do not match.
In the first case, you now know that your modell conforms to reality
_in this specific experiment_; you still do not know whether seemingly
similar experiments will also match the predictions of the model. In
the second case, your model is simply wrong; you do not know whether
a seemingly similar model is also wrong.
Obviously, you cannot prove that your model is _identical_ to reality
since this requires to perform _all_ experiments at the same time
(time is part of the experiment conditions). Your model may predict
that some parameters (such as time, location, amount of light etc.)
are not important. However, reality is not bound to follow your model
even if the model has been shown to correctly predict a huge number
of experiment results. So, it is important to be aware that reality
and any of the models that we have are not identical. Our models
can use OO features; reality is free to use others.
Now, what does "objective" mean? We could look up a dictionary,
e.g. my copy of Oxford Advanced Learner's Dictionary of Current
English (not that I think it is definitive in any way):
objective: 1. (in philosopy) having existence outside of mind;
real. see "subjective". 2. (of persons, writings, pictures)
uninfluenced by thought or feeling; dealing with outward
things, actual facts, etc uninfluenced by personal feelings
or opinions. 3. (gram) of the object: the objective case, in
Latin and other inflected languages.
I interpreted it like (2) above, and on second thought, you
probably meant it as described by (1) above.
In our particular case, both meanings are quite related: I think
that the OO features that you see are part of your model of the
world, while you think that they are part of the world itself.
With meaning (2), something is "objective" if everyone's model
agrees on it, so that personal opinion is not important. With
meaning (1), everyone's model _must_ agree on it, or be
inconsistent with the world.
I hope to have shown in the discussion of the scientific method,
that one cannot prove that one's model is identical to the world
(you may want to read the "Axioms etc" thread again, in which
several other people have explained the same thing; I realize that
you ended the discussion with a remark that you believe something
to be proven if enough experiments have been performed and shown
comformance of the model with the world; see messages
<5rt09q$ib5$1...@news2.digex.net>, <EE47n...@ecf.toronto.edu>. It
is just that, namely a believe, not a proof.).
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>"I look" is equivalent to "I have a view", which is subjective
>>(you being the subject that looks around).
Ell <e...@access.digex.net> wrote:
>OK.
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>Looking around is an experiment,and you then test whether
>>the result fits into your mental model.
Ell <e...@access.digex.net> wrote:
>Taking for granted that the "looking around" is controlled experiment.
I'm not sure what you want to control here. Can you explain
what you mean by "controlled experiment" in this context? We are
talking about a simple look out of the window.
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>If it does, then this proves
>>that your model conforms to the observations. This does not say
>>anything about objectivity or subjectivity.
Ell <e...@access.digex.net> wrote:
>If the evidence from "looking around" comes from controlled experiment
>why isn't that objective?
Because you are subjective: you are the one interpreting the results.
Basically, what this experiment yields is a pair of images (from your
two eyes). Everything beyond that is already interpretation and
subject to error. Let's drop basic levels of interpretation (such as
reckognizing physical elements) and assume that we both see a car in
front of your house. You call it an object; however, "object" is a
term of your model, the world only has a car. To what term or category
in your model this car corresponds, is beyond the realm of the world.
The world simply _is_, it does not interpret; humans do, based on
models.
To clarify [note: I do not want to put words into your mouth, so the
following is a gross simplification of how I think you would phrase
certain things; I may be wrong (I'm not a native speaker of English)
and apologize in advance]:
you might say "there are objects in the world", I would rephrase this
to "there are elements in the world that can be called `objects'".
The same holds for, e.g., polymorphism: you'd say "objects are treated
polymorphically in the world", while I'd say "the treatment of objects
in the world is modelled by the concept `polymorphism'".
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>You also have to be careful about direct results and indirect
>>conclusions. For example, everyone will agree that there are
>>apples, and if you put an apple 1m above ground and open your
>>fingers, it will fall down; these are all direct results. This
>>experiment does not prove that gravity exists, because that is
>>a conclusion, based on a model. The commonly agreed model for this
>>physicalprocess is the force of gravity with its laws, and all we
>>can do is test whether our perception conforms to this model. If
>>the test succeeds, then this does not say anything about the real
>>nature of the physical process that we observed; it merely shows
>>that your model is able to explain this aspect of the world.
>>The history of science is full of models that turned out to be
>>wrong or inaccurate after they had been accepted for quite a
>>while.
Ell <e...@access.digex.net> wrote:
>Yes. to confirm a theory it typically requires more than one result
>confirming the predictions of the theory.
One cannot finally confirm a theory, one can only falsify it.
If there are enough positive experiments, we put some faith
into the theory. Nevertheless, it may still turn out to be
inaccurate or wrong.
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>Of course the world has not changed because we withdrew
>>some of our models; it simply exists, and will do so independently
>>of all our attempts to describe it properly.
Ell <e...@access.digex.net> wrote:
>True, the world will exist independently of our theories and
>hypotheses, but that doesn't mean that our theories and hypotheses are
>never correct either.
Of course. Still, a theory about the world is not the world
itself, so any evidence that the OO model of the world is correct
(in the sense that it has correctly predicted the results of a
large number of experiments) does not mean that the world is
object-oriented. In particular, it does not exclude that there
are other models of the world that are also correct in the same
sense.
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>PS: All this has been discussed at length in the thread `What does
>> "Real" mean?' in June of this year. A related thread in July/August
>> of 1997 (Subject: Axioms etc.) shows that you use a different
>> interpretation of `proof' than most others do (i.e. something can be
>> proven temporarily).
Ell <e...@access.digex.net> wrote:
>The point I making is that sometimes the world exists objectively in a
>certain way for a limited time period. So that once the conditions
>change it may not even be possible to test a theory/hypothesis. But I
>want to make clear that as long as specific conditions hold the
>predictions of a correct theory/hypothesis should hold.
?? This sounds like you believe that the world changes, depending
on the correctness (in the sense described above) of our theories.
Above you agree that the world exists independent of our theories.
Under what conditions may it be impossible to test a theory/hypothesis?
Just as a thought experiment imagine that someone proves the OO model
to be wrong: does then the world transmogrify into a different world,
with no objects and no polymorphism? When some Greek proved that the
world is not flat, did it suddenly become a globe? Before that event,
did the world exist objectively in a flat way, from the very start up
to that moment? In the case of two competing theories, does the world
exist objectively in two different ways at the same time?
[ Note: this is meant to equate the OO model with the model of
a flat earth in any way other than that they are both theories
about the world. ]
schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>PPS: One has to be careful even with direct results because our
>> senses are not perfect. There are several visual tricks that
>> show that our eyes betray us in certain circumstances. Even
>> our nerves are not perfect: neurologists can induce visual
>> patterns into our conciousness using electric impulses. Taken
>> to the extreme, one cannot even be sure that there exists
>> anything except the own conciousness (read Hofstadter and
>> Dennett "The Mind's I" for a very interesting discussion of
>> all this).
Ell <e...@access.digex.net> wrote:
>That is why a.) we run controlled experiments, and b.) we run
>controlled experiments on controlled experiments {i.e. we test how
>experiments were carried out}.
That does not help: how do you control your senses? Just think
of your brain swimming in a nutrition liquid, with lots of cables
wired to its nerves which give you the illusion of the existence
of a world where light travels at 300.000km per second. How can
you prove that this is simply an illusion?
[ I'm playing devil's advocate here, but in principle the
above scenario cannot be excluded ]
Seriously, as I wrote at the beginning, every experiment has two
possible outcomes, where the negative one would falsify the theory
being under examination. Can you describe an experiment that could
falsify the OO model of the world? Which result would not be
interpretable in the OO model?
Regards,
Jürgen
Tim Ottinger
This whole OO/Real world thing would be better suited by actually
having some of the results of the "controlled experiments" provided
here. What theory has been tested in controlled conditions? What
exactly was the theory, and what assumptions/axioms were involved?
What was the experiment, and how is it conducted? What was the model
it was trying to support? What were the results, and how do we prove
"objectively" (not by feel) that the results support the theorum?
We talk about OO existing in the real world and that fact being
supported by controlled experiments to prove objectivity, but I
see no such experiment being performed. I see obvious things
restated as being analogous to OO only. A whole different thing.
The need for testability in theory is the whole drive behind the
search for OO principles that apply objectively, not just based
on an untested or untestable idea that one way or another is
"best" or "true". With the OO principles, and even with the
definition of "maintainability" I've been pushing here, we can
see and measure something.
Scientific theory is bloodless and sterile in one way-- science
doesn't much care what you like or think, it looks for facts
that are (not easily un)provable. But in this, it is quite
beautiful.
If there is a difference between art and science, it's that
art reaches for subjective interpretation as a form of
meaningfulness, and science bemoans it. Definition 2, I
believe.
Most of the arguments that eat most of the bandwidth here
most of the time tend to end with cries from one group to
please provide something hard and measurable, something
concrete and concise, and the other group becoming abusive
and claiming to "just know" and to have "seen the truth".
Which is scientific, and which is interpretation?
I think interpretation and artistry is a fine thing, but
there are rules and principles. These are critical for
judging (objectively) the quality of software design and
implementation.
1) Requirements must be met. Regardless of aesthetics
and preferences, the system must really work or there
is no point in writing it. All else is subordinate to
this one goal.
2) The software must be maintainable
a) You can find the part(s) where the change is made
easily.
b) You can make the change in that place.
c) You can prevent or control the propagation of the
change through the system.
d) Predicatable types of changes are especially easy
to make in terms of 4a, 4b, and 4c.
To me, all of software engineering has been focusing on
these simple ideas. A lot of opinion is based on focusing
on one or two of these ideas (sometimes to the exclusion
of all others).
Tim
mmo...@wiretel.com
gper...@dmu.ac.uk wrote:
> mmo...@wiretel.com wrote:
> > I sincerely believe that if we have a good understanding of any domain,
> > including interactions within a human, we can model that domain with OO.
>
> objects interact. you'd also need a load of stats, maths, chemistry,
> and maybe even quantum mechanics .. whether or not you bundle your
> model in OO fashion is a minor issue!
>
> however, you are argument falls down for other reasons. Most
> biological systems are extremely complex. Axiomatic models
> of them would undoubtedly be more powerful than Peano Natural
> Number axioms. Probably they would need framework as powerful
> as 2nd order logic, and probably you could describe systems
> as complex as Turing machines. So you'd be stuck with
> inconsistency, incolpleteness, intractability, undecidability,
> NP-completeness....
>
> ----------------------------------------------------------------
> Graham Perkins, De Montfort University, Milton Keynes
> http://www.mk.dmu.ac.uk/~gperkins/
>
I wish I understood more about Axiomatic and other modelling techniques in
order to continue this discussion. I would be curious to discover whether or
not OO techniques could encapsulate or mimic these other methodologies you
mentioned. Ignorance is bliss.
--
Mike Moore
Wireless Telecom
mmo...@wiretel.com
mmo...@wiretel.com
> Thank you for being a kindred spirit. This is what I've been trying to
> say, and lacking eloquence for, for years. Everything is not an
> object,
> but you can make an object for anything.
>
Not bad eloquence with that statement.
I would prefer to say: I can make an abstraction for anything - I hope.
Ell
>>>>Ell <e...@access.digex.net> wrote:
>>>>>
>>>>> ... My own and others experience has
>>>>> demonstrated to us that OO features, some major ones being:
>>>>> a.) entities with behavior and properties
>>>>> b.) encapsulation
>>>>> c.) polymorphism
>>>>> d.) identity
>>>>> e.) state
>>>>> f.) commonality
>>>>> exist objectively in the world.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>I see them in the world or, more precisely, I see aspects of the
>>>world that I explain to myself using the terms that you listed
>>>(not all of them but that is just a matter of degree). But that
>>>is only a view with a mental model behind it to explain features
>>>that are not directly visible. Whether the world really is built
>>>that way, is an open question, and I have no way to answer it
>>>definitely since _all_ my experiences are indirect, via my senses.
>>>So, there is no way to determine whether any of these features
>>>exists `objectively' in the world.
>Ell <e...@access.digex.net> wrote:
>>The way to determine what is objective is the scientific method.
>There seem to be two points of disagreement:
>- What the scientific method achieves
>- What "objective" means
I raise them together because I know science is the way to verify what
is objective.
>The scientific works like this: You have a model (a set axioms and
>inference rules) which you want to test. You then choose one of the
>theorems, design an experiment for it, and perform that experiment.
>There are only two possible outcomes: either the results of the
>experiment match the predictions of the theorem, or they do not match.
>In the first case, you now know that your modell conforms to reality
>_in this specific experiment_; you still do not know whether seemingly
>similar experiments will also match the predictions of the model. In
>the second case, your model is simply wrong; you do not know whether
>a seemingly similar model is also wrong.
>Obviously, you cannot prove that your model is _identical_ to reality
>since this requires to perform _all_ experiments at the same time
>(time is part of the experiment conditions).
Often time has no bearing; the conditions that have a bearing are
often the same regardless of time. This is true even for most quantum
physics. We can often show that regardless of time the other factors
that are truly meaningful are the same.
>Your model may predict
>that some parameters (such as time, location, amount of light etc.)
>are not important.
Yes, and as per my last paragraph we can often show that conditions
are the same for what is relevant regardless of time and possibly
other factors like location and light.
>However, reality is not bound to follow your model
>even if the model has been shown to correctly predict a huge number
>of experiment results.
Yes, our theory/model may not hold up as the full explanation. Or,
and this is important to note, our theory may be relegated to a
subordinate role within what is known at the time to be a better
overall theory. Just as Newtonian mechanics is subordinate to
Relativity.
>So, it is important to be aware that reality
>and any of the models that we have are not identical.
One can not infer that our theories/models are essentially inaccurate
in how they describe reality simply because they *may* not always be
accurate. Further I know that no theory can exactly match reality
because reality is infinite and changing. But often for what we are
interested in (a context) our theories may be essentially correct.
And we can say that for given conditions 'x', the result will be 'y'.
>Our models can use OO features; reality is free to use others.
Of course. But this doesn't mean that reality can't objectively have
OO features, as I and others have seen proof that it does.
Ell
>Now, what does "objective" mean? We could look up a dictionary,
>e.g. my copy of Oxford Advanced Learner's Dictionary of Current
>English (not that I think it is definitive in any way):
>
> objective: 1. (in philosopy) having existence outside of mind;
> real. see "subjective". 2. (of persons, writings, pictures)
> uninfluenced by thought or feeling; dealing with outward
> things, actual facts, etc uninfluenced by personal feelings
> or opinions. 3. (gram) of the object: the objective case, in
> Latin and other inflected languages.
>
>I interpreted it like (2) above, and on second thought, you
>probably meant it as described by (1) above.
Correct, I more mean #1. Though I can see how #2 might be a corollary
of #1, if one doesn't act on the things existing outside of the mind.
>In our particular case, both meanings are quite related: I think
>that the OO features that you see are part of your model of the
>world, while you think that they are part of the world itself.
I see OO as both. It is exists in the world independently of the
mind, and it is a modelling paradigm--a way of seeing world.
>With meaning (2), something is "objective" if everyone's model
>agrees on it, so that personal opinion is not important. With
>meaning (1), everyone's model _must_ agree on it, or be
>inconsistent with the world.
I'll accept your interpretation of #1 and #2 because I can see how
one would think that and to move on.
>I hope to have shown in the discussion of the scientific method,
>that one cannot prove that one's model is identical to the world
I think one can show for conditions 'x' then 'y' will occur and that a
theory has 'z' probability of being true.
>(you may want to read the "Axioms etc" thread again, in which
>several other people have explained the same thing; I realize that
>you ended the discussion with a remark that you believe something
>to be proven if enough experiments have been performed and shown
>comformance of the model with the world; see messages
Well I if I said that then I would correct that as per my last
sentence:
I think one can show for conditions 'x' then 'y' will occur and that a
theory/model has 'z' probability of being true.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>Looking around is an experiment,and you then test whether
>>>the result fits into your mental model.
>Ell <e...@access.digex.net> wrote:
>>Taking for granted that the "looking around" is controlled experiment.
>I'm not sure what you want to control here. Can you explain
>what you mean by "controlled experiment" in this context? We are
>talking about a simple look out of the window.
If "looking around" is an experiment as you just said then it isn't
simply looking out the window, right? All experiments should be
controlled. Or at least the statistical analysis of "looking around"
should have controls applied within it. You are saying look out the
window. at what? when? by what means?
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>If it does, then this proves
>>>that your model conforms to the observations. This does not say
>>>anything about objectivity or subjectivity.
>Ell <e...@access.digex.net> wrote:
>>If the evidence from "looking around" comes from controlled experiment
>>why isn't that objective?
>Because you are subjective: you are the one interpreting the results.
But if you perform experiments scientifically then the results are not
subjective.
>Basically, what this experiment yields is a pair of images (from your
>two eyes). .... assume that we both see a car in
>front of your house. You call it an object; however, "object" is a
>term of your model, the world only has a car.
I might also call it a car myself. I might call it a "car object". I
might also call it a platform for standing upon. It depends on
context.
>To what term or category
>in your model this car corresponds, is beyond the realm of the world.
Right it is a thing in the world.
>The world simply _is_, it does not interpret; humans do, based on
>models.
Obviously, but that doesn't mean we can't define terms and call things
in the world by those terms. I have defined objects as things with
data and behavior. The car probably meets those criteria, so I call
it an object.
But you are attempting to set this "looking out of a window" as a test
for existence of objects. Simply doing that is not always the best
way to do such a test. Generally it requires more info than looking
out a window provides.
>To clarify [note: I do not want to put words into your mouth, so the
>following is a gross simplification of how I think you would phrase
>certain things; I may be wrong (I'm not a native speaker of English)
>and apologize in advance]:
> you might say "there are objects in the world", I would rephrase this
> to "there are elements in the world that can be called `objects'".
> The same holds for, e.g., polymorphism: you'd say "objects are treated
> polymorphically in the world",
I would say polymorphism exists objectively in the world via objects.
>while I'd say "the treatment of objects
> in the world is modelled by the concept `polymorphism'".
OK, yes we differ.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>You also have to be careful about direct results and indirect
>>>conclusions. For example, everyone will agree that there are
>>>apples, and if you put an apple 1m above ground and open your
>>>fingers, it will fall down; these are all direct results. This
>>>experiment does not prove that gravity exists, because that is
>>>a conclusion, based on a model. The commonly agreed model for this
>>>physicalprocess is the force of gravity with its laws, and all we
>>>can do is test whether our perception conforms to this model. If
>>>the test succeeds, then this does not say anything about the real
>>>nature of the physical process that we observed; it merely shows
>>>that your model is able to explain this aspect of the world.
>>>The history of science is full of models that turned out to be
>>>wrong or inaccurate after they had been accepted for quite a
>>>while.
>Ell <e...@access.digex.net> wrote:
>>Yes. to confirm a theory it typically requires more than one result
>>confirming the predictions of the theory.
>One cannot finally confirm a theory, one can only falsify it.
One can say that it has a probability of being true.
>If there are enough positive experiments, we put some faith
>into the theory.
That's what I mean by probability. So we *can* do more than just
falsify.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>Of course the world has not changed because we withdrew
>>>some of our models; it simply exists, and will do so independently
>>>of all our attempts to describe it properly.
>Ell <e...@access.digex.net> wrote:
>>True, the world will exist independently of our theories and
>>hypotheses, but that doesn't mean that our theories and hypotheses are
>>never correct either.
>Of course. Still, a theory about the world is not the world
>itself, so any evidence that the OO model of the world is correct
>(in the sense that it has correctly predicted the results of a
>large number of experiments) does not mean that the world is
>object-oriented.
It means there is a high probability.
>In particular, it does not exclude that there
>are other models of the world that are also correct in the same
>sense.
Yes. I said that earlier, that world seems to have both OO and non-OO
features.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>PS: All this has been discussed at length in the thread `What does
>>> "Real" mean?' in June of this year. A related thread in July/August
>>> of 1997 (Subject: Axioms etc.) shows that you use a different
>>> interpretation of `proof' than most others do (i.e. something can be
>>> proven temporarily).
>Ell <e...@access.digex.net> wrote:
>>The point I making is that sometimes the world exists objectively in a
>>certain way for a limited time period. So that once the conditions
>>change it may not even be possible to test a theory/hypothesis. But I
>>want to make clear that as long as specific conditions hold the
>>predictions of a correct theory/hypothesis should hold.
>?? This sounds like you believe that the world changes, depending
>on the correctness (in the sense described above) of our theories.
The world changes so out theories may no longer be correct.
>Above you agree that the world exists independent of our theories.
>Under what conditions may it be impossible to test a theory/hypothesis?
The world changes.
>schl...@pella.informatik.uni-rostock.de (Juergen Schlegelmilch) wrote:
>>>PPS: One has to be careful even with direct results because our
>>> senses are not perfect. There are several visual tricks that
>>> show that our eyes betray us in certain circumstances. Even
>>> our nerves are not perfect: neurologists can induce visual
>>> patterns into our conciousness using electric impulses. Taken
>>> to the extreme, one cannot even be sure that there exists
>>> anything except the own conciousness (read Hofstadter and
>>> Dennett "The Mind's I" for a very interesting discussion of
>>> all this).
>Ell <e...@access.digex.net> wrote:
>>That is why a.) we run controlled experiments, and b.) we run
>>controlled experiments on controlled experiments {i.e. we test how
>>experiments were carried out}.
>That does not help: how do you control your senses?
You determine if the senses are correct by testing the knowledge they
give you.
>Just think
>of your brain swimming in a nutrition liquid, with lots of cables
>wired to its nerves which give you the illusion of the existence
>of a world where light travels at 300.000km per second. How can
>you prove that this is simply an illusion?
>[ I'm playing devil's advocate here, but in principle the
> above scenario cannot be excluded ]
Do scientific tests.
>Seriously, as I wrote at the beginning, every experiment has two
>possible outcomes, where the negative one would falsify the theory
>being under examination. Can you describe an experiment that could
>falsify the OO model of the world?
You test for the features of OO, among others:
a.) entities with behavior and properties
b.) encapsulation
c.) polymorphism
d.) identity
e.) state
f.) commonality
and see if the features never turn up.
That would show a high probability they don't exist. Because just as
you say an experiment of this kind doesn't say for sure than things
may not be different in the future. :-}
So you can't say for sure that OO doesn't exist objectively in the
real world. Ahah! :-}
Ell
>Tim Ottinger <otti...@oma.com> wrote:
>
>> Thank you for being a kindred spirit. This is what I've been trying to
>> say, and lacking eloquence for, for years. Everything is not an
>> object,
What doesn't have behavior and properties (state) (attributes)?
Elliott
>> but you can make an object for anything.
>>
>
>Not bad eloquence with that statement.
>
>I would prefer to say: I can make an abstraction for anything - I hope.
--
Ell
>mmo...@wiretel.com wrote:
>> I sincerely believe that if we have a good understanding of any domain,
>> including interactions within a human, we can model that domain with OO.
>well you need more, since OO is largely about how encapsulated
>objects interact. you'd also need a load of stats, maths, chemistry,
>and maybe even quantum mechanics .. whether or not you bundle your
>model in OO fashion is a minor issue!
All those things are not relevant to every field, or domain. I think
the writer's assertion holds: if you *understand the domain* - with
the stuff you mention *where necessary* - then typically we can model
the domain using OO.
Elliott
Joachim Durchholz
>
> >If it does, then this proves
> >that your model conforms to the observations. This does not say
> >anything about objectivity or subjectivity.
>
> If the evidence from "looking around" comes from controlled experiment
> why isn't that objective?
1) Because different people perceive different things. So "looking
around" is subjective (done by a subject, influenced by a subject's
perception and interpretation). There have been many cases where
different people disagreed on what they saw; and there are many cases
where everbody saw the same in discarded it as measurement error (recent
example: nanobacteria, the effects of which were discarded as signal
noise so they escaped detection until very recently).
2) Because we don't know whether we are humans, naked brains fed with
artificial stimuli by some mad scientist (well, mad only by our
definition), or computer programs (consciousness and free will *might*
be an illusion).
Of course I don't believe in being a computer program, or being fooled
by artificial stimuli. My model of the world is probably quite near to
yours. But I can't be sure, and whenever I see somebody claim he knows
what objective reality is, I flinch - his reality is probably 99% the
same as mine, but the remaining 1% will erupt into regular holy wars
because we can't agree on what the world is.
An objective reality may or may not exist. All I can say for sure is
that I have a set of models that have been matching my perceptions for
some time now. I'm reasonably confident that they will work for a bit
longer, but that's just my expectation; I wouldn't have to cast away all
my beliefs if I got switched off tonight, or if the laws of quantum
physics broke down, or whatever. (Not that I would appreciate any of
these events :)
BTW you should define what you mean. In particuclar, I couldn't make out
what a "controlled experiment" should be in your opinion, and what the
significance of "controlledness" is.
> >I'd call something `objective' if everyone agrees on it;
>
> I disagree. For me what is objective is what has been proved through
> controlled experiment.
There is no such thing as a rigorous, mathematical proof when it comes
to assertions on the real worlds. There is just a set of observations,
with more-or-less good confidence levels.
> The point I making is that sometimes the world exists objectively in a
> certain way for a limited time period. So that once the conditions
> change it may not even be possible to test a theory/hypothesis. But I
> want to make clear that as long as specific conditions hold the
> predictions of a correct theory/hypothesis should hold.
How can that be, if you cannot even prove that the world is more than a
simulation?
[follow-up set to comp.object - not an Eiffel topic, if it ever were
one]
Regards,
Joachim
--
Please don't send unsolicited ads.
Rolf F. Katzenberger
<363e202c...@news.erols.com> e...@access.digex.net (Ell) wrote:
> mmo...@wiretel.com wrote:
>
> >Tim Ottinger <otti...@oma.com> wrote:
> >
> >> Thank you for being a kindred spirit. This is what I've been trying to
> >> say, and lacking eloquence for, for years. Everything is not an
> >> object,
>
> What doesn't have behavior and properties (state) (attributes)?
Everything we don't want to manipulate.
Behavior and properties are manifestations of our desire to manipulate
something, and we mistake our ability to manipulate something for
understanding it.
Regards,
Rolf
--
______________________________________________________________________
Rolf F. Katzenberger | Software Developer | Trainer 1998-09-21
Home: http://www.geocities.com/SiliconValley/Park/9557
PGP : http://wwwkeys.pgp.net:11371/pks/lookup?op=get&search=0x3B39491F
(Fingerprint F1C0 3116 F6D4 DA33 E61D D2E4 2FB8 D6B6 3B39 491F)
Alex Rice
anthropomorphize computers, they don't like it."
Since DNA does not have a subjective, it cannot be "object oriented"...
What good is half a dichotomy?
It seems that you have applied artificial evolution to your computer
programming practice. See definition 1:
genetic: 1: relating to or determined by the origin, development or
causal antecedents of something 2: of, relating to, or involving
genetics. (Webster)
That you have evolved code which is OO-like is fascinating, but not
surprising really. There is a lot of cognitive evidence that perception
and categorization is based in our physical embodiment and interaction
with the environment. From gestalt psychology to color perception, the
structure of categories, and the use of metaphor, there is a great deal
of evidence that categories are not purely objective and they are not
random either... rather they emerge out of our embodied and directly
understood experience. See Lakoff *, Rosch, Wittgenstein.
I think this is related because I believe that OO programming concepts
have emerged from the same processes for the same reasons.
Does DNA exhibit object oriented aspects? I would expect it to, in that
we, as observers, will see that in the systems we are studying. If we
see OO aspects in digital computers systems than we will see OO aspects
in molecular computer systems (cellular automatons, chemical reactions,
autocatalytic molecular systems). Stated differently, if we can describe
any one Turing complete computing system (digital computer) in OO terms
then we can describe any Turing complete computing system (molecular
computer) in OO terms. We could also describe such systems in other,
non-OO ways.
These are big, big questions. I am a complexity hack. I digest and
regurgitate other peoples work. You asked for references so here you are
:-)
On the psychology of categories I really like:
Women, Fire and Dangerous Things
What Categories Reveal about the Mind
by George Lakoff (Univ. Chicago, 1987)
On complexity studies I really like:
Frontiers of Complexity
by Peter Coveney and Roger Highfield (Fawcett Columbine 1995)
I am compiling book and website reviews on these and other complexity
topics. I welcome references or URLs on OO design, patterns, complexity
and cognitive science. Or just about any content what-so-ever. TIA!
Regards,
Alex Rice alr...@swcp.com al...@SolarAge.org
P.S. I would love to see more analogies thrown about!
Alex Rice
After I posted, I realized how poorly worded this paragraph is:
> Does DNA exhibit object oriented aspects? I would expect it to, in that
> we, as observers, will see that in the systems we are studying. If we
> see OO aspects in digital computers systems than we will see OO aspects
> in molecular computer systems (cellular automatons, chemical reactions,
> autocatalytic molecular systems). Stated differently, if we can describe
> any one Turing complete computing system (digital computer) in OO terms
> then we can describe any Turing complete computing system (molecular
> computer) in OO terms. We could also describe such systems in other,
> non-OO ways.
What I meant was that we are describing our environment, or problem
domains, or whatever you choose to call it, in OO terms. This seems to
be a natural tendency for us. Whether the underlying computational
engine is a molecular computer or digital computer should not matter.
But by asking the question "Is our DNA objected oriented", you jump into
an interesting philosophical situation.
Alex
Juergen Schlegelmilch
J Schlegelmilch wrote:
>>>>So, there is no way to determine whether any of these features
>>>>exists `objectively' in the world.
>
>>Ell <e...@access.digex.net> wrote:
>>>The way to determine what is objective is the scientific method.
>
>>There seem to be two points of disagreement:
>>- What the scientific method achieves
>>- What "objective" means
>>The scientific works like this: You have a model (a set axioms and
>>inference rules) which you want to test. You then choose one of the
>>theorems, design an experiment for it, and perform that experiment.
>>There are only two possible outcomes: either the results of the
>>experiment match the predictions of the theorem, or they do not match.
>>In the first case, you now know that your modell conforms to reality
>>_in this specific experiment_; you still do not know whether seemingly
>>similar experiments will also match the predictions of the model. In
>>the second case, your model is simply wrong; you do not know whether
>>a seemingly similar model is also wrong.
>
>>Obviously, you cannot prove that your model is _identical_ to reality
>>since this requires to perform _all_ experiments at the same time
>>(time is part of the experiment conditions).
>
>Often time has no bearing; the conditions that have a bearing are
>often the same regardless of time. This is true even for most quantum
>physics. We can often show that regardless of time the other factors
>that are truly meaningful are the same.
Who tells you that time has no bearing? It is an assumption,
possibly derived from your model. As such, you cannot be sure
beforehand that time does not matter. To test that, you'd really
have to perform tests simultaneously. Remember for example the
Heisenberg Uncertainty principle; there, time does matter alot.
Again, what you can show is that your model is independent of
time; for the world, you can only guess, based on your model
and previous experiments.
Just a primitive example: assume that you measure something
at various points in time, and the outcome is always the same,
i.e. the graph is a horizontal line. This does not show that
time does not matter, because there are infinitely many possible
other graphs that pass through the points that you measured.
Each additional measurement excludes some of those, but you
cannot exclude them all; so, you have to accept an uncertainty.
To conclude that the measurement is independent of time
requires faith in the model that you assumed.
>>So, it is important to be aware that reality
>>and any of the models that we have are not identical.
>
>One can not infer that our theories/models are essentially inaccurate
>in how they describe reality simply because they *may* not always be
>accurate. Further I know that no theory can exactly match reality
>because reality is infinite and changing. But often for what we are
>interested in (a context) our theories may be essentially correct.
>And we can say that for given conditions 'x', the result will be 'y'.
According to our model, yes. Reality may look different if our
model is wrong. We put faith in a model if it has been tested
for quite a while, but even after some hundred years, it may turn
out to be wrong. Its successor can look radically different from
the deprecated model, only the predictions will be more or less
the same since the world does not change that much (well, that is
an assumption of mine).
>>Our models can use OO features; reality is free to use others.
>
>Of course. But this doesn't mean that reality can't objectively have
>OO features, as I and others have seen proof that it does.
There is no proof that reality has OO features. All you can say
is that there is no proof to the contrary, and that you have
enough faith in your model to say that the world most probably
has OO features. Any experiment in natural sciences can only
potentially falsify a hypothesis, but not proof it, because
there may be alternative models that result in the same predictions.
Whatever experiment you set up, you have to decide beforehand
what you consider a failure, and what a success. In case the
measured results meet the criteria for failure, then you can
be sure that the hypothesis was wrong; this is a real proof.
In case of a success, you have not excluded failures in some
future experiment, which will falsify the hypothesis under
consideration.
Ell
>e...@access.digex.net wrote:
>>
>>J Schlegelmilch wrote:
>>>>>
>>>>>So, there is no way to determine whether any of these features
>>>>>exists `objectively' in the world.
>>>Ell <e...@access.digex.net> wrote:
>>>>
>>>>The way to determine what is objective is the scientific method.
Ell continues:
From the moment we turn to our own use of these objects according to
the qualities we perceive in them, we put to an infallible test the
correctness or otherwise of our sense perceptions. If these
perceptions have been wrong then our estimate of the use to which an
object can be turned must also be wrong, and our attempt must fail.
But if we succeed in accomplishing our aim, if we find that the object
does agree with our idea of it and does answer the purpose we intended
it for then that is positive proof that our perception of it, and of
its qualities, so far, agree with reality outside ourselves.
Schlegelmilch writes:
>>>Obviously, you cannot prove that your model is _identical_ to reality
>>>since this requires to perform _all_ experiments at the same time
>>>(time is part of the experiment conditions).
>>Often time has no bearing; the conditions that have a bearing are
>>often the same regardless of time. This is true even for most quantum
>>physics. We can often show that regardless of time the other factors
>>that are truly meaningful are the same.
>Who tells you that time has no bearing? It is an assumption,
>possibly derived from your model.
When we draw the conclusion that time does not matter we base that
upon scientifically summing up prior experience and experiments. We
draw the conclusion not from a model, but from real experience and
experiment. The experience and experiment has given us knowledge that
has been and continues to provide success and then it also tells us
that in some cases time doesn't matter.
>As such, you cannot be sure
>beforehand that time does not matter. To test that, you'd really
>have to perform tests simultaneously. Remember for example the
>Heisenberg Uncertainty principle; there, time does matter alot.
Other experience and experiment that provides success shows us that in
this case time *does* matter.
>Just a primitive example: assume that you measure something
>at various points in time, and the outcome is always the same,
>i.e. the graph is a horizontal line. This does not show that
>time does not matter, because there are infinitely many possible
>other graphs that pass through the points that you measured.
>Each additional measurement excludes some of those, but you
>cannot exclude them all; so, you have to accept an uncertainty.
>To conclude that the measurement is independent of time
>requires faith in the model that you assumed.
But formerly successful experience often tells us whether or not for a
given size in, or position of, the gaps, whether or not there is a
high or low chance of the measurements between them being radically
different from what we have already measured. Frequently former
experience is a good guide, but sometimes we are surprised.
>>>So, it is important to be aware that reality
>>>and any of the models that we have are not identical.
>>One can not infer that our theories/models are essentially inaccurate
>>in how they describe reality simply because they *may* not always be
>>accurate. Further I know that no theory can exactly match reality
>>because reality is infinite and changing. But often for what we are
>>interested in (a context) our theories may be essentially correct.
>>And we can say that for given conditions 'x', the result will be 'y'.
>According to our model, yes. Reality may look different if our
>model is wrong.
Well I prefer to say that reality may be somewhat different even with
mostly correct theories/models. That is because we often don't know
absolutely everything about a context and reality is infinite. My
other point in the above paragraph is that even mostly correct,
theories can become mostly incorrect due to changes, growth,
development, etc in reality.
>We put faith in a model if it has been tested
>for quite a while, but even after some hundred years, it may turn
>out to be wrong. Its successor can look radically different from
>the deprecated model, only the predictions will be more or less
>the same since the world does not change that much (well, that is
>an assumption of mine).
Not sure of your point, but predictions of the past may be changed due
to new facts becoming known since the old theory was proposed and
shown to have a high probability of being accurate. Also the new
theory may predict new things.
>>>Our models can use OO features; reality is free to use others.
>>Of course. But this doesn't mean that reality can't objectively have
>>OO features, as I and others have seen proof that it does.
>There is no proof that reality has OO features.
No proof, but given the evidence there is an extremely high chance
that reality has OO features in it. I'll lay out some over next 2, or
3 days.
>All you can say
>is that there is no proof to the contrary, and that you have
>enough faith in your model to say that the world most probably
>has OO features.
Right a very high probability.
>Any experiment in natural sciences can only
>potentially falsify a hypothesis, but not proof it,
>because there may be alternative models that
>result in the same predictions.
1) Experiment, and more so a number of related experiments, can show
the probability of a theory, or hypothesis being false, or *true*.
2) Experiment can also show that given condition 'x' then 'y' will
occur.
It is also clear that you can not definitively state that reality
does not have OO features.
>Whatever experiment you set up, you have to decide beforehand
>what you consider a failure, and what a success. In case measured
>results meet the criteria for failure, then you can
>be sure that the hypothesis was wrong; this is a real proof.
I disagree. More than anything:
1) If we are testing a theory, we want to see if given certain
conditions, if what the theory predicts for those conditions will
happen.
2) If we are testing conditions, we note what happens given
specific conditions.
>In case of a success, you have not excluded failures in some
>future experiment, which will falsify the hypothesis under
>consideration.
But what you specify as falsity before the experiment is not
necessarily how an experiment *must* fail. It may fail in ways, or
for reasons other than what you have specified before the experiment.
If you are testing like my #1 just above then you are mainly looking
for verification of what the theory predicts. Failure may happen in
any way
Richard MacDonald
>
>>Any experiment in natural sciences can only
>>potentially falsify a hypothesis, but not proof it,
>>because there may be alternative models that
>>result in the same predictions.
>
>1) Experiment, and more so a number of related experiments, can show
>the probability of a theory, or hypothesis being false, or *true*.
Now you are talking probabilities, *not* proof.
>2) Experiment can also show that given condition 'x' then 'y' will
>occur.
But not *prove* that the cause-and-effect is due to model A or some
as-yet-unknown model B.
>It is also clear that you can not definitively state that reality
>does not have OO features.
Well this came out of nowhere. Is there some cause-and-effect you are trying
to show, or are you just stating the obvious?
>>Whatever experiment you set up, you have to decide beforehand
>>what you consider a failure, and what a success. In case measured
>>results meet the criteria for failure, then you can
>>be sure that the hypothesis was wrong; this is a real proof.
>
>I disagree. More than anything:
> 1) If we are testing a theory, we want to see if given certain
>conditions, if what the theory predicts for those conditions will
>happen.
Then you have decided beforehand what you consider a failure and a success.
Given certain conditions, it is a success if the theory predicts within X,
it is a failure if not.
> 2) If we are testing conditions, we note what happens given
>specific conditions.
If all you are doing is noting what happens, you are gathering data, not
testing a theory.
>>In case of a success, you have not excluded failures in some
>>future experiment, which will falsify the hypothesis under
>>consideration.
>
>But what you specify as falsity before the experiment is not
>necessarily how an experiment *must* fail. It may fail in ways, or
>for reasons other than what you have specified before the experiment.
>If you are testing like my #1 just above then you are mainly looking
>for verification of what the theory predicts. Failure may happen in
>any way
>
Point? Yes, failure is open-ended, but that is a confirmation of the
scientific method and what Schlegelmilch was saying. I thought we were
discussing whether or not a theory can be "correct".
Tim Ottinger
>
> e...@access.digex.net (Ell) wrote:
>
> > mmo...@wiretel.com wrote:
> >
> > >Tim Ottinger <otti...@oma.com> wrote:
> > >
> > >> Thank you for being a kindred spirit. This is what I've been trying to
> > >> say, and lacking eloquence for, for years. Everything is not an
> > >> object,
> >
> > What doesn't have behavior and properties (state) (attributes)?
>
> Everything we don't want to manipulate.
Wow, it's like an argument with a ghost. I only see the replies that
are quoted by others.
But anyhow, I have a height. It's described by a positive real number,
a
measure of vertical magnitude, but the integer isn't a height. My
height
can be observed, but if you were to slice me open, you'd not see any
5'11" integer inside me. You could measure me, and see that I'm just
about 6' tall, but you cant find a part of me that is my height.
Anything can be *described* by an object.
This is a silly discussion, IMHO. The world is not a bunch of oil
paintings, but an artist can paint anything, even abstract things
like motion and emotion and imaginary things. But the world is not
a bunch of oil paintings. It's much richer and more real than
that.
Oil paintings are interpretations of real world things. There is
no requirement that a painting of me should be 6' tall and weigh
220 pounds, but it should make me look as though I do if it is
trying to be realistic. There is no requirement that it is realistic,
either. It could be impressionistic, or a cartoonly picture of me
with an enormous head driving a tiny car (you've seen those).
Not everything painted is real, but anything real and observable
may be painted.
Tim
Ell
>On Fri, 30 Oct 1998 20:24:57 GMT, in article
><363e202c...@news.erols.com> e...@access.digex.net (Ell) wrote:
>
>> mmo...@wiretel.com wrote:
>>
>> >Tim Ottinger <otti...@oma.com> wrote:
>> >
>> >> Thank you for being a kindred spirit. This is what I've been trying to
>> >> say, and lacking eloquence for, for years. Everything is not an
>> >> object,
>> What doesn't have behavior and properties (state) (attributes)?
>Everything we don't want to manipulate.
Holy cow! This the same as saying if a tree falls in forest and no
one is around then it doesn't really fall. That idea is a really
untrue and potentially very harmful.
>Behavior and properties are manifestations of our desire to manipulate
>something,
To think that an object's behavior and properties don't exist
objectively, just when we desire to manipulate the object can get us
into a lot of trouble. One might feel because we do not desire to
manipulate object A we can do things in another part of the program
that would be harmful to object A if we were desiring to manipulate
it. But in fact we shouldn't do this in the other part of the program
because the behavior and properties of object A do exist whether we
desire to manipulate it or not.
>and we mistake our ability to manipulate something for
>understanding it.
The fact that we can manipulate something without understanding it is
widely known. You can do a lot to mess up something without really
understanding it. People do that all the time. One example is the
current movement afoot to mess with comp.object discussion.
But what does the fact we can manipulate an object without
understanding it have to do with your assertion that an object's
behavior and properties don't exist until you desire to manipulate it?
Don Harrison
:> So the question is with our genome, since it is the oldest
:> program known, is it possible it is objected oriented in design?
:
:Sorry, that is a non-question. Our genetic code is not designed.
:Nor is it a program. That's just a shorthand analogy useful in
:certain cases.
Our genetic code has a purpose (that we have a particular form etc.).
Purpose implies intent.
Intent implies design.
Therefore, our genetic code was designed.
QED.
:> Or perhaps our genome structure is far more
:> advanced than OO.
:
:Probably .. it's been "brewing" for four billion years.
Actually, I expect God (YAHWEH) conceived it even before he created time.
So, in that sense, it took no time at all. :)
:> Anybody know of any books on this subject?
:
:Check out the evolutionary biology popularisation texts by Stephen J
:Gould.
:Then read some Richard Dawkin (The Selfish Gene) to ensure your
:remaining
:optimism is completely stamped out.
Surprisingly, Dawkins, an avowed atheist, acknowledges the existence of
purpose in the universe (see Climbing Mount Improbable) but has evidently
not twigged to the fact that purpose implies a creator.
Some interesting quotes from "Not a Chance" by R.C. Sproul (Baker Books)
related to Darwinism, of which Dawkins is a champion:
"The phrase 'random selection' (or 'random mutation') is also somewhat
confusing. To make a 'selection' suggests some form of *intentionality*,
a trait usually associated with intelligence. Is it possible to have
unintentional intention?
...
[Carl] Sagan goes on to speak of 'fate', 'accident' and 'chance' as
factors involved in natural selection. These words seem to militate
against the idea of actual *selection". If selection involves
intentionality, then the concepts collide. We cannot have an intentional
accident or an accidental intention."
Don (Harrison). donh at syd.csa.com.au
Eiffel - Software engineering with class
Juergen Schlegelmilch
>>I think that the OO features that you see are part of your model of
>>the world, while you think that they are part of the world itself.
>
>I see OO as both. It is exists in the world independently of the
>mind, and it is a modelling paradigm--a way of seeing world.
IMHO (really, it's only my personal opinion), OO is only
a modelling paradigm. However, this does not limit its
usefulness in any way, especially not with regard to
software engineering where we have to construct models anyway;
whether these are models of the world, or simplified versions
of the model of the world cannot be determined afterwards.
>>I hope to have shown in the discussion of the scientific method,
>>that one cannot prove that one's model is identical to the world
>
>I think one can show for conditions 'x' then 'y' will occur and that a
>theory has 'z' probability of being true.
The important point here is `probability'. One cannot show that
for conditions 'x' some 'y' will occur, not even with a probability,
because that probability requires a model that allows to measure
the relative quantity of success and failure (remember, the
probability of something is the quotient of these numbers). Now,
if you assume that the implication `x -> y' has a certain
probability, then you already assume that your theory is wrong
on some occasions, so you can withdraw it anyway.
[ I note that below (*) you write that by `probability' you do
not mean the technical term, but that one can put faith in a
theory, i.e. believe in it. However, faith cannot be
measured by some `z'.
]
So, technically, there is no chance to overcome the fallibility
of natural sciences; nature may reveal new twists at any time.
Intuitively, all of us do put faith in the models used in
natural sciences because they have been useful and reliable in
the past.
>If "looking around" is an experiment as you just said then it isn't
>simply looking out the window, right? All experiments should be
>controlled. Or at least the statistical analysis of "looking around"
>should have controls applied within it. You are saying look out the
>window. at what? when? by what means?
Sorry, I came up with the "looking around" because you often use the
phrase that you "see objects in the real world" (e.g. in messages
<35702116...@news.erols.com>,<6h0sro$ejt$1...@newsreader.digex.net>).
Now I see that you did not mean that literally; sorry about the
misunderstanding.
Now that you mention it, what controlled experiments did you perform
to make sure the world has object-oriented features? What did you
measure, and how did you do it? Where was the boundary between success
and failure in that experiments?
>>Ell <e...@access.digex.net> wrote:
>>>If the evidence from "looking around" comes from controlled experiment
>>>why isn't that objective?
>
>>Because you are subjective: you are the one interpreting the results.
>
>But if you perform experiments scientifically then the results are not
>subjective.
Then, who interprets them? Without interpretation, experiments
just result in numbers, or images. It requires interpretation
to regard them as either failure or success according to a
hypothesis and chosen criteria, and interpretation is subjective.
There can be more than one model which allows a consistent
interpretation.
>>The world simply _is_, it does not interpret; humans do, based on
>>models.
>
>Obviously, but that doesn't mean we can't define terms and call things
>in the world by those terms. I have defined objects as things with
>data and behavior. The car probably meets those criteria, so I call
>it an object.
Yes, of course. You are free to define terms, but those are
part of your model, not of the world.
>But you are attempting to set this "looking out of a window" as a test
>for existence of objects. Simply doing that is not always the best
>way to do such a test. Generally it requires more info than looking
>out a window provides.
I completely agree. The "looking around" was taken from some of
your postings (see above for message IDs), and I mistakenly took
this phrase literally.
(*) Here is your comment on `probability':
>>One cannot finally confirm a theory, one can only falsify it.
>
>One can say that it has a probability of being true.
>
>>If there are enough positive experiments, we put some faith
>>into the theory.
>
>That's what I mean by probability. So we *can* do more than just
>falsify.
Yes, one can believe in a theory.
>>In particular, it does not exclude that there
>>are other models of the world that are also correct in the same
>>sense.
>
>Yes. I said that earlier, that world seems to have both OO and non-OO
>features.
I agree that the world has features that are best modelled in
an OO way, and other features that are not so easily modelled
with the OO model. It all depends on the application that we
have in mind.
>>?? This sounds like you believe that the world changes, depending
>>on the correctness (in the sense described above) of our theories.
>
>The world changes so our theories may no longer be correct.
I prefer to say that the world does not change just because
a human model of it does not match it; this view is less
anthropocentric, but nevertheless philosophical. Both are
valid.
>>Seriously, as I wrote at the beginning, every experiment has two
>>possible outcomes, where the negative one would falsify the theory
>>being under examination. Can you describe an experiment that could
>>falsify the OO model of the world?
>
>You test for the features of OO, among others:
> a.) entities with behavior and properties
> b.) encapsulation
> c.) polymorphism
> d.) identity
> e.) state
> f.) commonality
>and see if the features never turn up.
>
>That would show a high probability they don't exist. Because just as
>you say an experiment of this kind doesn't say for sure than things
>may not be different in the future. :-}
Well, if the world changes under our hands we have to drop
any science because it may just have changed after we performed
an experiment :-)))
Concerning the test, it is simpler: if a single experiment
yields results that are not interpretable in the OO model,
then this would falsify the theory (that the world is OO);
this would still not make it useless.
However, we would need some kind of yardstick to set up the
experiment, so that we can say beforehand what criteria
something must fulfill to count as an entity with behaviour
and properties, or what it means to have encapsulation. All
these terms are used colloquially and technically, and
unfortunately their technical meaning is not unambiguous (it
often depends on the programming language).
>So you can't say for sure that OO doesn't exist objectively in the
>real world. Ahah! :-}
I agree completely. Just as you can't say for sure that OO
does exist objectively in the real world. It is all a matter
of interpretation, guided by models.
I'll try a summary:
- We agree that experiments cannot proof that the real world
is object-oriented, ie that the OO model and the world are
identical. A large number of successful experiments is a
good base to put faith into the model. We differ in the
amount of faith, though.
- There are other models which model the world, and what
model one uses depends on the context.
- We differ in that you think the world changes according to
the models one adopts, while I think that the world exists
independently of the models.
Thanks for the discussion,
Jürgen Schlegelmilch
Darren New
> Our genetic code has a purpose (that we have a particular form etc.).
> Purpose implies intent.
I would say that if purpose implies intent, then our genetic code has no
purpose. Or do you think that molecules can become concious just because
they're made out of genetic material?
> Intent implies design.
Again, wrong. Unless you want to say that the design is to mix up a
bunch of random stuff to see what happens?
>
> Some interesting quotes from "Not a Chance" by R.C. Sproul (Baker Books)
> related to Darwinism, of which Dawkins is a champion:
>
> "The phrase 'random selection' (or 'random mutation') is also somewhat
> confusing. To make a 'selection' suggests some form of *intentionality*,
Yes. Making a selection implies intentionality. But selection isn't
what's implying the intentionality. It's the "making" part. These are
word games we're both playing; such word games say more about your mind
than about the real world or what's beyond it.
If I say "You have a selection of three colors", where's the intent in
that?
> [Carl] Sagan goes on to speak of 'fate', 'accident' and 'chance' as
> factors involved in natural selection. These words seem to militate
> against the idea of actual *selection". If selection involves
> intentionality, then the concepts collide. We cannot have an intentional
> accident or an accidental intention."
Sure can! I reach into an opaque bag with red and blue marbles in it,
mixed up. I select a red marble from the bag. Does that mean I intended
to select a red marble? Not at all. I intended to select a marble.
Whether it's red or not was not intentional, even tho red was selected.
Hence, selection without intention.
To say we can't have an intentional accident is easy to disprove with
silly word games, too. If you annoy me on the road and I dent your car
on purpose, then we've had an intentional accident.
Feel free to respond, but I've wasted many hours arguing silly word
games with folks who can't see they're just silly word games and think
it reveals something about a real and objective God, so I probably won't
answer unless you come up with something better, to which I look forward
if you do. :-)
Just for the record, no offense intended, I don't really believe in
evolution either (as I feel it's a non-disprovable theory), I really am
looking forward to hearing something other than wordgames, and I'd never
post something like this publicly except that Don doesn't give out his
private email address. Follow-ups directed directly to me.
Graham Perkins
> Oil paintings are interpretations of real world things...
> There is no requirement that it is realistic,
> either. It could be impressionistic, or a cartoonly picture of me
> with an enormous head driving a tiny car (you've seen those).
No I haven't.
Where are the pictures of Tim with an enormous
head driving a tiny car? I think we should be told.
Robert C. Martin
Graham Perkins wrote in message <363EF5...@dmu.ac.uk>...
>Tim Ottinger wrote:
>> Oil paintings are interpretations of real world things...
>> There is no requirement that it is realistic,
>> either. It could be impressionistic, or a cartoonly picture of me
>> with an enormous head driving a tiny car (you've seen those).
>
>No I haven't.
>
>Where are the pictures of Tim with an enormous
>head driving a tiny car? I think we should be told.
Tim... Are you really Peter Gabriel???
Robert C. Martin | Design Consulting | Training courses offered:
Object Mentor | rma...@oma.com | Object Oriented Design
14619 N Somerset Cr | Tel: (800) 338-6716 | C++
Green Oaks IL 60048 | Fax: (847) 918-1023 | http://www.oma.com
Tim Ottinger
>
> Tim Ottinger wrote:
> > Oil paintings are interpretations of real world things...
> > There is no requirement that it is realistic,
> > either. It could be impressionistic, or a cartoonly picture of me
> > with an enormous head driving a tiny car (you've seen those).
>
> No I haven't.
>
> Where are the pictures of Tim with an enormous
> head driving a tiny car? I think we should be told.
:-)
No pictures of me have ever been publically posted, nor
published (except for maybe yearbooks and church calendars).
--
Tim Ottinger Object Mentor OO Training and
otti...@oma.com www.oma.com Mentoring
-------------------------------------------------------
We can interpret a bad temper as a sign of inferiority.
-- Alfred Adler
Rolf F. Katzenberger
<363f9264...@news.erols.com> e...@access.digex.net (Ell) wrote:
> rf...@ibm.net (Rolf F. Katzenberger) wrote:
>
> >On Fri, 30 Oct 1998 20:24:57 GMT, in article
> ><363e202c...@news.erols.com> e...@access.digex.net (Ell) wrote:
> >
> >> mmo...@wiretel.com wrote:
> >>
> >> >Tim Ottinger <otti...@oma.com> wrote:
> >> >
> >> >> Thank you for being a kindred spirit. This is what I've been trying to
> >> >> say, and lacking eloquence for, for years. Everything is not an
> >> >> object,
>
> >> What doesn't have behavior and properties (state) (attributes)?
>
> >Everything we don't want to manipulate.
>
> Holy cow! This the same as saying if a tree falls in forest and no
> one is around then it doesn't really fall.
Not at all. Even if I were around, I could not *prove* whether there
is an objectively real tree at all or whether me and the tree are a
pet dog's dream.
I'm not quite sure whether I'd model fall() as a method of class Tree,
but even if I did, amoeba might have a fundamentally different concept
of a tree, not caring at all about height or falling. In fact, it
might not even know a class Tree. That is why I think that attributes
and methods are in the eye of the beholder, and not features
"inherent" to "real things".
> That idea is a really
> untrue and potentially very harmful.
>
> >Behavior and properties are manifestations of our desire to manipulate
> >something,
>
> To think that an object's behavior and properties don't exist
> objectively, just when we desire to manipulate the object can get us
> into a lot of trouble. One might feel because we do not desire to
> manipulate object A we can do things in another part of the program
> that would be harmful to object A if we were desiring to manipulate
> it. But in fact we shouldn't do this in the other part of the program
> because the behavior and properties of object A do exist whether we
> desire to manipulate it or not.
My statement dealt with "everything" as defined by the other posters,
i.e. with "real things". I did not say that e.g. C++ classes don't
have attributes, I rather said that things like C++ classes are
manifestations of our desire to manipulate things, i.e. that providing
such classes with attributes and methods serves a certain purpose.
Obviously, programs as collections of classes serve certain purposes,
too, so it is not so clever to simply forget about that purpose while
programming classes. It is, however, not harmful to be aware that our
designs reflect our curent interests and that it might change
(including attributes and methods) as soon as our interests change.
Such change does not imply, however, that the "real things" change,
too - in fact, it does not even imply that there were such things and
that we "understand" them much better after the change.
> >and we mistake our ability to manipulate something for
> >understanding it.
>
> The fact that we can manipulate something without understanding it is
> widely known. You can do a lot to mess up something without really
> understanding it. People do that all the time. One example is the
> current movement afoot to mess with comp.object discussion.
Since comp.object has been created because a relative majority voted
for a certain concept, it should be clear that e.g.
comp.object.moderated can be created the very same way - the existence
and further development of newsgroups results from votes. One can both
mess with votes and with the democratice principles as laid down in
the standard Usenet procedures for the creation and development of
newsgroups like c.o. and c.o.m. I haven't seen anybody mess with them
so far, however.
> But what does the fact we can manipulate an object without
> understanding it have to do with your assertion that an object's
> behavior and properties don't exist until you desire to manipulate it?
IMHO the acts of analysis and design result in models which reflect
our current interests in something. We don't model things we are not
interested in, it is impossible to think about attributes and methods
of something without having any interest in that something.
From time to time (sometimes even instantly), the models we create are
invalidated by counter-examples, although so far they had been useful
in manipulating the "real world". This implies that what we had
formerly modeled as objects having certain attributes and behaviors
might suddenly change: maybe we drop some attributes and methods and
introduce some others instead; maybe we drop a concept (class)
completely and introduce others instead.
A model that works is obviously no proof that its attributes and
behaviors really match any "objective reality" The existence of a
model that works demonstrates our interests in the work. No interest -
no model, no classes, no attributes, no methods.
Rolf F. Katzenberger
Tim Ottinger <otti...@oma.com> wrote:
> Rolf F. Katzenberger wrote:
> >
> > e...@access.digex.net (Ell) wrote:
> >
> > > mmo...@wiretel.com wrote:
> > >
> > > >Tim Ottinger <otti...@oma.com> wrote:
> > > >
> > > >> Thank you for being a kindred spirit. This is what I've been trying to
> > > >> say, and lacking eloquence for, for years. Everything is not an
> > > >> object,
> > >
> > > What doesn't have behavior and properties (state) (attributes)?
> >
> > Everything we don't want to manipulate.
>
> Wow, it's like an argument with a ghost. I only see the replies that
> are quoted by others.
>
> But anyhow, I have a height.
As soon as somebody (including you) is interested in it...
> It's described by a positive real number,
> a
> measure of vertical magnitude, but the integer isn't a height. My
> height
> can be observed, but if you were to slice me open, you'd not see any
> 5'11" integer inside me. You could measure me, and see that I'm just
> about 6' tall, but you cant find a part of me that is my height.
Measuring it changes it. It even changes over the nanoseconds. Your
"height" as noted in your passport is just an attribute needed for
certain purposes. Your kids care for other purposes, that is why you
are 5'11" tall in your passport, simply "big" for your kids and
infinitely tall according to fractal math. It all depends on who needs
your "height" for what. For an influenza virus, your height does not
matter at all, although your body as a whole matters, especially the
cells. One could even say that for a virus, you are not an object of
class HumanBeing at all.
> Anything can be *described* by an object.
>
> This is a silly discussion, IMHO. The world is not a bunch of oil
> paintings, but an artist can paint anything, even abstract things
> like motion and emotion and imaginary things. But the world is not
> a bunch of oil paintings. It's much richer and more real than
> that.
>
> Oil paintings are interpretations of real world things. There is
> no requirement that a painting of me should be 6' tall and weigh
> 220 pounds, but it should make me look as though I do if it is
> trying to be realistic.
Not at all. The paintings of you made by your children are very
realistic, too. I doubt they look as if you were 6' tall and weighed
220 pounds, but they surely show the attributes and behaviors your
children love ;-)
> There is no requirement that it is realistic,
> either. It could be impressionistic, or a cartoonly picture of me
> with an enormous head driving a tiny car (you've seen those).
>
> Not everything painted is real, but anything real and observable
> may be painted.
The painting will reflect the way the painter sees the painted,
showing the traits of it as regarded important by the painter.
Therefore, the painting can be understood by others having similar
preferences or at least understanding those preferences.
Richard MacDonald
>
>No pictures of me have ever been publically posted, nor
>published (except for maybe yearbooks and church calendars).
>
Nice legs :-)
Told you not to have that 5th tequila shot on Bourbon St last Mardi Gras.
Tim Ottinger
D'OH!
Matt Kennel
:Graham Perkins wrote:
:
::> So the question is with our genome, since it is the oldest
::> program known, is it possible it is objected oriented in design?
::
::Sorry, that is a non-question. Our genetic code is not designed.
::Nor is it a program. That's just a shorthand analogy useful in
::certain cases.
:Our genetic code has a purpose (that we have a particular form etc.).
:Purpose implies intent.
:Intent implies design.
:Therefore, our genetic code was designed.
:QED.
Are you serious or poking fun at some people?
If you're being serious, then I'll dispute the first one; the
"purpose" is something that human cognition has attached to the
genetic code, and thus any "intent" we abscribe to the purpose,
another artifact of human thought.
Does the moon have a 'purpose' in creating tides? Does its mass
distribution have some "purpose" and thus "intent" and thus "design"?
It would be hard to argue so, yet it's no different from the genetic
code except in complexity.
It's just there, and does its stuff.
This is an important issue in this group; there are two broad beliefs:
*) object-oriented features "really" are in the "real world"
*) object-orientation is a useful approximation, but wholly constructed
by human thought.
I'm 100% in the second category. How we model something in order to
better answer issues we deem useful to us, governs what we see to be
the "design" or, if we are more bold, the "purpose".
My opinion: Human conciousness is intrinsically and irreducibly
"craftite" and "pragmatic".
--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD
*
* "To chill, or to pop a cap in my dome, whoomp! there it is."
* Hamlet, Fresh Prince of Denmark.
Luis Ibanez
I would like to reinforce the sound
question that has been touched by this
message.
DNA is absolutely an Information Processing problem.
DNA has an alphabet, a syntaxis and a semantic.
DNA is a code which should be interpreted to guide
not only the development but also the continuous
work of every living being.
We have 3.000 Gigabytes of DNA code running inside.
It specifies what kinds of enzimes, proteines to produce
in each cell at every moment.
Any cell behaves like a "state machine" which reads
the DNA code stored in its nucleous, copy segments
of code in RNA chains, and produces proteins with them.
If you read any basic book of genetics or embryology
with "informatic" eyes, you will find inmediately that
the study of DNA concerns not only molecular biology
but also informatics and information processing fields.
At every cell division (reproduction), cells are able
to copy these 3000 Gigabytes in less than one second with
an error rate that any hard disk fabricant would like to
emulate.
The great problem of genetic engineers, embryologist
and paleontologist who are now strongly interested in
DNA comprehnsion is that they dispose of the "low level code"
of DNA.
The way they work is as arcane as trying to understand
how UNIX works, by reading the binary code of the kernel,
without any reference to systems architecture, design,
programming languages, devices, drivers.... nothing !
What they can do, by now, is to take the genetic
equivalent of a debuger (like gdb or the old
DOS/debug..) change a byte, and "rerun" the program
to see what changes. If the new embryo loss his
eyes, they say:
"Hey, this byte has something to do with embryo eyes...!!"
so the "label" this byte as "eye related" and pass to next byte.
--------------------
As an example:
The equivalent of a "Hello World" for genetic begginers
is an excercise over the drosophyla DNA (the fruit fly)
(http://www.fruitfly.org). It is know that the same gene
codes the morphology of the fly legs and antenae. It is
possible to "switch" this gene to make grow legs instead
of antenae in the fly's head. (http://flybase.bio.indiana.edu)
If you want to see the equivalent of a "man page" for
this gene, the "antennapedia" gene, visit:
http://flybase.bio.indiana.edu/.bin/fbidq.html?FBgn0000095
--------------------------------------------
DNA understanding is a reverse engineering problem !!
Don't trust me, take a look at:
Berkely human genome project
http://www-hgc.lbl.gov/
http://www-gsd.lbl.gov/
Cancer genome project
http://www.ncbi.nlm.nih.gov/ncicgap/
Human genome sequencing (Japan)
http://www-alis.tokyo.jst.go.jp/HGS/top.html
for a clear example see how a gene looks like:
http://www-alis.tokyo.jst.go.jp/HGS/chrom_3/all.html
If you want to see how a DNA debugger looks like:
http://www.ncbi.nlm.nih.gov/science96/
(...may be it is more like a browser...)
Visit the "tutorial of DNA code" of the Departement of Energy !!!
http://www.ornl.gov/TechResources/Human_Genome/home.html
The National Human Genome Research Institute
http://www.nhgri.nih.gov/
-----------------------------------------------------
Why the same gene control the formation of a leg and an antena ?
Simply because it is REUSING CODE !
morphology of leg and antena have been factorized somewhere,
and an "if" instruction selects from one or the two possible
development subprograms. The factorization is possible because
both structures are articulated by segments, so the morphological
process is very similar. (perhaps is only a question of parameters
passed to the right subroutine)
DNA (and all its decoding process) shows object oriented
behaviours like: inheritance, polymorphism and encapsulation.
(the derivation of the class "fruitfly" can be seen in :
http://www.exploratorium.edu/exhibits/mutant_flies/mutant_flies.html)
(the last two in this page, illustrate "eyeless" fly, and
"leg-headed" fly)
--------------------------------------------------
Personally I think that DNA is like that this,
simply because this kind of coding enhace the
robustness of code, while increasing its compresion
and performance at the same time.
Information processing comunity and Informatics comunity
have important contributions to do in DNA understanding.
....If they take it seriously.....
(Any way... MOTOROLA is building DNA chips...)
(http://www.spectrum.ieee.org)
________________________________________________________________
Luis Eduardo IBANEZ
PhD Student
Departement Image et Traitement de l'Information
Ecole Nationale Superieure des Telecommunications de Bretagne
Technopole de Brest-Iroise BP 832, 29285 BREST - FRANCE
Tel (33)2.98.00.13.57 Fax (33)2.98.00.10.98
e-mail: luis....@ieee.org
home page: http://www-iti.enst-bretagne.fr/~ibanez
________________________________________________________________
"no esperes que un hombre muera
para saber que todo corre peligro,
ni a que te cuenten los libros
lo que esta pasando ahi fuera
no esperes a que te den los planos
ni a que te proporcionen un manual
no esperes el consentimiento
ni a que el cielo te mande una senal..."
Joan Manuel Serrat
Jeff Younker
Luis Ibanez <luis...@club-internet.fr> writes:
> DNA is absolutely an Information Processing problem.
The information processing aspects of DNA are interesting,
but they are not the only aspects.
> DNA has an alphabet, a syntaxis and a semantic.
In some regions this is true. In others it is not.
The product of some sections of your genomic DNA
is (according to current evidence) catalytic RNA.
These products are never translated into proteins.
One such kind of RNA is ribosomal RNA.
There are lengths of DNA that provide binding sites
for repressor proteins. These binding sites don't
necessarily code for any products.
There is evidence that large stretches of DNA are
there strictly for it's structural properties.
> DNA is a code which should be interpreted to guide
> not only the development but also the continuous
> work of every living being.
But you can't isolate DNA from it's environment (a cell).
Simply having the DNA won't allow you to reconstitute
the organism. Proof by example: Some organisms have
differences in the codes. There are a few critters
whose DNA codes for a 21st amino acid called seleno
cysteine. Drop that DNA into our cells and it won't
produce proteins containing seleno-cysteine.
> We have 3.000 Gigabytes of DNA code running inside.
>
> It specifies what kinds of enzimes, proteines to produce
> in each cell at every moment.
Don't know how you arrived at that figure. I get a ball
park figure of 7.5*10^8 bytes per genome. That number
is based upon
2.0 bits/base pair * 3.0*10^9 base pair/human genome
---------------------------------------------------
8.0 bits per byte
> Any cell behaves like a "state machine" which reads
> the DNA code stored in its nucleous, copy segments
> of code in RNA chains, and produces proteins with them.
Except that the state machine functioning is soley
dependent upon random events. The apparent state-machine
behavior is observed average behavior of the components.
> If you read any basic book of genetics or embryology
> with "informatic" eyes, you will find inmediately that
> the study of DNA concerns not only molecular biology
> but also informatics and information processing fields.
And that is widely acknowledged by the folks who work
with DNA. The problem is that the informatics aspects
don't tell you much about the functional aspects that
many molecular biologists/biologist/biochemists/etc.
are concerned with.
The one big thing to come out of informatics is a method
for identifying coding versus non coding lengths of DNA.
(Non coding lengths are *less* random than coding lengths)
> At every cell division (reproduction), cells are able
> to copy these 3000 Gigabytes in less than one second with
> an error rate that any hard disk fabricant would like to
> emulate.
You're wrong on both counts. Replication is performed
before cell division, and it takes quite a while to do.
The fidelity of copying is also impressive, but any
modern disk copy will beat it.
The fastest replication rates (per replication fork)
are among Bacteria. They can replicate DNA at 50000
pase pairs per *minute*, but they only have one for.
Organisms such as Mice operate much more slowly
(at each fork), but even with all 25000 forks going
at the same time (not a realistic occurance) they
can only replicate 5.5*10^7 base pairs per *minute*.
(Lewin, "Genes V", p536)
After proof reading systems have acted the error rate
has been measured to be around one per 5x10^9 base pairs
copied. (Lewin, "Genes V", p576). I copy large
databases around all the time and I get better
fidelity.
> The great problem of genetic engineers, embryologist
> and paleontologist who are now strongly interested in
> DNA comprehnsion is that they dispose of the "low level code"
> of DNA.
Actually they don't dispose of it, it's that the DNA
doesn't tell you anything *useful* about it's products.
From DNA you can obtain the primary structure(s) of a
the gene product, but what you need to know is the
the three dimensional structure of the product.
Getting from the primary structure to the three
dimensional (tertiary structure) is known as the
"Protein Folding Problem". Solving the "Protein
Folding Problem" is one of the holy grails of
computational chemistry. If you can do it then
you have just won a nobel prize, and your name
will be found in every biology and biochemistry
textbook for the next several hundred years.
Even if you can do this they you have the problem
of how all the other machinery interacts with your
bit. Knowing pro-insulins three primary sequence
and three dimensional structure of won't tell you
that the translated product is chopped into peices
by another protein.
> ...
> The way they work is as arcane as trying to understand
> how UNIX works, by reading the binary code of the kernel,
> without any reference to systems architecture, design,
> programming languages, devices, drivers.... nothing !
> ...
In the above sections you complained that researchers
were throwing away the code, and now you're complaining
that they shouldn't be examining it? I fail to
see the point that you're now trying to make.
As you state further on, figuring out how a living
thing works is a reverse engineering problem, and
most biologists/biochemists/molecular biologists/etc.
that I've met understand this.
The problem is that living things were evolved
by a process that has respect for neither the cognitive
limitations nor the technological limitiations of
humans. If there is a physical law or phenomenon
that can be used in some warped way to accomplish a
biological process then in some pathway in some organism
it's going to be in use.
> ...
> What they can do, by now, is to take the genetic
> equivalent of a debuger (like gdb or the old
> DOS/debug..) change a byte, and "rerun" the program
> to see what changes. If the new embryo loss his
> eyes, they say:
> ...
A debugger is a very fine instrument. You're doing
it injustice comparing it with the methods used to
create mutants. The methods of biologists are more
akin to passing an electromagmet or a powerful xray
source over a field full of computer systems and then
examining all of the computer systems for errant
behaviors.
> ..
> DNA understanding is a reverse engineering problem !!
> ...
That it is, but we don't even understand the principles
of assembly yet. In only a handful of cases have we even
figured out how the machinery (enzymes) actually work.
> ..
> for a clear example see how a gene looks like:
> http://www-alis.tokyo.jst.go.jp/HGS/chrom_3/all.html
> ...
Keep in mind that this is a map. The real thing
is only grossly shaped like this only when it is
being packaged and separated during cell division.
The rest of the time it is packed in ways that are
not visible to imaging equipment, and not yet
understood by human beings.
> ...
> If you want to see how a DNA debugger looks like:
> http://www.ncbi.nlm.nih.gov/science96/
> (...may be it is more like a browser...)
> ...
Actually more or less the same thing as above. This
is (like the other) a compilation of the results
from other researchers. The real tools are things
like gels and the absorbance graphs from sequencing
machines. They are much messier and look nothing
like what is shown here.
> Visit the "tutorial of DNA code" of the Departement of Energy !!!
> http://www.ornl.gov/TechResources/Human_Genome/home.html
A nice basic overview, but it doesn't touch on the complications.
> The National Human Genome Research Institute
> http://www.nhgri.nih.gov/
Not sure what you were trying to point out here
>
>
> -----------------------------------------------------
>
>
> Why the same gene control the formation of a leg and an antena ?
>
> Simply because it is REUSING CODE !
Because the only realy way to generate new functionality in an
organism is to modify existing processes.
> morphology of leg and antena have been factorized somewhere,
> and an "if" instruction selects from one or the two possible
> development subprograms. The factorization is possible because
> both structures are articulated by segments, so the morphological
> process is very similar. (perhaps is only a question of parameters
> passed to the right subroutine)
You're ignoring other possibilies. In fact there reason
that one can be turned off one place and turned on another
may be a reflection of the 'if' mechanism and not an underlaying
similarity in gross physical structure. The fact is that
all structures are developed by
> DNA (and all its decoding process) shows object oriented
> behaviours like: inheritance, polymorphism and encapsulation.
It also shows non-oriented behaviors like high coupling
between code modules, multiple implementation from single
representations, self modifying code, intermixing of code
and data, high coupling between representation and
execution environment, high coupling between code and
data, etc..
> ...
> Personally I think that DNA is like that this,
> simply because this kind of coding enhace the
> robustness of code, while increasing its compresion
> and performance at the same time.
I don't think that "comprehensability" was one one
of the design criteria. Once you get down to the
nitty gritty it's a mess. DNA and the things that
it codes for are like they are simply because it
was what worked with the materials that were at
hand at a given instant in history.
> ...
> Information processing comunity and Informatics comunity
> have important contributions to do in DNA understanding.
>
> ....If they take it seriously.....
Most geneticists that I have known takes the informatics
aspects seriously. I seem to get the impression that
you're making a divide between geneticists/biochemists/
moleculary biologists and information theory specialists
that just isn't there in the world of genetics.
Jeff Younker
> In any case, DNA has no source code, which makes
> object-oriented metaphors very hard to apply.
Well put.
- Jeff Younker - je...@mdli.com - These are my opinions, not MDL's -
Jason Stokes
<je...@snipe.mdli.com> wrote:
>> DNA (and all its decoding process) shows object oriented
>> behaviours like: inheritance, polymorphism and encapsulation.
>
>It also shows non-oriented behaviors like high coupling
>between code modules, multiple implementation from single
>representations, self modifying code, intermixing of code
>and data, high coupling between representation and
>execution environment, high coupling between code and
>data, etc..
The X chromosome might be considered a gigantic module of global
variables.
In any case, DNA has no source code, which makes
object-oriented metaphors very hard to apply.
--
Jason Stokes: js...@bluedog.apana.org.au
Don Harrison
:Don Harrison wrote:
:> Our genetic code has a purpose (that we have a particular form etc.).
:> Purpose implies intent.
:
:I would say that if purpose implies intent, then our genetic code has no
:purpose.
I was asked recently to estimate the amount of effort to develop some
software using two different strategies. I happened to favour a particular
strategy and was tempted to reverse engineer the analysis so that it would
inevitably lead to the conclusion I favoured.
If we start with the conclusion thar "God does not exist", and reverse
engineer our analysis of the universe so that we inevitably reach that
conclusion - for example, by convincing ourselves that "nothing in the
'natural' world is designed" - then we are deluding ourselves.
As another example, suppose someone drew a diagram of a propulsion system
having the following characteristics:
- A long corkscrew-like propeller is attached via a universal
joint to a drive shaft
- The drive shaft rotates freely within a set of bushes mounted
in a casing
- The other end of the drive shaft has a rotor moving within a
stator, driven by a chemical process.
They then asked a bunch of engineers to comment on it. The group might make
various comments and enquire what it was intended to propel and who designed
it. None would even contemplate the possibility that the system was not
designed.
However, the system described is that of a bacterial flagellum (see P. 70 of
"Darwin's Black Box" by Michael J. Behe, Touchstone).
The fact that this is a biological propulsion system doesn't alter the
obvious fact that it was designed, but prior knowledge of that would bias
those who pretend that God doesn't exist.
:> Intent implies design.
:
:Again, wrong. Unless you want to say that the design is to mix up a
:bunch of random stuff to see what happens?
Well, no, actually. Although, surprisingly, that's basically what Darwinists
expect you to believe - order emanating from disorder! Even the worst designed
software I've seen shows an inherent order! (even if it's hard to find).
Some, apparently, would have us believe that order is all in the mind.
Personally, I have trouble getting enough order *into* my mind, rather than
having to get it *out* of my mind! :)
:> Some interesting quotes from "Not a Chance" by R.C. Sproul (Baker Books)
:> related to Darwinism, of which Dawkins is a champion:
:>
:> "The phrase 'random selection' (or 'random mutation') is also somewhat
:> confusing. To make a 'selection' suggests some form of *intentionality*,
:
:Yes. Making a selection implies intentionality. But selection isn't
:what's implying the intentionality. It's the "making" part.
"making a selection" means the same as "selecting" - which is a single action.
You're attempting to make a distinction which doesn't exist.
:If I say "You have a selection of three colors", where's the intent in
:that?
The colours selected were intended to be included in the selection and the
ones that weren't selected were deliberately excluded.
:> [Carl] Sagan goes on to speak of 'fate', 'accident' and 'chance' as
:> factors involved in natural selection. These words seem to militate
:> against the idea of actual *selection". If selection involves
:> intentionality, then the concepts collide. We cannot have an intentional
:> accident or an accidental intention."
:
:Sure can! I reach into an opaque bag with red and blue marbles in it,
:mixed up. I select a red marble from the bag. Does that mean I intended
:to select a red marble? Not at all. I intended to select a marble.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
:Whether it's red or not was not intentional, even tho red was selected.
:Hence, selection without intention.
You just contradicted yourself. You intended to select a single marble
of unknown colour from the bag. Importantly, you selected a *marble*, not
a colour. The colour just happens to be a property the marble you selected
and you didn't select according to that property.
:To say we can't have an intentional accident is easy to disprove with
:silly word games, too. If you annoy me on the road and I dent your car
:on purpose, then we've had an intentional accident.
As you well know, Sproul is using 'accident' in the strict sense of "event
without apparent cause". His statement is true when 'accident' is interpreted
in that sense.
You've used the term 'accident' in the loose sense of 'mishap'.
:Just for the record, no offense intended, I don't really believe in
:evolution either (as I feel it's a non-disprovable theory),
Well, you're right here, at least. As you suggest, the theory of evolution
(by natural selection) isn't falsifiable - one of the criteria for a
scientific theory, according to philosopher of Science, Karl Popper.
However, I'm curious. By evolution, do you mean *any* kind of evolution?
There are at least three kinds:
- Darwinism (evolution by 'natural' selection)
- Deistic Evolution (an impersonal god created the universe with a
set of physical laws to run it, wound it up like a clock, and left
it to run)
- Theistic Evolution (a personal God created the universe and
continually sustains it (including directing an evolutionary process) ).
If none of these, what do you believe?
:I really am
:looking forward to hearing something other than wordgames, and I'd never
:post something like this publicly except that Don doesn't give out his
:private email address.
Don (Harrison). donh at syd.csa.com.au
^^^^^^^^^^^^^^^^^^^^^^
Replace " at " with "@". In any case, we've corresponded in the past so
you already have my email address.
Don Harrison
:On Tue, 3 Nov 1998 06:23:52 GMT, Don Harrison <nos...@thanks.com.au> wrote:
::Purpose implies intent.
::Intent implies design.
::Therefore, our genetic code was designed.
::QED.
...
:If you're being serious, then I'll dispute the first one; the
:"purpose" is something that human cognition has attached to the
:genetic code, and thus any "intent" we abscribe to the purpose,
:another artifact of human thought.
Okay, I think I know what you mean.
So, when I see Darren New's car hurtling down the road in my direction
and think he intends to run me over, I should be unconcerned because it's
just my human congnition playing tricks on me, making me think his purpose
is to cause me grievous bodily harm? :)
I'm not sure how far I'll be able to continue with this line of
rationalisation if I subsequently find myself in hospital in traction. :(
:Are you serious or poking fun at some people?
Mostly serious, but I'm now poking fun at you. :)
Darren New
> :"purpose" is something that human cognition has attached to the
> :genetic code, and thus any "intent" we abscribe to the purpose,
> :another artifact of human thought.
>
> Okay, I think I know what you mean.
>
> So, when I see Darren New's car hurtling down the road in my direction
> and think he intends to run me over, I should be unconcerned because it's
> just my human congnition playing tricks on me, making me think his purpose
> is to cause me grievous bodily harm? :)
If there's no driver behind the wheel, and it just looks like the brakes
let go, wouldn't you react differently than if there was someone behind
the wheel giving you the finger as he accellerated towards you?
I expect you would, if you were rational enough and had enough time to
compute that. If it's just empty, dodging to the side will help.
Otherwise, getting into a narrow alley is probably your better bet.
Hence, it isn't the *car* having the intentions, it's the *person* who
is driving, if any. Or do you think the car intends to hurt you, and
hence purposely blew the brake line so it could coast down the hill at
you?
If you fall in the lake, is the water out to get you, upset, perhaps,
that you disturbed its serenity? Do you think a virus really intends to
hurt you, and actively persues you? How about the bees that sting you
when you accidentally bump against their hive? How high up the chain do
you have to go before inatimate objects start having intentions? If you
think the lake intends to hurt you, I think you're using the word in a
less-than-standard way.
Now that this is *way* off topic, I'll stop posting.
--
Darren New / Senior Software Architect / First Virtual Holdings Inc
http://www.fv.com or in...@fv.com -=|=- PGP Key: ftp://ftp.fv.com/pub/fv
Fingerprint: 61 7D AF 9E 00 CC C2 ED / D8 4C D7 AA E4 C2 A0 73
Darren New
>
> Darren New wrote:
> :Just for the record, no offense intended, I don't really believe in
> :evolution either (as I feel it's a non-disprovable theory),
>
> Well, you're right here, at least. As you suggest, the theory of evolution
> (by natural selection) isn't falsifiable - one of the criteria for a
> scientific theory, according to philosopher of Science, Karl Popper.
>
> However, I'm curious. By evolution, do you mean *any* kind of evolution?
Oh, obviously I don't mean evolution in terms of disbelieving that
Windows98 is an evolutionary step from Windows95.
> There are at least three kinds:
>
> - Darwinism (evolution by 'natural' selection)
> - Deistic Evolution (an impersonal god created the universe with a
> set of physical laws to run it, wound it up like a clock, and left
> it to run)
> - Theistic Evolution (a personal God created the universe and
> continually sustains it (including directing an evolutionary process) ).
And of course, you missed out on all the others, like there's a bunch of
gods who toy with humanity as peices on a big game board, and there's a
bunch of gods each of which embody an aspect of the natural world and
keep it in harmony, and the one where the universe just started out and
then there's some intelligent species that arises and thinks into
existance a god or two who then keep the universe working, not unlike
"Midnight at the Well of Souls".
Of course, maybe the "personal god" you mention is really aliens, who
came and left geneticly seeded beings on Earth, and when the dinosaurs
got wiped out, came back and seeded again, periodically returning to
tweak new species into existance as old ones fail to work out. This
seems quite consistant with current knowledge of old species (as far as
I know), but of course pushes off to the question of where the aliens
came from.
> If none of these, what do you believe?
Unlike many religeos folks, I'm perfectly comfortable saying "I don't
know." I see no good evidence in favor of any of the proposed theories
above. Since the physicists can't even figure out what makes the
universe real (in terms of collapsing the state vector) I don't see
evolution as all that important in retrospect. I've also decided that
the existance or nonexistance of God is going to be irrelevant to my
life, as I wouldn't follow an evil God and I already do what I think a
good God would want. Now, of course, if God chooses to disprove
darwinism by addressing me in person, then I'd cchange my mind. I'd be
glad to go into more depth in email.
> Don (Harrison). donh at syd.csa.com.au
> ^^^^^^^^^^^^^^^^^^^^^^
> Replace " at " with "@". In any case, we've corresponded in the past so
> you already have my email address.
Erf. Sorry about that. D'oh!
(And no, I don't keep every email address I've corresponded with. :-)
OK, really now, this is the last public post.
Matt Kennel
:Don Harrison wrote:
:>
:> Darren New wrote:
:
:> :evolution either (as I feel it's a non-disprovable theory),
:>
:> Well, you're right here, at least. As you suggest, the theory of evolution
:> (by natural selection) isn't falsifiable - one of the criteria for a
:> scientific theory, according to philosopher of Science, Karl Popper.
Well, if one observed significant Lamarckian inheritance and consequences
of that, then most people would consider Darwinian evolution mostly falsified.
If genomes weren't both inherited and mutable, then it would be falsified.
But you can argue that all various theories of stellar and galactic
evolution can't be falsified. And yet a few simple explanations from
nuclear physics combined with sophisticated fluid dynamics end up
explaining a very wide diversity of odd and freaky observed phenomena.
That's power. Evolution is in the same category.
:Unlike many religeos folks, I'm perfectly comfortable saying "I don't
:know." I see no good evidence in favor of any of the proposed theories
:above. Since the physicists can't even figure out what makes the
:universe real (in terms of collapsing the state vector) I don't see
:evolution as all that important in retrospect.
There's been lots of progress on the "collapsing the state vector
problem", mainly in the realization that that's a lousy description of
what really happens, and the state vector "appears" to collapse, but
everything still evolves unitarily and continually by the Heisenberg
equation. There is No Man Behind The Curtain.
Everything isn't settled, but the contemporary "decoherence"
theoretical effort appears to have a good chance at making quantum
mechanical theory just like another part of regular physics.
(I.e. Einstein was right).
--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD
*
* "do || !do; try: Command not found"
* /sbin/yoda --help
Kai Quale
>
>Juergen Schlegelmilch wrote:
>
>>Ell <e...@access.digex.net> wrote:
>>>
>>>We'll have to agree to disagree. My own and others experience has
>>>demonstrated to us that OO features, some major ones being:
>>> a.) entities with behavior and properties
Which could just as well be interpreted as streams and functions with
parameters and local variables, or expressions of God's thought, or baloney
sandwiches, or whatever is useful at the moment.
>>> b.) encapsulation
Where ?
>>> c.) polymorphism
Where ?
>>> d.) identity
Where ? Social security numbers ?
>>> e.) state
Could just as well be modeled by functions and streams.
>>> f.) commonality
I'm not sure what you mean by this.
>>>exist objectively in the world.
>
>>You and others may
>>see the world from an OO point of view, but that is _your_
>>view on the world
>
>You mean to tell me that you don't see these features existing
>objectively in the world?
>
>I can look around right now at the wastebasket, the chair, my pants,
>the phone, the keyboard and they all exhibit OO features. I can test
>to see if the OO features are objective by formulating my OO feature
>observation as hypothesis which can be tested. When I test those
>hypothesis, the hypothesis are proven as true. Hence I know that OO
>features exist objectively. My view of OO features in the things
>around me is not just subjective gut feeling.
>
>Elliott
What you see is sensory impressions your brain interprets/makes a model of:
"That's a wastebasket, that's a chair..." It's extremely USEFUL to have these
models (otherwise we would have to decide from scratch every time), but that's
what they are - models. You COULD make other, less useful, models: "That's a
spacecraft, that's a torture instrument". They wouldn't be WRONG, just not
useful.
kai
Darren New
> >>> f.) commonality
> I'm not sure what you mean by this.
I think he meant "commonality" in the sense that all instances from the
same class have a certain commonality, and all heirs of the same class
have a certain commonality. All pants have a certain commonality of one
big hole and at least two smaller holes....
> >I can look around right now at the wastebasket, the chair, my pants,
> >the phone, the keyboard and they all exhibit OO features. I can test
> >to see if the OO features are objective by formulating my OO feature
> >observation as hypothesis which can be tested. When I test those
> >hypothesis, the hypothesis are proven as true. Hence I know that OO
> >features exist objectively. My view of OO features in the things
> >around me is not just subjective gut feeling.
I think you're seeing physical properties and using them to deduce OO
properties, when in reality the things you can measure don't imply
OOness any more than they imply God's intentions or anything else.
If I rub a rubber comb thru my hair and pick up bits of paper with it,
are the comb and the paper now the same object? Is the picture bolted to
the wall actually part of the wallness? You'd say no, but only because
the intermolecular forces between paper and comb, between picture and
wall, are less than the intermolecular forces between paint and wall,
between tooth and handle of the comb.
What makes the collection of copper and rubber and paint and glass and
high explosives a "space shuttle"? If you can give me an objective
physics-type answer to that, without reference to what it's supposed to
do, or what it will do in the future if we add in a bunch of stuff
that's distinctly not a space shuttle (like ground control and
astronauts), then you might have a tennable position.
--
Darren New / Senior Software Architect / First Virtual Holdings Inc
http://www.fv.com or in...@fv.com -=|=- PGP Key: ftp://ftp.fv.com/pub/fv
Fingerprint: 61 7D AF 9E 00 CC C2 ED / D8 4C D7 AA E4 C2 A0 73
The vase the maid the agency the neighbor recommended hired dropped
broke.
Robin Y. Millette
[big snip]
> PPS: One has to be careful even with direct results because our
> senses are not perfect. There are several visual tricks that
> show that our eyes betray us in certain circumstances. Even
> our nerves are not perfect: neurologists can induce visual
> patterns into our conciousness using electric impulses. Taken
> to the extreme, one cannot even be sure that there exists
> anything except the own conciousness (read Hofstadter and
> Dennett "The Mind's I" for a very interesting discussion of
> all this).
I would suggest, by the same author, the book "Godel, Escher, Bach",
where he devotes chapter 26 to Typogenetics, a simplification of
genetics with only 16 codons, each made up from only 2 bases. It's a
simplification, yet it remains close to the nature of genetics, where
data and code play each others role, at different levels.
One nice url would be:
http://www.nada.kth.se/~kai/lectures/geb.html
Search for "Self-Ref and Self-Rep" on that page, and you shall find :)