Data has no meaning...
Jan Roland Eriksson
[required pause]
(and you can all stop laughing now :)
Preface:
A discussion in a opera.* NG about the 'block' and 'inline' groups of
element declarations in HTML, something Arjun once said about "money"
some time back, and my "Schrödingers Cat" screen saver at work, are
things that have contributed to this "thinking" activity.
I may not be the best of writers, but what the heck, I will give it a
try anyway to try to illustrate what my "thinking" came up with.
(and it will be on topic for this NG, just read on :)
About the Subject; "Data has no meaning..."
Let me qualify that as to mean data that is in one way or another in
"transition" between points of origin and an arbitrary number of
targets. Data that is in this state of transition is nothing but a
series of pretty meaningless state variations within the transfer
system, it may be variations in air pressure, ones and zeros in an
electrical wire or light impulses in a glass fibre, that's all.
But some one may object; "These state variations may appear in some
systematic sequence (encoding), that enables them to carry some meaning
with them while they are in transfer".
Sure, encode/decode info may be there as a part of the data stream but
the data stream needs to be finally "decoded and recognized" by a
receiver before it can be reassembled into something that has a
"meaning" attached to it.
So, my take is still that; "while data is in transit there is no meaning
to it...". If some one talks to me in Swahili, I will not be able to
"decode and recognize" so effectively the speakers data is without
meaning to me from the very instance it leaves his lips and I'm made
aware of its existence.
The roots:
In order for us humans to be able to communicate with each other, it is
required that we have some form of common ground as in common habits,
common family procedures, common cultural based activities etc...
I.e. there must be a level of individual recognition of activities that
results in a common basic knowledge of what to expect from other people,
and from where we can build methods (encode/decode schemas) for
communication.
We are all equipped with a set of assets (with different levels of
usability depending on what life has kept as a surprise for us)...
...but most of us can see, hear, feel, smell, taste, talk and move.
(where "talk" can be seen as just a special case of "move")
On top of all this there is a "brain" that controls the output devices
as in muscles where ever they are located in our bodies, and reads the
input from our five senses.
As an example we can look at the process of "talk - listen" where one
"brain" originates data, sends it into a transfer state through the air
to be picked up by a listening ear for further transfer to a receiving
"brain"
My take here is that the actual transfer part of neuronic activity on
both sides, and the intermediate variations in air pressure between the
two, has no inherent meaning to it.
Meaning, is at the originating side according to some, for the
originator known, way to create meaning.
The same, or close to same, meaning _may_be_recreated_ at the receiving
side. The level of success of that process of recreation will all depend
on if there is a level of common experience among the persons involved.
In essence, human communication is all about being able to _recognize_
given input in terms of what it meant "yesterday" and go on to assume
that it means just about the same today, and further to interpret a
detailed meaning of a new message based on that assumption.
So let's sum that up as "all communication between humans takes place
over a process based on recognition", and this is totally independent of
what muscles and senses that are used in the communication process.
No recognition means communication problems, simple as that.
So where does "Schrödingers Cat" come into this?
Well, just in the same way as we can not know if the cat is dead or
alive, when we decide to let the wave function collapse by making an
observation of the current status of the cat, there is no way to know
what data in transfer really means. The "experiment" must be brought to
a decisive point of "opening the box" but this time in the sense of
letting received data be decoded and recognized against some level of
common experience between the originator and the receiver.
Arjun once mentioned "money" as representing a very well known "common
level of experience". Most humans know pretty well that money can be
used to let the result of human activities in one part of the world
create new human activity in another part of the world.
But money in it self has no value, its only at the ends of the transfer
chain where money is produced or consumed that it creates a value, and
then only because we _recognize_ that usage of money.
Author stylesheets are required!?!?
A "provoking" statement/question in this NG, right? :)
Let me illustrate why I feel, more and more, that it is correct to say
that an author stylesheet shall be sent along with the actual data, to
be used by the receiver of that data.
We have a teacher in history standing at the pulpit in front of his
students, and the agenda of today may be "The battle of Lützen in
November 1632" or maybe the "Storming of the Bastille in 1789" or some
such event of important historical interest.
The historical facts around just those two events are pretty well
established through research, so we could easily say that our teacher
has a good foundation for his presentation in the form of verified data.
E.g. we know details like that the Swedish King Gustav II Adolf was
wounded to death on November 6, 1632; or that "Free the prisoners of the
Bastille" is an after construct, and was not the real reason for the
event that came to start the French Revolution in 1789, etc...
One version of our hypothetical teacher knows nothing about how to
"style" his presentation, he gives his speech on the subject as a stream
of highly correct content, "marked up" to present the data in chunks of
syntactic components; headlines, paragraphs, lists, tables, etc...
His view is that it's his students job to "decode and recognize" his
presentation, (i.e. attach the correct semantic to the components) he
only provides the material to feed that process.
The other version of our hypothetical teacher has a living interest in
history, gives his speech using exactly the same "content and markup"
but adds colorful intonation, effect full pausing, some "body language",
expressive faces; all of it geared up to trigger the fantasy of his
students and make them start creating their own "user stylesheets on the
fly" based on that.
This type of presentation will catch the interest of the students first
of all, but it will also trigger a better "recognition of input" based
on the level of "common experience" that exists between the teacher, the
students and the historical people involved in scenarios described.
Needless to say, our first teacher will have a few sleeping students at
the end of his speech, while our second teacher will most probably
trigger and "after debate" among some of his students where they
continue to discuss the content of the presentation.
The WWW and what it is good for...
Now apply this teacher/student model to data made available on the WWW.
It's pretty much the same as I think of it. If some person thinks that
s/he has something to say and puts the energy in to actually make the
info available to others, it would be only natural for this person to
try to make the data available in a form where it becomes easy to
"decode and recognize" by some one that find reason to take part of it.
At the basic level we still have a markup syntax, and at least in the
case of defined SGML applications, also a semantic definition for the
syntactical components. Descriptive markup is simplicity it self.
Now if I make only the marked up content available for a user, what I'm
effectively doing is to place my self in the role of "our first
teacher". My markup and content may be ever so correct but since I'm
letting the users first impression of it be through his clients default
presentation mode (possibly somewhat enhanced by some standard user
stylesheet) it is still all up to the user to do his own work to try to
find out about the more subtle details that I already know about.
This situation has always been at hand in the typographical world as
well as in the SGML community. A typographer that is assigned the task
of typesetting some publication _must_ spend a considerable number of
hours to read the text before he starts working. He may even at some
points make contact with the author to discuss some parts of the content
to find out about author intentions that may not be described in the
original manuscript. All of that to come up with a best possible
presentation in print, as to punctuation, white space balance, page
disposition etc...
The same thing holds true for an SGML based document; there is no way we
can put just any available DSSSL programmer at the work of creating a
specific output from content in SGML markup.
That guy needs to go through the same process as all others involved,
"decode and recognize" the content, build a mental impression of what is
really said, and from there come up with a transformation of the content
into a good presentation.
But the higher level of what a presentation _really_ is, seems to be
forgotten most of the time when the WWW is involved...
The author stylesheet, and where it can help...
If I present some marked up information on the www and decisively sends
a stylesheet along with it, I am basically placing my self in the role
of the "second hypothetical teacher".
First; a user will get an initial presentation that helps to keep up the
interest for the content (i.e. not fall asleep :)
Second; a good author stylesheet may be of help for the user to "read
between lines" and find out more from the content as a result of better
input to his "decode and recognize" process, i.e. trigger an extra level
of fantasy and association that may produce a better understanding of
the original content.
And my take here is that this last level of triggered "fantasy and
association" is what constitutes the ultimate "user stylesheet", as the
final result of the "decode and recognize" process will now be the
cascaded result of my author stylesheet and all levels of "defaults in
between" plus the final interpretation in the users brain.
Epilog...
Communication is a process that goes on between human brains, totally
regardless of what media that is in use to convey a message, and "data
in transit between brains" has no meaning really, while it is in
"transit".
"Meaning" is created at point of origin, and recreated to something more
or less similar at point of reception, through a process of "decode and
recognize".
It is the responsibility of a message originator to make all information
available for a potential message receiver so that the receivers final
interpretation of the message can be done with a minimum of
"distortion". Be that all of syntax, semantics and style.
Just as in Schrödingers proposed experiment with the cat, we have no
possibility to know if an "author message" ever gets received, much less
"decoded and recognized" in a correct or distorted way; but it is still
an obligation of authors to send all available information out on the
WWW that could help the recognition process along if it ever gets
initiated.
That is why author stylesheets are required additional components of a
document that is made available over the WWW.
--
Jan Roland Eriksson <r...@css.nu> .. <URL:http://css.nu/>
Jerry Muelver
<r...@css.nu> wrote:
>I have been thinking!
>
<snip thinking, and paste into my "Keep" file...>
>
>That is why author stylesheets are required additional components of a
>document that is made available over the WWW.
Pattern Resonance Theory (my unpublished invention) models
communication interactions along similar lines, but without
the "data has no meaning" gap between sender and receiver.
Claude Shannon died only a couple of weeks ago, so it's too
early to throw information theory out entirely.
Stylesheets in the PRT context are pattern-clarifiers,
simplifying harmonics and strengthening the pattern's
persistence in order to enhance resonance in the receiver's
"common ground" patterns.
We really have to find a pub somewhere and sit down to talk
this out....
---- jerry
--