TRM - Morbidity has set in, or not?
Frank Hamersley
http://www.dbms2.com/2005/10/10/17/
It was authored by a "Curt Monash" - is he known to any CDT'ers and/or
credible?
Is there any veracity in _any_ of the claims made by _any_ of the parties?
Given lots of mud gets flung as the discussion proceeds so I wondered if
there was any corroborative or contrary sources.
Cheers, Frank.
J M Davitt
Curt Monash has been discussed a few times at dbdebunk.com.
[This is my remembrance.]
There was, a while ago, a Curt Monash online article in Computerworld
in which he proposed a DBMS2. The tenet, as I recall, was that SQL
databases - which he described as relational - were ill-suited for
'modern' data storage needs.
Fabian Pascal, I believe, took him to the woodshed in a vigorous
exchange of "comment on this article" posts that, apparently, were
edited by some keeper of the website.
Somewhere along the line, Curt brought in Chris Date's name. I don't
recall the specifics, but it amounted to "Curt said Chris said..."
when, in fact, Date hadn't participated at all.
Date took issue with that and posted something on dbdebunk.com.
Monash then obliquely engaged Date in the article you cited.
{/remembrance]
If you read it - and have any familiarity with what was supposed to
be Date's "Go Faster!" - you might realize that many of the criticisms
are off-point and that everything that's presented as a revelation was
previously and carefully disclosed by Date himself.
I personally heard Date do this two summers ago...
Frank Hamersley
OK - the scene is set for a scrap!
> If you read it - and have any familiarity with what was supposed to
> be Date's "Go Faster!" - you might realize that many of the criticisms
> are off-point and that everything that's presented as a revelation was
> previously and carefully disclosed by Date himself.
>
> I personally heard Date do this two summers ago...
Yup - I was particularly interest to know if the "rational market
economy" had brought the TRM adventure to a sticky end?
Cheers, Frank.
Marshall Spight
>
> Is there any veracity in _any_ of the claims made by _any_ of the parties?
>
> Given lots of mud gets flung as the discussion proceeds so I wondered if
> there was any corroborative or contrary sources.
The "transrelational" stuff doesn't have much written about it. I can't
find
anything to suggest that it's anything besides a traditional column
store.
Various parties, including FP himself, have on occasion said, "oh no,
it's much more than that" but they don't back it up at all, so their
claims
are unevaluable.
Michael Stonebreaker has a small company that is selling a column
store; it looks quite interesting.
As an aside, I note that Mr. Pascal spends quite a lot of time telling
us
how smart and logical he is, and how all of his opponents are stupid,
ignorant, and illogical. He also spends a lot of time name calling, and
even making fun of other people's names! I wonder: has he ever
accomplished anything that would back up his claims? Has he ever
published a proof? Published a paper in a peer-reviewed journal?
Made use of any formal methods? Written any software? To my
knowledge, he has published two books, one of them "Practical
Issues is Database Management". Which was a fine book, although
the last person I leant it to noted "you could hear the axe grinding
on every page."
Marshall
paul c
> Frank Hamersley wrote:
>
>>Is there any veracity in _any_ of the claims made by _any_ of the parties?
>>
>>Given lots of mud gets flung as the discussion proceeds so I wondered if
>>there was any corroborative or contrary sources.
>
>
> The "transrelational" stuff doesn't have much written about it. I can't
> find
> anything to suggest that it's anything besides a traditional column
> store.
Funny thing that. Although the patents are written in language that
tries to be all-encompassing (maybe all patents these days try to claim
every possible technique - I don't know), the central idea does seem
obvious (storage adjacency used for column order with links to relate
values in 'rows' versus the typical impl'n that uses storage adjacency
for relating values in rows and links for ordering). Still, I've seen a
few implementations towards similar ends and at one time followed a lot
of the literature and I've never actually seen anybody implement or
describe anything quite like it, so even though it seems obvious it
doesn't look mainstream traditional to me. But one person couldn't
survey the whole field and it seems plausible to me that somebody else
has done similar, perhaps in programs that aren't involved with
conventional databases, even though I can't point to such an effort.
> Various parties, including FP himself, have on occasion said, "oh no,
> it's much more than that" but they don't back it up at all, so their
> claims
> are unevaluable.
> ...
I believe he is bound by some non-disclosure agreement and since the
business involved has gone awry, he is stuck. So you can hardly blame
him except perhaps for not having the foresight to sign a NDA that had
no expiry date.
What is more mysterious to me is whether all the secrecy is because TRM
is not entirely patentable, eg., is it more trade secret than novel
technique?
> Michael Stonebreaker has a small company that is selling a column
> store; it looks quite interesting.
> ...
After a promising start many years ago, Stonebraker has long been a tech
salesman for various fads. Is that the column store out of MIT dating
from the early 1990's?
> As an aside, I note that Mr. Pascal spends quite a lot of time telling
> us
> how smart and logical he is, ...
Well, I've read at least a few hundred pages written by him and although
I didn't understand parts of it, I can't recall him boasting about
himself, not even once.
> and how all of his opponents are stupid,
> ignorant, and illogical. He also spends a lot of time name calling, and
> even making fun of other people's names! ...
From what I've read, most of his "opponents" haven't got the foggiest
of what he is talking about.
> I wonder: has he ever
> accomplished anything that would back up his claims? Has he ever
> published a proof? Published a paper in a peer-reviewed journal?
> Made use of any formal methods? Written any software? To my
> knowledge, he has published two books, one of them "Practical
> Issues is Database Management". Which was a fine book, although
> the last person I leant it to noted "you could hear the axe grinding
> on every page."
> ...
Yes, he has performed a public service, mostly without gain to himself
whereas many of the "opponents" puff fools-gold because it is in their
own interest to make systems bigger and more grandiose than they need to
be. IT being a modern-day goldrush is full of carpet-baggers always
promising more than they can deliver. Some chemistry and other Ph.D's
got into programming because there's more money in it plus they were
second-rate in their first field anyway. Peer review means less and
less now and is often a joke - more like a cover to protect "jobs for
the boys". CS credentials are usually a tawdry peerage - most of those
"peers" should demand refunds of their tuition fees but only a few have
the brains to see this and even fewer have the guts.
p
Gene Wirchenko
wrote:
>Marshall Spight wrote:
>> Frank Hamersley wrote:
>>
>>>Is there any veracity in _any_ of the claims made by _any_ of the parties?
>>>
>>>Given lots of mud gets flung as the discussion proceeds so I wondered if
>>>there was any corroborative or contrary sources.
>> The "transrelational" stuff doesn't have much written about it. I can't
>> find
>> anything to suggest that it's anything besides a traditional column
>> store.
> > ...
>
>Funny thing that. Although the patents are written in language that
>tries to be all-encompassing (maybe all patents these days try to claim
>every possible technique - I don't know), the central idea does seem
That seems to be it. The few that I have checked myself seem to
have related claims so that if one part of the patent is invalidated,
another part may still apply.
>obvious (storage adjacency used for column order with links to relate
>values in 'rows' versus the typical impl'n that uses storage adjacency
>for relating values in rows and links for ordering). Still, I've seen a
>few implementations towards similar ends and at one time followed a lot
>of the literature and I've never actually seen anybody implement or
>describe anything quite like it, so even though it seems obvious it
>doesn't look mainstream traditional to me. But one person couldn't
>survey the whole field and it seems plausible to me that somebody else
>has done similar, perhaps in programs that aren't involved with
>conventional databases, even though I can't point to such an effort.
>> Various parties, including FP himself, have on occasion said, "oh no,
>> it's much more than that" but they don't back it up at all, so their
>> claims
>> are unevaluable.
>> ...
>I believe he is bound by some non-disclosure agreement and since the
>business involved has gone awry, he is stuck. So you can hardly blame
>him except perhaps for not having the foresight to sign a NDA that had
>no expiry date.
An NDA is my understanding, too. I do not know about the expiry
date bit though. BTW, I am a regular reader of dbdebunk.com.
>What is more mysterious to me is whether all the secrecy is because TRM
>is not entirely patentable, eg., is it more trade secret than novel
>technique?
Maybe, but it could be covered by simply being cautious until you
have something. Certainly, FP has a lot of bad to say about those
trumpeting the Latest Thing.
>> Michael Stonebreaker has a small company that is selling a column
>> store; it looks quite interesting.
>> ...
>After a promising start many years ago, Stonebraker has long been a tech
>salesman for various fads. Is that the column store out of MIT dating
>from the early 1990's?
FP has said that.
>> As an aside, I note that Mr. Pascal spends quite a lot of time telling
>> us
>> how smart and logical he is, ...
As a regular reader of dbdebunk.com, I call bullshit!
>Well, I've read at least a few hundred pages written by him and although
>I didn't understand parts of it, I can't recall him boasting about
>himself, not even once.
>> and how all of his opponents are stupid,
>> ignorant, and illogical. He also spends a lot of time name calling, and
>> even making fun of other people's names! ...
As a regular reader of dbdebunk.com, I call bullshit!
> From what I've read, most of his "opponents" haven't got the foggiest
>of what he is talking about.
>> I wonder: has he ever
>> accomplished anything that would back up his claims? Has he ever
>> published a proof? Published a paper in a peer-reviewed journal?
>> Made use of any formal methods? Written any software? To my
>> knowledge, he has published two books, one of them "Practical
>> Issues is Database Management". Which was a fine book, although
>> the last person I leant it to noted "you could hear the axe grinding
>> on every page."
>> ...
Perish the thought that someone should have an agenda. I like
FP's of supporting the RM. It is far better than these slimy Latest
Things that keep popping up from under rocks.
>Yes, he has performed a public service, mostly without gain to himself
>whereas many of the "opponents" puff fools-gold because it is in their
>own interest to make systems bigger and more grandiose than they need to
>be. IT being a modern-day goldrush is full of carpet-baggers always
>promising more than they can deliver. Some chemistry and other Ph.D's
>got into programming because there's more money in it plus they were
>second-rate in their first field anyway. Peer review means less and
>less now and is often a joke - more like a cover to protect "jobs for
>the boys". CS credentials are usually a tawdry peerage - most of those
>"peers" should demand refunds of their tuition fees but only a few have
>the brains to see this and even fewer have the guts.
FP is plainspoken, and he backs up his statements. The
namecalling is on the other side.
Sincerely,
Gene Wirchenko
J M Davitt
> Frank Hamersley wrote:
>
>>Is there any veracity in _any_ of the claims made by _any_ of the parties?
>>
>>Given lots of mud gets flung as the discussion proceeds so I wondered if
>>there was any corroborative or contrary sources.
>
>
> The "transrelational" stuff doesn't have much written about it.
Yeah, other than the patent application - which is tortuous - there's
little out there. To confuse the issue, other products are described
as transrelational - Cache, IIRC, is one.
I can't
> find
> anything to suggest that it's anything besides a traditional column
> store.
> Various parties, including FP himself, have on occasion said, "oh no,
> it's much more than that" but they don't back it up at all, so their
> claims
> are unevaluable.
It *is* much more that a column store storage scheme. I don't know
whether you've read a description of TRM, but it features (a) a
not-so-surprising ordered collection of observed values, (b) a mildly
clever permutation and inverse permutation index, and (c) a very clever
"record reconstruction table."
> Michael Stonebreaker has a small company that is selling a column
> store; it looks quite interesting.
If we're talking about the C-Store he was involved with, it does feature
a column-wise storage scheme. But, unlike TRM, values will appear in
storage just as many times as they appear in the "logical" records
being represented. C-Store makes extensive use of compression and,
IIRC, is able to performs restricts and selects based on the compressed
representations of values. Besides that, one of C-Store's big features
is a technique for replicating a data store at different sites and
knowing, at all sites, the most recent instant for which all sites have
the same values.
> As an aside, I note that Mr. Pascal spends quite a lot of time telling
> us
> how smart and logical he is,
I'm not sure that's right; he's certainly not shy about voicing his
opinions -- and presenting them as certainties. Many of them probably
are, although the context in which he makes his remarks is often
opaque.
and how all of his opponents are stupid,
> ignorant, and illogical.
That's certainly true -- but he explains why. At least, every now and
again.
He also spends a lot of time name calling, and
> even making fun of other people's names!
Hmm.. I've never seen him cast the first stone.
I wonder: has he ever
> accomplished anything that would back up his claims? Has he ever
> published a proof? Published a paper in a peer-reviewed journal?
> Made use of any formal methods? Written any software? To my
> knowledge, he has published two books, one of them "Practical
> Issues is Database Management". Which was a fine book, although
> the last person I leant it to noted "you could hear the axe grinding
> on every page."
Maybe so. I've read the book (even caught a mistake which resulted in
an entry in his Errata) and I have to admit: at time, his tone could be
described as "reactionary." But, you know, Codd's RM/V2 was somewhat
like that, too. Reading between the lines, I could sense that some of
what was in that volume - T-joins, for example - were reactions to
features the network model implementations offered while others - what
Codd called "required features," I think - were meant to point out
defects in other database implementations.
HungryLion
The whole Pascal/Date/Darwen thing is a cult. No suprise that their
members here abide by the motto "Go ye, and debunk". Typical cult
behaviour. And like all cults the idea is to seperate the fools from
their money. The TRM baloney serves two purposes:
a) Hype to sell books (Dates 8th version) b) IMplementation issues in
regards to their manifesto
Why waste time with this cult when they have nothing to offer in
comparison with *real* researchers.
Go to the leonid libkin site, or seek out hartmann, link, etc. Thats
just a few among many. Anyways this from the alphora newsgroup
server:
alphora.dataphor.dae
Johan,
Let me expand a little. I've seen an impressive prototype of the
technology
and have gleaned about as much about the technology as we've been able
to.
Though we do think the technology is interesting, we do not believe it
is
necessarily the ultimate solution for Physical Data Independence.
Instead,
in truth it seems to represent an optimization for access patterns of a
certain variety (just as B-Trees do for another variety). It is our
position that the ultimate solution is not a certain access pattern,
but
rather unlimited variability between the logical and the physical. Do
a
search for GMAP for a taste.
Best,
--
Nathan Allan [Alphora]
"Johan Dufour" <johan....@datarutin.se> wrote in message
news:gqd1vjwC...@windu.softwiseonline.com...
> Has Alphora any contact with Required Technologies Inc. or plans to get
> Dataphor working with some TransRelational (TM) device?
>
dawn
> Marshall Spight wrote:
> > Frank Hamersley wrote:
> > As an aside, I note that Mr. Pascal spends quite a lot of time telling
> > us
> > how smart and logical he is,
>
> I'm not sure that's right; he's certainly not shy about voicing his
> opinions -- and presenting them as certainties. Many of them probably
> are, although the context in which he makes his remarks is often
> opaque.
>
> and how all of his opponents are stupid,
> > ignorant, and illogical.
>
> That's certainly true -- but he explains why. At least, every now and
> again.
>
> He also spends a lot of time name calling, and
> > even making fun of other people's names!
>
> Hmm.. I've never seen him cast the first stone.
I took several years of stones and did not toss one until a
light-hearted lob in a blog entry a few weeks ago named "Pascal Loses
Wager." I removed it when he removed his V.I. page. It was a rebus
that included a picture of an ass. That is/was my entire punch as best
I recall. I do not wish the man ill, and given his approach to fellow
travelers he seems a sad sort, but I would prefer not to continue to
take his kicks and I know that others feel the same way. (Yes, I said
the word "feel" although I know that is not part of the active
vocabulary of everyone here.)
I have no problems with logical arguments if he would like to take any
one of mine and lay out a logical rebuttal. I might even retract
something as I'm certain I have written some less-than-stellar
arguments and know I've asked "dumb questions" in my effort to learn
more. He has done a lot of "see how stupid she is" writing, without
countering any particular point with sound reasoning.
I recognize I'm a heretic among relational theorists, and I also know
what happens to heretics. Why are my questions and arguments not
strewn with mathematics? Because I have no disagreement with the
mathematics itself. I disagree with some choices typically deemed
standard in the industry related to how to model data for "large shared
data banks." I write about that which interests me, which might or
might not be of interest to one person or another -- how to write
quality data-based software in a fiscally responsible way. I recall a
comment on cdt about how if I cared about cost then nothing I had to
say would be of interest here -- very telling, but clearly false (e.g.
performance concerns and optimization are often mentioned, not to
mention costs of data redundancy, poor quality data, etc). I have
enough experience with budgets related to software development to have
formed opinions about the relative cost of various database solutions
for an organization (including data quality, personnel, configuration
and change management... costs), but not enough to prove my hunches, so
I continue my search, in spite of the bruises from the ad hominem
attacks.
Now while I sit in my oldest daughter and her husband's comfortable,
happy, recently-built home, I'm online to look up Phi Beta Kappa so I
understand the honor my youngest is receiving upon her graduation this
weekend. It makes me think I have at least dome some things right, and
I wonder if those who feel a need to burn heretics at the stake have
happy lives, given their need to throw stones at all. Maybe we need to
throw Pascal a party or something and thank him for the work he has
done that we do appreciate. I worry about him (and BB too). --a mom
Marshall Spight
> Marshall Spight wrote:
> >
> > The "transrelational" stuff doesn't have much written about it. I can't
> > find anything to suggest that it's anything besides a
> > traditional column store.
>
> few implementations towards similar ends and at one time followed a lot
> of the literature and I've never actually seen anybody implement or
> describe anything quite like it, so even though it seems obvious it
> doesn't look mainstream traditional to me.
My understanding is that the column store technique dates from perhaps
the 1970s, and has been used in many special-purpose stores over
the years. (This is not the same as being part of a dbms, but it's
still use.)
> > Various parties, including FP himself, have on occasion said, "oh no,
> > it's much more than that" but they don't back it up at all, so their
> > claims are unevaluable.
> > ...
>
> I believe he is bound by some non-disclosure agreement and since the
> business involved has gone awry, he is stuck. So you can hardly blame
> him except perhaps for not having the foresight to sign a NDA that had
> no expiry date.
I don't "blame" him, but I do note that he does not back up
any of the claims he makes relative to "transrelational". So his claims
are unevaluable. They might be true, or they might be false; we have
no way of knowing.
My policy towards unevaluable statements is to ignore them
until they become evaluable.
> > Michael Stonebreaker has a small company that is selling a column
> > store; it looks quite interesting.
>
> After a promising start many years ago, Stonebraker has long been a tech
> salesman for various fads.
Do you speak from personal experience here?
Stonebraker started studying relational in 1973. DBMSs that he's lead
the develompent of include Ingres, Illustra, Postgres, Cohera, and
StreamBase. He is the recipient of the IEEE Von Neuman medal,
which has also gone to such people as Gordon Bell, Fred Brooks,
Don Knuth, and Alfred Aho.
> > As an aside, I note that Mr. Pascal spends quite a lot of time telling
> > us
> > how smart and logical he is, ...
>
> Well, I've read at least a few hundred pages written by him and although
> I didn't understand parts of it, I can't recall him boasting about
> himself, not even once.
I will admit that FP writes much more about the deficiencies of his
opponents than about his own virtues. But doesn't a statement like
"I BACK UP my claims with evidence and logic" sound familiar?
In other words, he uses evidence and logic, and his opponents
don't. But what does he mean when he says he uses logic in his
arguments? He certainly doesn't mean predicate calculus. Have
you ever seen him use, say modus ponens? So in that context,
when he says "I use logic" what he is really just saying is "I'm
right." He's not talking about the use of formal methods.
> From what I've read, most of his "opponents" haven't got the foggiest
> of what he is talking about.
So he keeps telling us. (Although I do not disagree with the sentiment
that there is a lot of ignorance to go around.)
> > I wonder: has he ever
> > accomplished anything that would back up his claims? Has he ever
> > published a proof? Published a paper in a peer-reviewed journal?
> > Made use of any formal methods? Written any software? To my
> > knowledge, he has published two books, one of them "Practical
> > Issues is Database Management". Which was a fine book, although
> > the last person I leant it to noted "you could hear the axe grinding
> > on every page."
>
> Yes, he has performed a public service, mostly without gain to himself
> whereas many of the "opponents" puff fools-gold because it is in their
> own interest to make systems bigger and more grandiose than they need to
> be. IT being a modern-day goldrush is full of carpet-baggers always
> promising more than they can deliver.
In other words, he's a consultant and public speaker. I've done
both of those things myself.
But what I asked was, has he ever accomplished anything?
Can you find a mathematical proof anywhere on dbdebunk?
> Some chemistry and other Ph.D's
> got into programming because there's more money in it plus they were
> second-rate in their first field anyway.
This statement is mostly unevaluable. But I will offer as a
counterexample
this guy:
http://en.wikipedia.org/wiki/Van_Jacobson
who I used to work with. Wikipedia says his work "is said to have saved
the Internet from collapsing due to traffic in 1988-1989." His PhD is
in physics.
> Peer review means less and
> less now and is often a joke - more like a cover to protect "jobs for
> the boys". CS credentials are usually a tawdry peerage - most of those
> "peers" should demand refunds of their tuition fees but only a few have
> the brains to see this and even fewer have the guts.
In other words, everyone is stupid, so the fact that FP doesn't
publish in peer reviewed journals is irrelevant.
Publishing in peer reviewed journals remains the gold standard
for academic achievement. In other words, it is how we judge
"evidence and logic." The system is imperfect: like democracy,
it is the worst thing possible, except for every alternative.
It is certainly preferable to, say, putting up a website where
you make fun of your debating opponents names, control
every aspect of the presentation of the debate yourself, and
refuse to publish any corrections or updates that they send
you.
Marshall
paul c
> Frank Hamersley wrote:
>
>>I just stumbled over this exchange from Oct 2005 on the TRM...
>>
>> http://www.dbms2.com/2005/10/10/17/
>>
>>It was authored by a "Curt Monash" - is he known to any CDT'ers and/or
>>credible?
>>
>>Is there any veracity in _any_ of the claims made by _any_ of the parties?
>>
>>Given lots of mud gets flung as the discussion proceeds so I wondered if
>>there was any corroborative or contrary sources.
>>
>>Cheers, Frank.
>
>
This is the kind of name-calling that fp's so-called opponents are into
(partly because their flim-flam is devoid of competent questions). The
main reason is that a small minority of them are brighter, in a devious
way, than the rest of the crowd and realize that the biggest threat to
the various nonsense they peddle is what he says. Should this get back
to their relatively naive employers and customers there would be big
questions to answer which they couldn't. So they try to tear down the
man with various slurs and drown him out with a lot of noise because
they haven't much else to offer. Talk about cults - the sinister
minority works hard at cultivating a loud crowd majority to be their
flunkies. I'll bet most of the flunkies slinging mud, such as this
apparent wannabe, don't even know their strings are being pulled.
Quoting the Alphora man on physical independence doesn't do much to
disparage fp as much of what everything he writes about is on the
logical and the rest is about conceptual modelling.
No suprise that their
> members here abide by the motto "Go ye, and debunk". Typical cult
> behaviour. And like all cults the idea is to seperate the fools from
> their money. The TRM baloney serves two purposes:
> a) Hype to sell books (Dates 8th version) b) IMplementation issues in
> regards to their manifesto
> ...
Good example of the usual anti-fp BS - manifesto doesn't talk about trm
at all, pre-dated it by quite a few years. Using TRM to attack p/d/d is
bogus.
p
Marshall Spight
> Marshall Spight wrote:
>
> > anything to suggest that it's anything besides a traditional column
> > store.
> > Various parties, including FP himself, have on occasion said, "oh no,
> > it's much more than that" but they don't back it up at all, so their
> > claims are unevaluable.
>
> It *is* much more that a column store storage scheme. I don't know
> whether you've read a description of TRM, but it features (a) a
> not-so-surprising ordered collection of observed values, (b) a mildly
> clever permutation and inverse permutation index, and (c) a very clever
> "record reconstruction table."
Your paragraph above seems to me to be a pretty good description
of a column store with a fully inverted index. My understanding
is that these techniques are decades old.
Now, it is possible that "transrelational" is something more than
that, or not; I have no way of evaluating any given statement
about it. So far.
Marshall
Marshall Spight
>
> FP is plainspoken, and he backs up his statements. The
> namecalling is on the other side.
I do not consider abusive and insulting language to be "plainspoken"
since I don't consider that kind of language "plain." However I suppose
you could stretch the definition to include it.
I invite anyone to follow the link in the OP and see if FP engages
in any namecalling. Here it is again:
http://www.dbms2.com/2005/10/10/17/
If you are short on time, you can ^F and search
for the word "crap" and see who is using it and how, or search
for "ignora", which will get you "ignorant", "ignorance" and the
ever-popular "ignorami". Search for the word "baby" to see if
anyone is being condescending to the opposite party.
See who, if anyone, is making any broad, insulting generalizations
about people of a particular nationality. See if such generalizations
are on-topic or off-topic.
See if FP provides any verifiable links to any particular claim
that could be called "evidence."
(Note that I am not claiming that anyone in the referenced
exchange comes off well.)
Marshall
Bob Badour
> HungryLion wrote:
>
>> Frank Hamersley wrote:
>
> Quoting the Alphora man on physical independence doesn't do much to
> disparage fp as much of what everything he writes about is on the
> logical and the rest is about conceptual modelling.
Actually, a better critique would be to note that it is a straw man.
Neither Chris nor Fabian has ever, to my knowlege, claimed that TRM was
the ultimate answer to physical independence. I have heard one or the
other of them call it exciting and remark on things like linearity of
cost for joins etc. In that sense, I strongly suspect both Chris and/or
Fabian largely agree with Nathan's opinion.
The ultimate answer to physical independence is physical independence
itself. ie. the greatest flexibility possible to map a logical model to
physical stores.
As for myself, I am in Marshall's camp on the topic of TRM -- I lack the
information necessary to evaluate the claims.
Bob Badour
http://www.tdan.com/sms_issue34.htm
What makes you think Fabian has to provide a verifiable link to the
substantive debunking of an ignorant every time he posts? Do you require
the same of self-aggrandizing ignorants like Monash?
Dan
Far be it for me to judge and apologies for going off topic here, but
you've come a long way over years, not only in expressing database
theory concepts, but in the use of substantiative but valid logical
argumentation. I enjoy your rebuttals immensely now -- full of logic
and knowledge.
Regards,
Dan
David Cressey
"dawn" <dawnwo...@gmail.com> wrote in message
news:1147490119.0...@j33g2000cwa.googlegroups.com...
> that included a picture of an ass. That is/was my entire punch as best
> I recall. I do not wish the man ill, and given his approach to fellow
> travelers he seems a sad sort, but I would prefer not to continue to
> take his kicks and I know that others feel the same way. (Yes, I said
> the word "feel" although I know that is not part of the active
> vocabulary of everyone here.)
Speaking just for myself, I disagree.
My active vocabulary includes both "think" and "feel". I try not to use
them as interchangeable concepts.
Many people in today's society do just that, and cheapen both concepts in
the process.
I'm not sure about any other c.d.t. regulars. It wouldn't surprise me if
"the leisure of the theory class" consisted mainly in thinking.
Marshall Spight
> Marshall Spight wrote:
>
> What makes you think Fabian has to provide a verifiable link to the
> substantive debunking of an ignorant every time he posts?
I don't think that.
However, if one is going to regularly claim that one uses "evidence"
in one's argumentation, one needs to on occasion provide verifable
external sources, aka evidence. A prose paragraph might be logical
on its own, but it isn't evidence on its own.
Marshall
Marshall Spight
I thank you sir! You have made my day. :-)
Marshall
Marshall Spight
> > the word "feel" although I know that is not part of the active
> > vocabulary of everyone here.)
>
> Speaking just for myself, I disagree.
>
> My active vocabulary includes both "think" and "feel". I try not to use
> them as interchangeable concepts.
> Many people in today's society do just that, and cheapen both concepts in
> the process.
Hear hear! Nicely put; strongly agree.
Marshall
Bob Badour
> "dawn" <dawnwo...@gmail.com> wrote in message
> news:1147490119.0...@j33g2000cwa.googlegroups.com...
>
>>that included a picture of an ass. That is/was my entire punch as best
>>I recall. I do not wish the man ill, and given his approach to fellow
>>travelers he seems a sad sort, but I would prefer not to continue to
>>take his kicks and I know that others feel the same way. (Yes, I said
>>the word "feel" although I know that is not part of the active
>>vocabulary of everyone here.)
One wonders why she feels it appropriate to intrude in the first place?
By ignoring substantive responses to her drivel, whether due to intent
or capability, and by continuing to repeat defeated arguments, she very
much intrudes where she has no place or business.
> Speaking just for myself, I disagree.
>
> My active vocabulary includes both "think" and "feel". I try not to use
> them as interchangeable concepts.
> Many people in today's society do just that, and cheapen both concepts in
> the process.
I am not sure who you think you refer to. I do not use them as
interchangeable concepts, neither does Fabian, and I am not aware of
anyone here who does. Certainly, I agree many people abuse the verb 'to
think' when they have no intention of thinking.
> I'm not sure about any other c.d.t. regulars. It wouldn't surprise me if
> "the leisure of the theory class" consisted mainly in thinking.
Theory is an intellectual topic where feelings are mostly irrelevant.
The fact that people seldom express feelings with respect to theory says
nothing about their emotional capacity. Emotions do come into play to
recognize aesthetics and elegance.
I have found few people who are intellectually damaged who want to hear
that fact. When one considers that the intellectually damaged lack the
capacity to understand that damage, one can only expect them to feel
greater and greater frustration as they refuse to accept the one
cognition that will overcome their cognitive dissonance.
The more Dawn fobs her ignorant drivel onto others, the more she
demonstrates her intellectual damage. Presumably she is trying to
demonstrate her intellectual strength and seeks recognition instead of
the inevitable rejection.
In short, whether Dawn feels any emotional discomfort is entirely under
her own control: "Doctor! Doctor! It hurts when I do that!"
David Cressey
"Marshall Spight" <marshal...@gmail.com> wrote in message
news:1147490980....@i40g2000cwc.googlegroups.com...
> paul c wrote:
> > Marshall Spight wrote:
> > >
> > > The "transrelational" stuff doesn't have much written about it. I
can't
> > > find anything to suggest that it's anything besides a
> > > traditional column store.
> >
> > [...] Still, I've seen a
> > few implementations towards similar ends and at one time followed a lot
> > of the literature and I've never actually seen anybody implement or
> > describe anything quite like it, so even though it seems obvious it
> > doesn't look mainstream traditional to me.
>
> My understanding is that the column store technique dates from perhaps
> the 1970s, and has been used in many special-purpose stores over
> the years. (This is not the same as being part of a dbms, but it's
> still use.)
>
>
Marshall,
I'm going to start a new discussion, rather than divert this one. There are
several ideas that come together here,
including some that you have brought up in the past.
First, the idea that the index IS the database, from the article the OP
referenced.
Second, the idea of "content based addressing" from some of your old posts.
Third, the idea that the index and the table are redundant, but not
"harmful redundancy" in the sense that term is usually used to mean.
Fourth, hardware associative memories.
Fifth the question of whether a "column store" is any more fundamental than
a "row store".
Sixth the question of whether Starkey didn't use the "index IS the database"
concept in Interbase.
All of the above has little to do with the TRM, but may merit discussion in
its own right.
J M Davitt
> J M Davitt wrote:
>
>>Marshall Spight wrote:
>>
>>
>>> I can't find
>>>anything to suggest that it's anything besides a traditional column
>>>store.
>>>Various parties, including FP himself, have on occasion said, "oh no,
>>>it's much more than that" but they don't back it up at all, so their
>>>claims are unevaluable.
>>
>>It *is* much more that a column store storage scheme. I don't know
>>whether you've read a description of TRM, but it features (a) a
>>not-so-surprising ordered collection of observed values, (b) a mildly
>>clever permutation and inverse permutation index, and (c) a very clever
>>"record reconstruction table."
>
>
> Your paragraph above seems to me to be a pretty good description
> of a column store with a fully inverted index. My understanding
> is that these techniques are decades old.
Like the "inverted hierarchy?" Yes, that is old enough to be well
known. The TRM difference is that a value in a column appears
exactly once -- no matter ho many times it appears in the
representation. A further point not made is that each value need
appear only once in a domain. In other words, if there are many
columns holding date values, with the same value appearing in not
only one but many columns, the value need to be stored only once in
TRM.
This has huge significance: all values in a date domain covering
hundreds of years require fewer than 100,000 values. Time-of-day
precise to a second requires only 86,000 values. Given these domains,
adding records to a system would require no new values -- the domains
can be established before the first data arrive Social security
numbers? There are far fewer that 10^9 possible. License plates
numbers? What, 36^6 or 36^8 -- times 50? That's not a big gulp.
Names? Far fewer than one might think. If domains such as these
are enumerated before the system requiring the database is turned on,
it could conceivably operate for years without seeing a "new" value.
The benefit in the physical layer -- which is where all commercial
products now have trouble when "big data" come to the party -- are
is that space required for storing values becomes a mere tiny fraction
of what modern systems require. 1/1,000,000 is not an unreasonable
expectation.
Marshall Spight
> Marshall Spight wrote:
> >>
> >>It *is* much more that a column store storage scheme. I don't know
> >>whether you've read a description of TRM, but it features (a) a
> >>not-so-surprising ordered collection of observed values, (b) a mildly
> >>clever permutation and inverse permutation index, and (c) a very clever
> >>"record reconstruction table."
> >
> > Your paragraph above seems to me to be a pretty good description
> > of a column store with a fully inverted index. My understanding
> > is that these techniques are decades old.
>
> Like the "inverted hierarchy?" Yes, that is old enough to be well
> known. The TRM difference is that a value in a column appears
> exactly once -- no matter ho many times it appears in the
> representation. A further point not made is that each value need
> appear only once in a domain.
In an inverted index, the value appears exactly once.
http://en.wikipedia.org/wiki/Inverted_index
Marshall
Keith H Duggar
I took several years of stones ...
[I] and did not toss one until ...
I would prefer not to continue to take [stones]
I know that others feel the same way ...
I know that is not part of the ... vocabulary ... here ...
I have no problems ...
I might even retract ...
I recognize I'm a heretic ...
I also know what happens to heretics ...
I have no disagreement with the mathematics ...
I disagree with some [standard] choices ...
I write about that which interests me ...
I have enough experience with budgets ...
[I] have formed opinions ...
[I have] not ... prove[n] my hunches ...
I continue my search, in spite of [you]
I sit in my daughter['s] ... comfortable ... home ...
I'm online to look up Phi Beta Kappa ...
I understand the honor my youngest is receiving ...
her graduation this weekend ...
I have at least done some things right ... and
I wonder [you] have happy lives ...
I worry about [Fabian Pascal]
[I worry about Bob Badour too]
[I'm] a mom
------
Dawn, I only "met" you several days ago through your
postings dating back a few years. Therefore, try to take
this as a concerned outsider's opinion. From the summary
of your posting above, do you not see how some have come to
see you as "self-aggrandizing"?
If not, allow me to suggest one simple technique for you to
experiment with. The next time you post, read over the draft
and try to remove all occurrences of the word "I", not in a
trivial fashion of course, but rather with introspection.
Keith
Frank Hamersley
[..]
> Like the "inverted hierarchy?" Yes, that is old enough to be well
> known. The TRM difference is that a value in a column appears
> exactly once -- no matter ho many times it appears in the
> representation. A further point not made is that each value need
> appear only once in a domain. In other words, if there are many
> columns holding date values, with the same value appearing in not
> only one but many columns, the value need to be stored only once in
> TRM.
>
> This has huge significance: all values in a date domain covering
> hundreds of years require fewer than 100,000 values. Time-of-day
> precise to a second requires only 86,000 values. Given these domains,
> adding records to a system would require no new values -- the domains
> can be established before the first data arrive Social security
> numbers? There are far fewer that 10^9 possible. License plates
> numbers? What, 36^6 or 36^8 -- times 50? That's not a big gulp.
> Names? Far fewer than one might think. If domains such as these
> are enumerated before the system requiring the database is turned on,
> it could conceivably operate for years without seeing a "new" value.
> The benefit in the physical layer -- which is where all commercial
> products now have trouble when "big data" come to the party -- are
> is that space required for storing values becomes a mere tiny fraction
> of what modern systems require. 1/1,000,000 is not an unreasonable
> expectation.
But none of this is completely novel - some DBMS systems use techniques
like this for compression in the physical layer - and the beholder of
the relation is not aware of it. AKAIK Teradata apparently uses this in
limited form with fixed length strings. I suspect however they don't
perform a universal gathering (like every date within a single database)
for a domain.
[..]
Cheers, Frank.
Frank Hamersley
>>> Frank Hamersley wrote:
[..]
>> I believe he is bound by some non-disclosure agreement and since the
>> business involved has gone awry, he is stuck. So you can hardly blame
>> him except perhaps for not having the foresight to sign a NDA that had
>> no expiry date.
>
> An NDA is my understanding, too. I do not know about the expiry
> date bit though. BTW, I am a regular reader of dbdebunk.com.
Is the business defunct (in liquidation) or has it been procured by
others? I guess the NDA has survived these events and has not been
extinguished hence he must remain silent. I suspect if he (et al) has
been conned (as was suggested in the past) by the TRM proponents he must
be very irritated by this state of affairs!
[..]
>>> As an aside, I note that Mr. Pascal spends quite a lot of time telling
>>> us how smart and logical he is, ...
>
> As a regular reader of dbdebunk.com, I call bullshit!
I raise you the bull... *
>> Well, I've read at least a few hundred pages written by him and although
>> I didn't understand parts of it, I can't recall him boasting about
>> himself, not even once.
and its holy cow! * :-)
>
>>> and how all of his opponents are stupid,
>>> ignorant, and illogical.
This aspect is a common thread in dbdebunk postings.
* Therefore contrary to your opinion Gene, I think it is reasonable to
conclude that FP considers, nay promotes albeit obliquely, the belief
that he is "smart and logical" i.e. the antithesis of his tormentors,
and further that he is also "boastful" given his preparedness to operate
dbdebunk.
All that said there is nothing wrong with this posture if you are
capable of substantiating this at every turn of events. Of course it
can go pear shaped if you meet your match or you pick a loser - but that
is FP's risk and his call to place his own bets.
[..]
> Perish the thought that someone should have an agenda. I like
> FP's of supporting the RM. It is far better than these slimy Latest
> Things that keep popping up from under rocks.
For sure. My only concern is having made himself a target of the
naysayers, if he stumbles for what ever reason, it has the potential to
tarnish the RM in the minds of under informed types. A lesser evil
perhaps than allowing snake-oil to be sold in super-markets?
[..]
Cheers, Frank.
Roy Hann
news:myy9g.3567$S7....@news-server.bigpond.net.au...
I admit I don't understand why I feel moved to rebut an ad hominem (and
wholly irrelevant) attack on someone else, but here goes anyway.
> * Therefore contrary to your opinion Gene, I think it is reasonable to
> conclude that FP considers, nay promotes albeit obliquely, the belief that
> he is "smart and logical"
Well duh! Is there *anyone* who writes for public consumption who doesn't
think that of themself? That is just a truism. You'd be right to howl
down anyone who set out to do that who knew they were stupid or illogical
and did it anyway.
> i.e. the antithesis of his tormentors, and further that he is also
> "boastful" given his preparedness to operate dbdebunk.
So everyone who is moved to operate a website for the instruction and
information of others is boastful? What about those who write books?
Articles? Who lecture? Who teach? Just how much humility is enough to
qualify you venture out in public then?
> All that said there is nothing wrong with this posture if you are capable
> of substantiating this at every turn of events. Of course it can go pear
> shaped if you meet your match or you pick a loser - but that is FP's risk
> and his call to place his own bets.
There is nothing wrong with being wrong occasionally. There is everything
wrong with having it demonstrated to one but refusing ever to concede. FP
does publish errata and does from time to time revise his views. All the
examples of the latter that I can think of he openly acknowledged to be
after discussion with others.
>> Perish the thought that someone should have an agenda. I like
>> FP's of supporting the RM. It is far better than these slimy Latest
>> Things that keep popping up from under rocks.
>
> For sure. My only concern is having made himself a target of the
> naysayers, if he stumbles for what ever reason, it has the potential to
> tarnish the RM in the minds of under informed types.
Tough. That's a chance I'm certainly willing to see him take. Anyway, in
the long run, one stumble couldn't tip the balance any more than one
elegantly proved truth could.
> A lesser evil perhaps than allowing snake-oil to be sold in super-markets?
You make out that it is an issue of towering importance to our age, to be
decided by the subtlest of arguments. It's just one guy and a website for
goodness sake!
Roy
Frank Hamersley
> "Frank Hamersley" <terabit...@bigpond.com> wrote
>
> I admit I don't understand why I feel moved to rebut an ad hominem (and
> wholly irrelevant) attack on someone else, but here goes anyway.
AH attack on who exactly? Does mentioning someone by name qualify a
post as AH these days?
>> * Therefore contrary to your opinion Gene, I think it is reasonable to
>> conclude that FP considers, nay promotes albeit obliquely, the belief that
>> he is "smart and logical"
>
> Well duh! Is there *anyone* who writes for public consumption who doesn't
> think that of themself? That is just a truism. You'd be right to howl
> down anyone who set out to do that who knew they were stupid or illogical
> and did it anyway.
Doh yourself! It was GW who made the bullshit claim - I simply felt
inclined to rebutt it using your "duh" analysis but using more words.
>> i.e. the antithesis of his tormentors, and further that he is also
>> "boastful" given his preparedness to operate dbdebunk.
>
> So everyone who is moved to operate a website for the instruction and
> information of others is boastful?
Basically anyone who publicly belittles and humiliates those who they
consider less capable or wrongly informed risks being tagged. Sure its
hard to modify a stance when we are strongly convinced of it, but
sometimes a little decorum goes a long way.
> What about those who write books?
> Articles? Who lecture? Who teach? Just how much humility is enough to
> qualify you venture out in public then?
You are taking the piss!
>> All that said there is nothing wrong with this posture if you are capable
>> of substantiating this at every turn of events. Of course it can go pear
>> shaped if you meet your match or you pick a loser - but that is FP's risk
>> and his call to place his own bets.
>
> There is nothing wrong with being wrong occasionally. There is everything
> wrong with having it demonstrated to one but refusing ever to concede.
Sure - horses, water - you have to get over it though if they drown or
die of thirst.
> FP
> does publish errata and does from time to time revise his views. All the
> examples of the latter that I can think of he openly acknowledged to be
> after discussion with others.
>
>>> Perish the thought that someone should have an agenda. I like
>>> FP's of supporting the RM. It is far better than these slimy Latest
>>> Things that keep popping up from under rocks.
>> For sure. My only concern is having made himself a target of the
>> naysayers, if he stumbles for what ever reason, it has the potential to
>> tarnish the RM in the minds of under informed types.
>
> Tough. That's a chance I'm certainly willing to see him take. Anyway, in
> the long run, one stumble couldn't tip the balance any more than one
> elegantly proved truth could.
LOL! The fickleness of human behaviour seems to place much greater
weight on a persons stuff ups than achievements. I accept however, that
the long run is a leveler - even Bill Clinton is back from the dead
although Dan Quayle seems to have gone to ground with the potatoe (sic).
>> A lesser evil perhaps than allowing snake-oil to be sold in super-markets?
>
> You make out that it is an issue of towering importance to our age, to be
> decided by the subtlest of arguments. It's just one guy and a website for
> goodness sake!
What ever gave you that idea - he's just this guy you know!
Cheers, Frank.
Roy Hann
> Roy Hann wrote:
>> "Frank Hamersley" <terabit...@bigpond.com> wrote
>>
>> I admit I don't understand why I feel moved to rebut an ad hominem (and
>> wholly irrelevant) attack on someone else, but here goes anyway.
>
> AH attack on who exactly?
Oh dear, we can't even agree we're talking about Fabian Pascal? I can see
this is going to be a real short conversation.
> Does mentioning someone by name qualify a post as AH these days?
Well I sure didn't see any abstract truth being attacked, so what else is
it?
>>> i.e. the antithesis of his tormentors, and further that he is also
>>> "boastful" given his preparedness to operate dbdebunk.
>>
>> So everyone who is moved to operate a website for the instruction and
>> information of others is boastful?
>
> Basically anyone who publicly belittles and humiliates those who they
> consider less capable or wrongly informed risks being tagged.
That's not the argument you used to conclude he is boastful. You wrote "he
is also 'boastful' given his preparedness to operate dbdebunk."
> Sure its hard to modify a stance when we are strongly convinced of it, but
> sometimes a little decorum goes a long way.
>
>> What about those who write books? Articles? Who lecture? Who teach?
>> Just how much humility is enough to qualify you venture out in public
>> then?
>
> You are taking the piss!
No, I really don't think I am. I'm just drawing some reasonable conclusions
from what you actually wrote. Maybe what you wrote wasn't what you intended
to say, or was incomplete?
[Much puzzling but probably unimportant stuff snipped.]
Roy
Frank Hamersley
> "Frank Hamersley" <terabit...@bigpond.com> wrote
>> Roy Hann wrote:
>>> "Frank Hamersley" <terabit...@bigpond.com> wrote
>>>
>>> I admit I don't understand why I feel moved to rebut an ad hominem (and
>>> wholly irrelevant) attack on someone else, but here goes anyway.
>> AH attack on who exactly?
>
> Oh dear, we can't even agree we're talking about Fabian Pascal? I can see
> this is going to be a real short conversation.
OK - IMO your original thrust was so tenuous I presumed it could well
have been associating GW as the unfortunate. And short is good!
>> Does mentioning someone by name qualify a post as AH these days?
> Well I sure didn't see any abstract truth being attacked, so what else is
> it?
Hmmm - not white, must be black! You are being simplistic, lazy
perhaps, or are just a religious acolyte. Were the analyses I mentioned
likely to be true(ish) - well yes, you confirmed them in your own post
with the "duh" analysis.
>>>> i.e. the antithesis of his tormentors, and further that he is also
>>>> "boastful" given his preparedness to operate dbdebunk.
>>> So everyone who is moved to operate a website for the instruction and
>>> information of others is boastful?
>> Basically anyone who publicly belittles and humiliates those who they
>> consider less capable or wrongly informed risks being tagged.
>
> That's not the argument you used to conclude he is boastful. You wrote "he
> is also 'boastful' given his preparedness to operate dbdebunk."
Its about the tone. Its his choice to operate it as he sees fit and
mine to form my own view on the material posted and assess the tenor of
the writings in doing so.
>> Sure its hard to modify a stance when we are strongly convinced of it, but
>> sometimes a little decorum goes a long way.
>>
>>> What about those who write books? Articles? Who lecture? Who teach?
>>> Just how much humility is enough to qualify you venture out in public
>>> then?
>> You are taking the piss!
>
> No, I really don't think I am. I'm just drawing some reasonable conclusions
> from what you actually wrote.
Hypocrite!
> Maybe what you wrote wasn't what you intended
> to say, or was incomplete?
Or maybe you didn't read it as it was intended. Ce la vie!
> [Much puzzling but probably unimportant stuff snipped.]
That cool with me! Far be it from me to tickle your neurones with an
oblique reference. Anyway thats enough OT prattle from moi!
Cheers, Frank.
x
"David Cressey" <dcre...@verizon.net> wrote in message
news:tdo9g.1728$_B5.1298@trnddc01...
> Marshall,
> I'm going to start a new discussion, rather than divert this one. There
are
> several ideas that come together here,
> including some that you have brought up in the past.
> First, the idea that the index IS the database, from the article the OP
> referenced.
I also said something similar at one time, intrigued by the statement in the
patent that there are not indices.
I said the data is the index. :-)
There is also that "saying" that the log is the database and the database is
an optimized access path to it. :-)
http://weblogs.asp.net/aaguiar/archive/2004/05/04/125857.aspx
http://flipdb.com/
> Second, the idea of "content based addressing" from some of your old
posts.
We are talking about relations, aren't we ? Codd said that maybe ?
Maybe many others before Codd ?
What is the difference between a function and a relation ?
I've read about relations in a book published around 1920.
The ideeas in that book were older than that.
Why do we name columns and not rows ? :-)
> Third, the idea that the index and the table are redundant, but not
> "harmful redundancy" in the sense that term is usually used to mean.
Redundancy at the storage level is not "harmful redundancy".
Redundancy at the logical level is not "harmful redundancy".
Unmanaged redundancy is "harmful redundancy".
> Fourth, hardware associative memories.
Cellular automata ?
> Fifth the question of whether a "column store" is any more fundamental
than
> a "row store".
Why do we name columns and not rows ? :-)
What is fundamental about an implementation issue ? :-)
> Sixth the question of whether Starkey didn't use the "index IS the
database"
> concept in Interbase.
I don't know about that.
I've heard that some people claim they can read a book by reading its index
table. :-)
Maybe we should say "THE index is THE database" ?
I heard Interbase/Firebird is based on multiversion, compressed btrees and
bitmaps.
I heard Netfrastructure is based on some kind of in memory "column
storage" - a kind of segmented storage.
I heard that the i386 can organize memory as we like - array, paged,
segmented, or a mix. :-)
I heard that some file systems allow sparce files and writing/deleting any
part of a file. :-)
I heard that one could put some kind of translator on top of that. :-)
An XML one ? :-)
> All of the above has little to do with the TRM, but may merit discussion
in
> its own right.
Very little, indeed. :-)
x
"Marshall Spight" <marshal...@gmail.com> wrote in message
news:1147490980....@i40g2000cwc.googlegroups.com...
> > Marshall Spight wrote:
> > >
> > I believe he is bound by some non-disclosure agreement and since the
> > business involved has gone awry, he is stuck. So you can hardly blame
> > him except perhaps for not having the foresight to sign a NDA that had
> > no expiry date.
> I don't "blame" him, but I do note that he does not back up
> any of the claims he makes relative to "transrelational". So his claims
> are unevaluable. They might be true, or they might be false; we have
> no way of knowing.
> My policy towards unevaluable statements is to ignore them
> until they become evaluable.
Full disclosure: set theory and predicate logic. :-)
Please evaluate. :-)
> > > Michael Stonebreaker has a small company that is selling a column
> > > store; it looks quite interesting.
> > After a promising start many years ago, Stonebraker has long been a tech
> > salesman for various fads.
> Do you speak from personal experience here?
> Stonebraker started studying relational in 1973. DBMSs that he's lead
> the develompent of include Ingres, Illustra, Postgres, Cohera, and
> StreamBase. He is the recipient of the IEEE Von Neuman medal,
> which has also gone to such people as Gordon Bell, Fred Brooks,
> Don Knuth, and Alfred Aho.
So ?
> > From what I've read, most of his "opponents" haven't got the foggiest
> > of what he is talking about.
> So he keeps telling us. (Although I do not disagree with the sentiment
> that there is a lot of ignorance to go around.)
Do you know about what he is talking ?
He is talking about the fact that you should not trust him or Date or Codd
or ... :-)
Question the axioms. :-)
> > > I wonder: has he ever
> > > accomplished anything that would back up his claims? Has he ever
> > > published a proof? Published a paper in a peer-reviewed journal?
> > > Made use of any formal methods? Written any software? To my
> > > knowledge, he has published two books, one of them "Practical
> > > Issues is Database Management". Which was a fine book, although
> > > the last person I leant it to noted "you could hear the axe grinding
> > > on every page."
> >
> > Yes, he has performed a public service, mostly without gain to himself
> > whereas many of the "opponents" puff fools-gold because it is in their
> > own interest to make systems bigger and more grandiose than they need to
> > be. IT being a modern-day goldrush is full of carpet-baggers always
> > promising more than they can deliver.
> In other words, he's a consultant and public speaker. I've done
> both of those things myself.
> But what I asked was, has he ever accomplished anything?
> Can you find a mathematical proof anywhere on dbdebunk?
Mathematical proofs are foolproof. :-)
> > Peer review means less and
> > less now and is often a joke - more like a cover to protect "jobs for
> > the boys". CS credentials are usually a tawdry peerage - most of those
> > "peers" should demand refunds of their tuition fees but only a few have
> > the brains to see this and even fewer have the guts.
> In other words, everyone is stupid, so the fact that FP doesn't
> publish in peer reviewed journals is irrelevant.
> Publishing in peer reviewed journals remains the gold standard
> for academic achievement. In other words, it is how we judge
> "evidence and logic." The system is imperfect: like democracy,
> it is the worst thing possible, except for every alternative.
> It is certainly preferable to, say, putting up a website where
> you make fun of your debating opponents names, control
> every aspect of the presentation of the debate yourself, and
> refuse to publish any corrections or updates that they send
> you.
I bring to your attention another "gold" standard:
"Now while I sit in my oldest daughter and her husband's comfortable,
happy, recently-built home, I'm online to look up Phi Beta Kappa so I
understand the honor my youngest is receiving upon her graduation this
weekend." -- a mom
x
"Keith H Duggar" <dug...@alum.mit.edu> wrote in message
news:1147573812....@g10g2000cwb.googlegroups.com...
That's funny.
I've read somewhere that the Renaissance was based mainly on the accent on
"I" instead of on "The Lord".
My dictionary say: "The style of architecture and decoration originating in
Italy in the 15th century, characterized by the revival and adaptation of
ancient Roman motifs and forms, including the classical orders, and by an
emphasis on *symmetry*."
x
"Bob Badour" <bba...@pei.sympatico.ca> wrote in message
news:xEf9g.7082$A26.1...@ursa-nb00s0.nbnet.nb.ca...
> paul c wrote:
> The ultimate answer to physical independence is physical independence
> itself. ie. the greatest flexibility possible to map a logical model to
> physical stores.
You should have patented this before stating it here. :-)
> As for myself, I am in Marshall's camp on the topic of TRM -- I lack the
> information necessary to evaluate the claims.
Why ? ;-)
x
"J M Davitt" <jda...@aeneas.net> wrote in message
news:Z_99g.24008$YI5....@tornado.ohiordc.rr.com...
> Marshall Spight wrote:
> > Frank Hamersley wrote:
> >
parties?
> >>
> >>Given lots of mud gets flung as the discussion proceeds so I wondered if
> >>there was any corroborative or contrary sources.
> >
> >
> Yeah, other than the patent application - which is tortuous - there's
> little out there. To confuse the issue, other products are described
> as transrelational - Cache, IIRC, is one.
I thought patents were supposed to be complete. Full disclosure!
At least here in Novelia we have this little requirement, I think.
Discoveries, ideas, equations and algorithms are not patentable.:-)
I wonder what is patentable. :-)
> I can't
> > find
> > anything to suggest that it's anything besides a traditional column
> > store.
> > Various parties, including FP himself, have on occasion said, "oh no,
> > it's much more than that" but they don't back it up at all, so their
> > claims
> > are unevaluable.
> It *is* much more that a column store storage scheme. I don't know
> whether you've read a description of TRM, but it features (a) a
> not-so-surprising ordered collection of observed values, (b) a mildly
> clever permutation and inverse permutation index, and (c) a very clever
> "record reconstruction table."
Is this patentable ?
> > Michael Stonebreaker has a small company that is selling a column
> > store; it looks quite interesting.
> If we're talking about the C-Store he was involved with, it does feature
> a column-wise storage scheme. But, unlike TRM, values will appear in
> storage just as many times as they appear in the "logical" records
> being represented. C-Store makes extensive use of compression and,
> IIRC, is able to performs restricts and selects based on the compressed
> representations of values. Besides that, one of C-Store's big features
> is a technique for replicating a data store at different sites and
> knowing, at all sites, the most recent instant for which all sites have
> the same values.
How is compression different from "each value - stored only once" ?
From what I've heard, compression algorithms are not patentable. :-)
And yet, compression is fundamental in "computer science".
x
What about all the combinations of the domain values ?
Is not this called compression (to ignition :) ?
Frank Hamersley
> "J M Davitt" <jda...@aeneas.net> wrote in message
> news:Z_99g.24008$YI5....@tornado.ohiordc.rr.com...
>> Marshall Spight wrote:
>>> Frank Hamersley wrote:
[..]
>> It *is* much more that a column store storage scheme. I don't know
>> whether you've read a description of TRM, but it features (a) a
>> not-so-surprising ordered collection of observed values, (b) a mildly
>> clever permutation and inverse permutation index, and (c) a very clever
>> "record reconstruction table."
>
> Is this patentable ?
Apparently.
>>> Michael Stonebreaker has a small company that is selling a column
>>> store; it looks quite interesting.
>
>> If we're talking about the C-Store he was involved with, it does feature
>> a column-wise storage scheme. But, unlike TRM, values will appear in
>> storage just as many times as they appear in the "logical" records
>> being represented. C-Store makes extensive use of compression and,
>> IIRC, is able to performs restricts and selects based on the compressed
>> representations of values. Besides that, one of C-Store's big features
>> is a technique for replicating a data store at different sites and
>> knowing, at all sites, the most recent instant for which all sites have
>> the same values.
>
> How is compression different from "each value - stored only once" ?
> From what I've heard, compression algorithms are not patentable. :-)
> And yet, compression is fundamental in "computer science".
My layman's understanding is that patents are about method rather than
outcomes so if the TRM is unique in that regard its patent will stand.
FWICR you have to pay an annual fee to keep the patent alive - is that
still happening?
Cheers, Frank.
Bob Badour
> "Bob Badour" <bba...@pei.sympatico.ca> wrote in message
> news:xEf9g.7082$A26.1...@ursa-nb00s0.nbnet.nb.ca...
>
>>information necessary to evaluate the claims.
>
> Why ? ;-)
Good ol' honest ignorance. Why else?
x
Because you have not buyed the knowledge ?
Just kidding.
x
How is an invention different from a discovery ?
Bob Badour
One is the product of imagination and the other is a product of
observation. The answers to many of your questions seem rather obvious.
Keith H Duggar
just flutters anonymously around dropping sometimes veiled
sometimes dull always irrelevant micro-postings rarely (if
ever) adding any substance to the thread. Small little "x"s
spatter all around like pigeon shit on a park bench.
Keith
Paul Mansour
> This has huge significance: all values in a date domain covering
> hundreds of years require fewer than 100,000 values. Time-of-day
> precise to a second requires only 86,000 values. Given these domains,
> adding records to a system would require no new values -- the domains
> can be established before the first data arrive
But wouldn't adding a row require an index for each (indexed) column?
In the case of a date domain, the index value will in all likelyhood be
the same size as the date value it points to. Correct? Or am I missing
something? How, in the case of a date or time domain, does this have
"huge significance"?
Gene Wirchenko
<marshal...@gmail.com> wrote:
[snip]
>See if FP provides any verifiable links to any particular claim
>that could be called "evidence."
He has stated that he does not do that to avoid giving them the
link count. Some search engines prioritise by link counts. Something
will be considered more authoritative if it has more links to it. If
FP is trying to debunk a point, giving the link is counterproductive
in a way, isn't it?
[snip]
Sincerely,
Gene Wirchenko
J M Davitt
In typical physical implementations, adding a value to an indexed
column results in adding the value to whatever represents the row -
or column - as well as adding an entry to an index. As an example:
Date of Birth for driver's license holders in New York State: there
are probably, what, 10 million current licenses? I'd guess that
there are fewer than 35,000 distinct Dates of Birth in that
population. In that case, there are more than 9,000,000 "extra"
copies of the same date values in the row representations.
If dates are maintained in a domain, they should well include Date
of Issue and Expiration Date -- maybe Date Suspended, too. If
these were all indexed columns and each index was maintained "in
order" - which would certainly be the case if the index were to be
at all useful - the typical implementation's physical storage would
be littered with the same values in several places. (What are we
up to: 27 million observations of some 36,000 values?)
TRM, on the other hand, would maintain exactly one ordered set of
values for the domain and everything referencing the same date
would refer to the same value. Indices aren't really needed. Index
maintenance - the dreaded B-tree "rotate the root" operation - would
never occur. Sure, as birth dates are corrected and licenses are
renewed, the value a given record refers to would change -- but the
values remain undisturbed and there's no need for index maintenance.
There is, of course, a trade-off: the record reconstruction table
has to be maintained. That's significant work and the techniques for
doing it efficiently are, AFAIK, a closely-held secret. (Not every-
thing's covered by the patent, you know. When you apply for a patent,
you have to tell the world how you did it. Some of the most
profitable industrial secrets are not patented.)
J M Davitt
[snip]
> FWICR you have to pay an annual fee to keep the patent alive - is that
> still happening?
Absolutely. Patent holders have to pay each government from which they
expect protection. Of course, protection only means that, when you find
an infringement, you have standing in that country's courts when you seek
to recover.
And the bite isn't small. I did an "intellectual property management"
warehouse for a large and well-known American electronics manufacturer
which was generating patents by the truck load. I was absolutely amazed
at how much money was being handed over every month.
Paul Mansour
"TRM, on the other hand, would maintain exactly one ordered set of
values for the domain and everything referencing the same date
would refer to the same value. Indices aren't really needed. "
But you still have to store the "everything referencing" stuff,
correct? That's just not going to magically disappear. It seems to me
that all of the advantage comes from having a column store over a
row-store, and that the only way to get less data than a typical
column-store would then be to compress a column with, say, run-length
encoding. In other words, maintaining an ordered set of dates and
having say, five columns referencing that ordered set does nothing for
you over just maintaining five columns of dates if the you use a 32 bit
int for storing the date and 32 bit int for storing the thing that
references the date column. If the column in question were on a wide
character domain then it would save space to have a integer to
reference the column (of just make a new table and have foreign keys
point to the values). So I guess my real question is what does TRM
offer over a conventional column-store, which has been around for well
over 30 years? I've read C.J. Date's small section in the 8th edition
of text book on it, but I don't get it.
Bob Badour
Paul, you seem to have missed the point that to keep the data ordered,
you would have to store the date in the heap and in the index using a
row store. With the column store, that is not required. So while the
reference pointer is on the order of the size of the date, it only
appears once.
With the index, the date appears at least twice. Once in the heap, and
once in the leaf pages of the index. It may also appear elsewhere in the
index.
Paul Mansour
"Paul, you seem to have missed the point ...."
Bob, I'm not disputing the improvement of TRM over a trad row store.
My question again is what does TRM have over a tradition column store?
Paul Mansour
wouldn't be too much trouble, would you or J M Davitt give me a small
example? Or maybe you can correct mine:
Condider, as in Davitt's example above, a Date domain D that enumerates
all the possible dates (leave it a 5 days for the example):
D <-> 2006-01-01 2006-01-02 2006-01-03 2006-01-04 2006-01-05
and a single relation with two columns Date-of-Birth (DOB) and
Expiration date (EXP) with indices pointing to D:
ID <-> 1 2 3 4
DOB <-> 3 2 5 4
EXP <-> 1 5 5 3
Are the columns stored like this in TRM? If not, how are they stored?
If they are stored like this, what advantage is there to pointing to D
instead of just storing actual date values in DOB and EXP?
I realize I must sound like an idiot, but I would appreciate any help
or explanation. Thanks.
Frank Hamersley
> Paul Mansour wrote:
>> J M Davitt wrote regarding inverted indexes:
>>
>>> This has huge significance: all values in a date domain covering
>>> hundreds of years require fewer than 100,000 values. Time-of-day
>>> precise to a second requires only 86,000 values. Given these domains,
>>> adding records to a system would require no new values -- the domains
>>> can be established before the first data arrive
>>
> TRM, on the other hand, would maintain exactly one ordered set of
> values for the domain and everything referencing the same date
> would refer to the same value. Indices aren't really needed. Index
> maintenance - the dreaded B-tree "rotate the root" operation - would
> never occur. Sure, as birth dates are corrected and licenses are
> renewed, the value a given record refers to would change -- but the
> values remain undisturbed and there's no need for index maintenance.
That seems OK - without presupposing the physical form of the mapping
structures, it is akin to the MS Access et al surrogate key
decomposition into "domain" tables. This is acceptable at the physical
level but quite theoretically suspect at the logical level where it is
commonly seen.
> There is, of course, a trade-off: the record reconstruction table
> has to be maintained. That's significant work and the techniques for
> doing it efficiently are, AFAIK, a closely-held secret. (Not every-
> thing's covered by the patent, you know. When you apply for a patent,
> you have to tell the world how you did it. Some of the most
> profitable industrial secrets are not patented.)
No slight intended, but they wouldn't be secrets if they were patented :-).
Does the TRM patent (or related writings) hint as to how this record
reconstruction (RR) occurs in when performing queries framed at the
logical level?
Thinking aloud I am guessing for some queries eg. "this_date =
that_date" you don't even have to refer to the "domain" elements because
equivalence implies an identical physical "pointer" (a term of
convenience, perhaps there is a better choice?). Similarly "this_date >
that_date" can be abridged if the domain physical pointers themselves
are ordered. Therefore if all the supported operations can be performed
at the physical level then RR can be deferred to the last step of
materialising the outcome.
One question that pops up though - how do you represent a relation where
no single attribute is a candidate key. eg. [Competitor;Date;Score]?
Cheers, Frank
x
Is Math a discovery or an invention ?
>The answers to many of your questions seem rather obvious.
What questions ?
You mean obvious like a nail ?
Obvious:
1. easily seen, recognized, or understood;
open to view or knowledge; evident.
2. lacking in subtlety.
3. Obs. being or standing in the way.
x
Thank you for your kind words.
x
"x" <x...@not-exists.org> wrote in message news:e4bs29$r9c$1...@emma.aioe.org...
I don't know if this is dull or veiled for you.
To clarify, I like pigeons. They are the symbol of peace.
It is said that if a pigeon shit land on you (from all the people it could
drop on), you are lucky. :-)
x
"Paul Mansour" <pa...@carlislegroup.com> wrote in message
news:1147749726....@j33g2000cwa.googlegroups.com...
> Bob, I'm sure I'm missing something, but I can't see it. If it
> wouldn't be too much trouble, would you or J M Davitt give me a small
> example? Or maybe you can correct mine:
>
> Condider, as in Davitt's example above, a Date domain D that enumerates
> all the possible dates (leave it a 5 days for the example):
>
> D <-> 2006-01-01 2006-01-02 2006-01-03 2006-01-04 2006-01-05
>
> and a single relation with two columns Date-of-Birth (DOB) and
> Expiration date (EXP) with indices pointing to D:
>
> ID <-> 1 2 3 4
> DOB <-> 3 2 5 4
> EXP <-> 1 5 5 3
> Are the columns stored like this in TRM? If not, how are they stored?
> If they are stored like this, what advantage is there to pointing to D
> instead of just storing actual date values in DOB and EXP?
Have you read the patent ?
How you store the relation is the whole problem.
The patent is supposed to give you all the details you cannot figure out by
yourself (full disclosure).
x
> Paul Mansour wrote:
> > J M Davitt wrote regarding inverted indexes:
> TRM, on the other hand, would maintain exactly one ordered set of
> values for the domain and everything referencing the same date
> would refer to the same value. Indices aren't really needed. Index
> maintenance - the dreaded B-tree "rotate the root" operation - would
> never occur. Sure, as birth dates are corrected and licenses are
> renewed, the value a given record refers to would change -- but the
> values remain undisturbed and there's no need for index maintenance.
This is just adding one level of indirection.
Why is this something new ?
> There is, of course, a trade-off: the record reconstruction table
> has to be maintained. That's significant work and the techniques for
> doing it efficiently are, AFAIK, a closely-held secret. (Not every-
> thing's covered by the patent, you know. When you apply for a patent,
> you have to tell the world how you did it. Some of the most
> profitable industrial secrets are not patented.)
This is the subject of the patent and cannot be a secret.
J M Davitt
Date, at least, has written descriptions and examples of how this is
done.
> Thinking aloud I am guessing for some queries eg. "this_date =
> that_date" you don't even have to refer to the "domain" elements because
> equivalence implies an identical physical "pointer" (a term of
> convenience, perhaps there is a better choice?). Similarly "this_date >
> that_date" can be abridged if the domain physical pointers themselves
> are ordered. Therefore if all the supported operations can be performed
> at the physical level then RR can be deferred to the last step of
> materialising the outcome.
Actually, if I understand the operations you're imagining, restrictions
and selections (I'm distinguishing between these, here) both are more
efficiently done in the "domain elements" rather than among the
"physical pointers."
But, you know,...
> One question that pops up though - how do you represent a relation where
> no single attribute is a candidate key. eg. [Competitor;Date;Score]?
...I've probably taken this as far as I can without looking something
up. There are a few documents available that describe TRM; they are a
better place to look for the answers you seek than this forum is.
David Cressey
"Bob Badour" <bba...@pei.sympatico.ca> wrote in message
news:jm1ag.8005$A26.2...@ursa-nb00s0.nbnet.nb.ca...
So, was "America" a discovery or an invention?
Frank Hamersley
I understand the amount escalates, presumably as you are expected to be
increasing your royalty income over time (if your invention is any good).
If you don't pay I guess its open slather and the patent lapses. I
presume there are no squatters rights in that event either so the
technique becomes public domain from that point?
Anyway - back to the real question - is the TRM patent still actively
being funded?
Cheers, Frank.
x
Why don't you search the patent database ?
dawn
> Dawn, I only "met" you several days ago through your
> postings dating back a few years. Therefore, try to take
> this as a concerned outsider's opinion. From the summary
> of your posting above, do you not see how some have come to
> see you as "self-aggrandizing"?
Hi Keith --
[After reading your last statement, I see that everything I wrote here
does not take into account your suggestion, but the topic is also not
database theory. I will see what I can do when discussing or asking
questions in the future, however. I do not know how to express an
opinion as if were the truth, as others do, and, in fact, have made it
a policy to try to ensure the reader can tell the difference, but I
will research that style of writing before I post again. My response
to your above statement is below.]
Because in my real life I never hear comments of this nature, because
the person who has said that of me says that about everyone who
approaches the topic from a standpoint other than his (the RM), and
because there is only one person who does say that, I was ignoring it
as a reflection of the author and not me. But if you are thinking this
sounds at all accurate, I certainly would want to rectify it.
I will try harder to be clear, as I have been in the past, that I
consider my own opinions flawed, that I am here seeking answers to
database theory questions rather than thinking I have all the answers,
that I know I am not yet even consistent in my opinions (typing being
an example where I have not made my own thinking clear), but that my
questions are the legitimate questions of someone looking to
understand.
My questions are also not within relational theory, but in the
application of relations as an exclusive way of modeling database data.
Because I am not writing database theory, but seeking to understand
why what I know of database theory does not align with what I have seen
with database practice, I give examples from practice when expressing
opinions and asking questions. Then it should be clear to the reader
that I am not stating an absolute, but rendering an opinion. There are
many opinions stated as facts in this forum and I have tried not to
contribute to that.
> If not, allow me to suggest one simple technique for you to
> experiment with. The next time you post, read over the draft
> and try to remove all occurrences of the word "I", not in a
> trivial fashion of course, but rather with introspection.
Stating opinions as absolutes is arrogant and misleading. Does that
sentence work? I trust you can see the humor in using that statement
to give this a spin. With my own style, I would have said "It seems to
me that stating opinions as absolutes is arrogant and misleading." BB
even makes his statements about me, never having met me, as if they
were the truth, rather than simply his opinion, flawed as it may be. I
recognize there are some who prefer that style of writing and will give
it a spin when I engage the subject matter again. If I find the
technique too hard to master or am too offended by it myself, I will
simply chat with those who are willing to engage in the dialogue while
letting me be me.
Thanks for your suggestion, as that is much more enlightening and
helpful than simply slamming me. Cheers! --dawn
Bob Badour
Paul, you left out an important element of J M's argument, and this
seems central to your misunderstanding what he said.
{ { 1, 2006-01-03, 2006-01-01 }, { 2, 2006-01-02, 2006-01-05 }
, { 3, 2006-01-05, 2006-01-05 }, { 4, 2006-01-04, 2006-01-03 }
}
In a traditional row store, the four rows above appear in a heap in some
essentially random order. To maintain the dates and the ids in order,
they will also appear in three indexes.
An index for four rows will occupy only a single page on disk. An index
for a billion rows will occupy many more pages.
The size of each of those indexes is O(log(N)) and the size of the heap
is O(N) where N is the cardinality. The size of the column store is O(1)
for the domain values and O(N) for the pointers.
Your whole point is the O(N)'s are the same.
J M's point is O(log(N)) >> O(1) for large N.
However, I just received an email from Fabian where he points out that
the "TRM is not at the physical level, so it's not about storage. It is
a layer between RM and storage, and can be physically implemented in any
number of ways, one of which is columnar."
While I do not have sufficient information to evaluate the claims of the
TRM, hopefully I have cleared up what you were missing from J M's argument.
Bob Badour
> "Bob Badour" <bba...@pei.sympatico.ca> wrote in message
> news:jm1ag.8005$A26.2...@ursa-nb00s0.nbnet.nb.ca...
>
>>x wrote:
>>
>>>"Frank Hamersley" <terabit...@bigpond.com> wrote in message
>>>news:qq_9g.4396$S7....@news-server.bigpond.net.au...
>>>
>>>>x wrote:
>>>>
>>>>>"J M Davitt" <jda...@aeneas.net> wrote in message
>>>>>news:Z_99g.24008$YI5....@tornado.ohiordc.rr.com...
>>>>>
>>>>>>Marshall Spight wrote:
>>>>>>
>>>>>>>Frank Hamersley wrote:
>>>
>>>>My layman's understanding is that patents are about method rather than
>>>>outcomes so if the TRM is unique in that regard its patent will stand.
>>>
>>>How is an invention different from a discovery ?
>
>>One is the product of imagination and the other is a product of
>>observation.
>
> Is Math a discovery or an invention ?
Math is an invention.
>>The answers to many of your questions seem rather obvious.
>
> What questions ?
> You mean obvious like a nail ?
Plonk.
Bob Badour
The continent was a discovery. The values and systems of government were
inventions.
Kenneth Downs
> Because I am not writing database theory, but seeking to understand
> why what I know of database theory does not align with what I have seen
> with database practice, I give examples from practice when expressing
> opinions and asking questions.
I've been pondering for some days starting a thread on the general placement
of db theory (in general and in specific) to app development. I've tried
it before but usually nobody bites. But I will probably try again.
--
Kenneth Downs
Secure Data Software, Inc.
(Ken)nneth@(Sec)ure(Dat)a(.com)
dawn
> "Frank Hamersley" <terabit...@bigpond.com> wrote in message
> news:qq_9g.4396$S7....@news-server.bigpond.net.au...
> > x wrote:
> > > "J M Davitt" <jda...@aeneas.net> wrote in message
> > > news:Z_99g.24008$YI5....@tornado.ohiordc.rr.com...
> > >> Marshall Spight wrote:
> > >>> Frank Hamersley wrote:
>
>
> > My layman's understanding is that patents are about method rather than
> > outcomes so if the TRM is unique in that regard its patent will stand.
>
> How is an invention different from a discovery ?
In theory, invention is a creative act of the mind, while discovery is
a scientific or sensory act. It is hard to draw a clean line between
them, however. Do we invent mathematics or discover it? --dawn
dawn
> dawn wrote:
>
> > Because I am not writing database theory, but seeking to understand
> > why what I know of database theory does not align with what I have seen
> > with database practice, I give examples from practice when expressing
> > opinions and asking questions.
>
> I've been pondering for some days starting a thread on the general placement
> of db theory (in general and in specific) to app development. I've tried
> it before but usually nobody bites. But I will probably try again.
Go for it. --dawn
Keith H Duggar
> > From the summary of your posting above, do you not see
> > how some have come to see you as "self-aggrandizing"?
> >
> > to experiment with. The next time you post, read over
> > the draft and try to remove all occurrences of the word
> > "I", not in a trivial fashion of course, but rather with
> > introspection.
dawn wrote:
> [I see that everything I wrote here does not take into
> account your suggestion ... I do not know how to express
> an opinion as if were the truth, as others do, and, in
> fact, have made it a policy to try to ensure the reader
> can tell the difference, but I will research that STYLE
> of writing before I post again.
> ...
> Stating opinions as absolutes is arrogant and misleading.
> Does that sentence work? I trust you can see the humor in
> using that statement to give this a spin. With my own
> STYLE, I would have said "It seems to me that stating
> opinions as absolutes is arrogant and misleading." BB
> even makes his statements about me, never having met me,
> as if they were the truth, rather than simply his opinion,
> flawed as it may be. I recognize there are some who
> prefer that STYLE of writing and will give it a spin when
> I engage the subject matter again. If I find the technique
> too hard to master or am too OFFENDED by it myself, I will
> simply chat with those who are willing to engage in the
> dialogue while letting me be me.
[emphasis added - KHD]
Dear Dawn,
Noting your repeated use of STYLE and the general content of
your message I believe you have missed my point. The
suggestion was not to follow a particular grammatical
style. Transforming sentences from one parse tree to another
falls under "trivial" editing "fashion" warned against. It
is not a prohibition of "I" or any other syntax. Repeated
use of "I" is simply sometimes a _signal_.
The _key_ to the suggestion was _introspection_. In other
words I'm suggesting one (as in one of many) simple mechanic
that helps to encourage a style of _thinking_ (not writing).
Differences in thinking of course will translate into
differences in communication. Thus since the suggestion was
for a method of introspection when communicating (and not a
grammatical style) you cannot be OFFENDED by the method
except to the extent that your thoughts offend yourself.
For example, you might ask yourself "What is the purpose of
these statements? What is my motivation? What is my goal?"
1) I took several years of stones ...
I recognize I'm a heretic ...
2) I also know what happens to heretics ...
3) I might even retract ...
I have no disagreement with the mathematics ...
I disagree with some [standard] choices ...
I have not ... prove[n] my hunches ...
4) I have enough experience with budgets ...
5) I have formed opinions ...
I write about that which interests me ...
I continue my search, in spite of [you]
6) I sit in my oldest daughter and her husband's
comfortable, happy, recently-built home
7) I'm online to look up Phi Beta Kappa ...
I understand the honor my youngest is receiving ...
her graduation this weekend ...
8) I have at least done some things right ... and
9) given his approach ... [FP] seems a sad sort
I wonder if [you] have happy lives ...
I worry about [Fabian Pascal]
I worry about Bob Badour too]
0) -- a mom
For my part I found them to be
1) appeal to pity, appeal to authority (in this case the
common knowledge that many heretics were right, Galileo
for example)
2) faulty generalization, guilt by association, poisoning
the well
In other words you want readers to feel sorry for you, to
believe that you are somehow special and revolutionary
and anyone who opposes you is a dogmatist, an inquisitor.
3) proof by assertion, it remains to be seen whether you can
accept reasoned arguments and change your beliefs.
4) appeal to authority, non-sequitur
5) meaningless or obvious statements, self-centered (why
should we care what interests you?)
Here's where self-aggrandizement, arrogance, and especially
price became blatant in my view.
6) PRIDE. Recently-built? I couldn't help but wonder how
much of nature they consumed to build this new home. Are
they greedy and gluttonous? How many of them drive FUVs
(yes the F is intentional)? Moreover, who bloody well
cares how fortunate you are? And how is this relevant.
7) PRIDE, non-sequitur. Her wonderful special family is
entirely irrelevant.
8) PRIDE, arrogance, non-sequitur. Perhaps her daughters
turned out well in spite of, rather than because of her.
And either way it is irrelevant. And did I mention "Who
cares?"
9) arrogance, ad hominem.
0) unmitigated price and arrogance.
At least those were my immediate impressions which of course
depended on many factors. Furthermore, what any particular
reader infers (in this case me) is not the important
point. You cannot control how _every_ reader perceives your
communication. What is important is that you _consider_ with
introspection what, how, and why you are communicating.
> Thanks for your suggestion, as that is much more
> enlightening and helpful than simply slamming me.
Hopefully this continues to be the case. Please understand
that I am not trying to slam _you_. As a mother enjoying the
graduation of her child, some amount of pride is certainly
expected. Thus I have no trouble understanding the very
human emotions reflected in your writing. So it truly is not
my goal to attack you. Rather it is to point out what I
perceived as deficiencies in the thinking behind a
particular communication you wrote.
-- Keith --
PS. Congratulations to your family. And to anyone out there
who happens to drive an FUV, please consider a more fuel
efficient lifestyle. Our culture of consumption is a very
bad thing.
Bob Badour
Check out the threads around here from three years ago. Nothing remains
to be seen on that score.
> -- Keith --
>
> PS. Congratulations to your family. And to anyone out there
> who happens to drive an FUV, please consider a more fuel
> efficient lifestyle. Our culture of consumption is a very
> bad thing.
I disagree. The sooner consumption forces us to find cheap domestic
energy sources the better.
Marshall
> [...]
Keith,
Two thoughts:
1) "I" statements are better than "you" statements.
2) Do you have anything that falls under the general topic
of database theory that you would care to discuss?
Marshall
JOG
> 1) appeal to pity, appeal to authority (in this case the
> common knowledge that many heretics were right, Galileo
> for example)
Alfred Wegener is a more recent example, academically crucified for his
theory of continental drift, which was widely accepted soon after his
death.
I feel this thread has gone somewhat off topic, which is a shame
because the progression of the TRM is interesting discussion.
Dan
values for the domain and everything referencing the same date
would refer to the same value. Indices aren't really needed. Index
maintenance - the dreaded B-tree "rotate the root" operation - would
never occur. Sure, as birth dates are corrected and licenses are
renewed, the value a given record refers to would change -- but the
values remain undisturbed and there's no need for index maintenance. "
To work well, your statements would seem to imply the following:
1) For any declared domain, we must enumerate all possible values in
advance.
Suppose we were to do this with a fixed character string domain of say?
CHAR(1000).
If I understand what you are saying, then for that one fixed width
domain, assuming 36 symbols for upper and lower case, etc., then we
would need 2.006784635490209538403217829194e+1556 values stored, either
in memory or on disk. Granted, all domains of smaller symbol list
widths could leverage the use of the largest set of possible values,
but then there seems a price to pay for that little flag fixed width
value consisting of 1 symbolic character that needs to pull its value
from such a large store.
When considering variable length domains, the problem gets a little
larger and more complex. And how about domains where ordering
rules/comparison operators are intended to be different than say,
perhaps the standard ANSI collation? A same domain with a different
ordering might have to be "redundantly" maintained?
- Dan
Paul Mansour
Bob, thanks for the reply. I guess I misundestood J M -- he seemed to
be saying that the TRM has some signficant space saving over a
tradition column store, when, as you explained, his point was that a
column store has significant space savings over a row store. I still
don't see what the TRM has over a trad column store, but It seems I'm
not the only one.
vldm10
Frank Hamersley wrote:
> I just stumbled over this exchange from Oct 2005 on the TRM...
>
> http://www.dbms2.com/2005/10/10/17/
>
> It was authored by a "Curt Monash" - is he known to any CDT'ers and/or
> credible?
>
> Is there any veracity in _any_ of the claims made by _any_ of the parties?
>
> Given lots of mud gets flung as the discussion proceeds so I wondered if
> there was any corroborative or contrary sources.
>
> Cheers, Frank.
You can try my solution. It effectively shows how to decompose
arbitrary relation into its binary relations. It is on my web site:
www.dbdesign10.com
Yesterday I added two new small chapters (4 and 5) that are about
"theoretical" side of this problem. I believe that my solution is
simple, correct and more general than TransRelatinal Model (TM).
Vladimir Odrljin
Frank Hamersley
> Keith H Duggar wrote:
>> PS. Congratulations to your family. And to anyone out there
>> who happens to drive an FUV, please consider a more fuel
>> efficient lifestyle. Our culture of consumption is a very
>> bad thing.
>
> I disagree. The sooner consumption forces us to find cheap domestic
> energy sources the better.
<OP_OffTopic_Licence :-)>
Not that I expect Bob to read this but yet again I call his "disagree"
(sic) and raise him to extinction!
I assert his statement is evidence of a "Jack man" attitude. For those
not familiar with the term it identifies a selfish individual and is
drawn from a colloquial phrase "Bugger you Jack, I'm OK!".
This is particularly evident in the use of the term "domestic" (ie.
stuff the rest of the world) and the inference that the consumption
itself does not need to stop regardless of its cost or value - only the
material expense of the consuming needs to be fixed.
Personally I don't rate the "don't sweat it, we will always find a way
to fix the problem when it finally arises" attitude as a somewhat
unreliable survival trait.
</OP_OffTopic_Licence :-)>
Frank Hamersley
Was trying to avoid that eventuality ... but have succumbed and sunk 30
mins getting to grips with a rather clunky but quite rich resource!
For the earliest patent it appears the 4 year window has been covered
which means the next payment (8 year window) is due sometime in 2007.
The later 2 patents were awarded in 2003 and are due for a payment
(approx USD500 each) in 2007 as well.
Cheers, Frank.
J M Davitt
> "TRM, on the other hand, would maintain exactly one ordered set of
> values for the domain and everything referencing the same date
> would refer to the same value. Indices aren't really needed. Index
> maintenance - the dreaded B-tree "rotate the root" operation - would
> never occur. Sure, as birth dates are corrected and licenses are
> renewed, the value a given record refers to would change -- but the
> values remain undisturbed and there's no need for index maintenance. "
>
> To work well, your statements would seem to imply the following:
>
> 1) For any declared domain, we must enumerate all possible values in
> advance.
Not necessarily. The examples cited are rather small domains - a couple
centuries worth of dates, states, words used in names or addresses - and
the supported system would undoubtedly benefit by having these known
ahead of time. But exhaustive enumeration isn't the only way one can
specify a domain.
> Suppose we were to do this with a fixed character string domain of say?
> CHAR(1000).
>
> If I understand what you are saying, then for that one fixed width
> domain, assuming 36 symbols for upper and lower case, etc., then we
> would need 2.006784635490209538403217829194e+1556 values stored, either
> in memory or on disk.
Well, you could do it that way if you wanted to -- but there are better
ways to handle this. And CHAR (1000) is a trite example: how about
JPEGs and MP3s and *.DOCs and XML? One needs to think in terms of
"billions and billions" of bits when considering modern data stores.
Granted, all domains of smaller symbol list
> widths could leverage the use of the largest set of possible values,
> but then there seems a price to pay for that little flag fixed width
> value consisting of 1 symbolic character that needs to pull its value
> from such a large store.
>
> When considering variable length domains, the problem gets a little
> larger and more complex.
Much larger, not more complex.
And how about domains where ordering
> rules/comparison operators are intended to be different than say,
> perhaps the standard ANSI collation?
Every domain should provide an operator that allows users to distinguish
different values. If ordering is necessary: supply it. (And yes, TRM
requires ordering; whether it is the same operator exposed to users is
a different question, but I don't see any need for them to differ.)
A same domain with a different
> ordering might have to be "redundantly" maintained?
I don't understand this part.
>
> - Dan
>
Bob Badour
I found your web article confusing. Are you defining a logical data
model? Or a conceptual model?
If it is a conceptual model, correctness is not particularly relevant.
If it was supposed to be a logical data model, it wasn't simple.
Keith H Duggar
> 1) "I" statements are better than "you" statements.
At face value this statement is impossible to evaluate
without context. Reading between the lines I take your
point and have no substantial quarrel with it.
> 2) Do you have anything that falls under the general
> topic of database theory that you would care to discuss?
Yes I believe I do. Unfortunately I do not yet possess the
knowledge and vocabulary needed to properly formulate the
questions here. Basically I want to know if and how the RM
can be leveraged to improve and simplify numerical
simulations. Typically I implement OO/network models using
C++. In the recent past slogging through tedious navigational
code (pointers, iterators, visitors, etc) led me to wonder
if there was a better way. Hence I wandered over to c.d.t.
looking for education and inspiration.
Dan
Thanks for the response. I have a couple of follow-on questions.
1. "Not necessarily. The examples cited are rather small domains - a
couple
centuries worth of dates, states, words used in names or addresses -
and
the supported system would undoubtedly benefit by having these known
ahead of time. But exhaustive enumeration isn't the only way one can
specify a domain."
Ok. Though understanding the alternative approaches to contrast to
exhaustive enumeration would be useful, especially in light of the
previous statement, "the values remain undisturbed and there's no need
for index maintenance." Without enumeration, how does one claim that a
set of values remain undisturbed?
2. "And CHAR (1000) is a trite example. How about JPEGs and MP3s and
*.DOCs and XML."
Let's stick with the trite example and demonstrate the effectiveness of
the trite example first. Isn't this just a restatement of the age old
claim that serialization of everything is the solution to everything.
Honestly, Claude Shannon covered this effectively and from an entirely
theoretical perspective in the 60's.
3. "Every domain should provide an operator that allows users to
distinguish
different values. If ordering is necessary: supply it. (And yes, TRM
requires ordering; whether it is the same operator exposed to users is
a different question, but I don't see any need for them to differ.) "
This is an interesting and open question for me. It inevitably leads
us back to whether domain and type are really the same thing. I tend
to think not, and admittedly it runs against Date and many proponents
here. An equivalent or equal domain can have different ordering
operators in my opinion and considering symbolic values independent of
some operators provides a greater degree of independence. But I
digress.
You wrote earlier that "TRM, on the other hand, would maintain exactly
one ordered set of values for the domain." If we refer to that same
ordered set but want an alternative ordering, we run into the same
problems that lack of physical independence provides in network,
hierarchical, and odbms data models, except that the problem might
become more extreme because of the massiveness of potential domains.
Don't get me wrong, I am not knocking the approach, but as with
everything, it appears that this provides the set of
benefit/shortcoming trade-offs that are dependent on context and on
application that we've experienced before. It might not be a great
"general" solution.
- Dan
x
> x wrote:
>
> > "Bob Badour" <bba...@pei.sympatico.ca> wrote in message
> > news:jm1ag.8005$A26.2...@ursa-nb00s0.nbnet.nb.ca...
> >
> >>>How is an invention different from a discovery ?
> >
> >>One is the product of imagination and the other is a product of
> >>observation.
> >
> > Is Math a discovery or an invention ?
> Math is an invention.
We should be allowed to patent theories and numbers then.
x
"dawn" <dawnwo...@gmail.com> wrote in message
news:1147803880.9...@i39g2000cwa.googlegroups.com...
invention [1300-50; ME < L inventio *discovery* = inven (ire) :-)
----------------^
in-vent (in vent') v.t. <-vent-ed, -vent-ing>
1. to originate as a product of one's own
ingenuity, experimentation, or
contrivance: to invent a better mousetrap.
2. to produce or create with the
imagination: to invent a story.
3. to make up or fabricate (something
fictitious or false): to invent excuses.
--------> 4. Archaic. to come upon; find.
[1425-75; late ME invented (ptp.) discovered < L
inventus, ptp. of invenire to encounter, come
upon, find = in- IN -2 + venire to COME]
dis-cov-er (di skuv'uhr) v.t. <-ered, -er-ing>
1. to gain sight or knowledge of (something
previously unseen or unknown).
2. to notice or realize.
-------> 3. Archaic. to make known; reveal; disclose.
[1250-1300; ME < AF discoverir, OF descovrir < LL
discooperire. See DIS -1, COVER]
Bob Badour
I disagree. If one has the same sets of representations but the less
than operation is defined differently, one has a different type
entirely. However, because the types have the same sets of
representations, the type caste should have all of the cost of a no-op.
x
"Keith H Duggar" <dug...@alum.mit.edu> wrote in message
news:1147842916.3...@38g2000cwa.googlegroups.com...
> Marshall wrote:
> > 2) Do you have anything that falls under the general
> > topic of database theory that you would care to discuss?
> Yes I believe I do. Unfortunately I do not yet possess the
> knowledge and vocabulary needed to properly formulate the
> questions here. Basically I want to know if and how the RM
> can be leveraged to improve and simplify numerical
> simulations. Typically I implement OO/network models using
> C++.
I don't have much experience with numerical simulations.
What part of them should not require many loops in your opinion ?
Matrix calculations ?
> In the recent past slogging through tedious navigational
> code (pointers, iterators, visitors, etc) led me to wonder
> if there was a better way. Hence I wandered over to c.d.t.
> looking for education and inspiration.
Some differences between network and relational databases are (in my opinion
:):
- the network ones seems ad hoc ; the relational ones have a foundation in
math right from the start
- the network ones exists ; the relational ones are incomplete and expensive
- the network ones are based on sequencial access ; the relational ones are
based on random access
The OO ones are supposed to be a cross of network and relational, easyer to
use than both of them.
The "semistructured" ones are supposed to cure the flaws in the OO ones.
I don't see how they do that given the "fact" that "we" have not been able
to implement the relational ones in the first place.
Best luck,
Mr. X :-)
J M Davitt
> Hi,
>
> Thanks for the response. I have a couple of follow-on questions.
>
> 1. "Not necessarily. The examples cited are rather small domains - a
> couple
> centuries worth of dates, states, words used in names or addresses -
> and
> the supported system would undoubtedly benefit by having these known
> ahead of time. But exhaustive enumeration isn't the only way one can
> specify a domain."
>
> Ok. Though understanding the alternative approaches to contrast to
> exhaustive enumeration would be useful, especially in light of the
> previous statement, "the values remain undisturbed and there's no need
> for index maintenance." Without enumeration, how does one claim that a
> set of values remain undisturbed?
In the illustration from which you plucked this fragment, the domains
were enumerated.
> 2. "And CHAR (1000) is a trite example. How about JPEGs and MP3s and
> *.DOCs and XML."
>
> Let's stick with the trite example and demonstrate the effectiveness of
> the trite example first. Isn't this just a restatement of the age old
> claim that serialization of everything is the solution to everything.
Your turn: elaborate, please!
> Honestly, Claude Shannon covered this effectively and from an entirely
> theoretical perspective in the 60's.
And he said, ... What?
> 3. "Every domain should provide an operator that allows users to
> distinguish
> different values. If ordering is necessary: supply it. (And yes, TRM
> requires ordering; whether it is the same operator exposed to users is
> a different question, but I don't see any need for them to differ.) "
>
> This is an interesting and open question for me. It inevitably leads
> us back to whether domain and type are really the same thing. I tend
> to think not, and admittedly it runs against Date and many proponents
> here.
Sounds like an opportunity for a new topic.
An equivalent or equal domain can have different ordering
> operators in my opinion and considering symbolic values independent of
> some operators provides a greater degree of independence.
How do "equivalent or equal" and "different ordering" work together?
> But I
> digress.
Yes, and...
> You wrote earlier that "TRM, on the other hand, would maintain exactly
> one ordered set of values for the domain." If we refer to that same
> ordered set but want an alternative ordering, we run into the same
> problems that lack of physical independence provides in network,
> hierarchical, and odbms data models, except that the problem might
> become more extreme because of the massiveness of potential domains.
...you seem to be trying to make some point about something that should
simultaneously be "the same" and "different" and I'm not sure what
that is.
> Don't get me wrong, I am not knocking the approach, but as with
> everything, it appears that this provides the set of
> benefit/shortcoming trade-offs that are dependent on context and on
> application that we've experienced before. It might not be a great
> "general" solution.
For some things, it might not be a great general solution. One
difficulty with many computerizable solutions is that they're discrete
representations of continuous systems. Obviously, there are going to
be cracks; I think you touched on this earlier with the reference to
symbolic values.
David Cressey
"dawn" <dawnwo...@gmail.com> wrote in message
> I recognize I'm a heretic among relational theorists, and I also know
> what happens to heretics.
When a heretic turns out to have been right, we remember. When a heretic
turns out to have been wrong, we forget.
This creates a skewed perspective on the inherent value of heresy.
My response to your writings has NOT been keyed on whether or not it was
heretical. It's been keyed on whether or not it squares with my own
experience. You say that your experiences have led you to some hunches that
you cannot prove. However, your experiences, as you describe them here,
have run afoul of some cost consuming snags that I haven't run afoul of.
In all the years I have been reading what you write, you have yet to
pinpoint a reason why I should be dissatisfied with what I think I have
learned. That's why I remain skeptical. It has nothing to do with either
group think or heresy.
dawn
Fair enough, David. As I've mentioned, I'm not here to convince
cdt'ers other than as needed to push myself further, learning
vocabulary and concepts and refining my own opinions on related topics.
On the other hand, I started the blog with an interest in influencing
others and know that I'm not yet at the point (and might never be)
where you might have any take-aways from it.
Hearing your experiences and other differing experiences has been
helpful. For example, I have thought many times about what made
datatrieve (as little as I can recall of it) such a nice tool. I was
going to ask you again to recall what might have been better about it
than SQL, for example, although that should likely be in a new thread.
Thanks. --dawn
Marshall
>
> Yes I believe I do. Unfortunately I do not yet possess the
> knowledge and vocabulary needed to properly formulate the
> questions here. Basically I want to know if and how the RM
> can be leveraged to improve and simplify numerical
> simulations. Typically I implement OO/network models using
> C++. In the recent past slogging through tedious navigational
> code (pointers, iterators, visitors, etc) led me to wonder
> if there was a better way. Hence I wandered over to c.d.t.
> looking for education and inspiration.
Ah, what an excellent topic!
I believe the answer to your question is yes. The most immediate
places the RM can help is physical independence and set-at-a-time
processing. I think the example of the J programming language
is quite instructive. Although it is more focused on ordered collection
rather than relations, it nonetheless has a solid mathematical
foundation
and no explicit loops.
Physical independence comes in when one considers how much work
many C++ programmers have to do to, for example, lay out their
data in a way that will satisfy a graphics coprocessor, or enable
them to use SIMD instructions. It would be better if this was
abstracted
from the code.
To achieve the big wins, though, we need a programming language that
uses the RM at its core, and that has support for physical
independence.
I am afraid that at this time this is just a wish.
Marshall
vldm10
Model (CM) level.
However they are very related to RM level.
Example 2.5 on my web article has relation with the states.
The states of a relation are defined similarly as the entity's states
(the states are knowledge).
Example 2.5 can be more complex in real life and it is appropriate to
introduce the states
Chapter 4 is about the decomposition into the binary relations.
The idea is to create the relations so that each relation has a simple
key.
Then the binary decomposition is by the simple key.
In this sense idea is simple.
Vladimir Odrljin
vldm10
It seems that I made mistake, sanding this massage to me, instead to
Bob Badour
Vl.Odrljin
David Cressey
"Marshall" <marshal...@gmail.com> wrote in message
news:1147886349.7...@38g2000cwa.googlegroups.com...
> To achieve the big wins, though, we need a programming language that
> uses the RM at its core, and that has support for physical
> independence.
> I am afraid that at this time this is just a wish.
I have a different take on this, slightly.
We need a programming language that has a data model as its core, and has
support for physical data independence.
The desired data model would incorporate all the benefits of the present
RDM, at the least. such a data model would probably have to incorporate the
RDM as a sub model.
The programming language also needs a highly developed process model at its
core. The object oriented process model provides a good starting place.
Here's where I would start:
Since the time OOP became popular, the design and construction of objects
has been revolutionized. But OOP depends on two fundamental concepts, not
just one. In an object oriented world, there are objects, and there are
messages. The messaging scheme of languages ranging from Smalltalk to Java
is woefully inadequate. There has been almost no fundamental advance here
in 30 years.
In order to build on the successes that OOP has acheived, the messaging
scheme is going to have to go through a profound shift. When people get
around to building a better messaging scheme, they will discover that the
fundamental question is: how can objects share data coherently?
This turns out to be the same question that database theory began working
on, back in 1970, when Codd published. It's in a different guise, but it's
the same question.
x
"David Cressey" <dcre...@verizon.net> wrote in message
news:iBYag.4589$a23.4395@trndny01...
> The programming language also needs a highly developed process model at
its
> core. The object oriented process model provides a good starting place.
> Here's where I would start:
> Since the time OOP became popular, the design and construction of objects
> has been revolutionized. But OOP depends on two fundamental concepts, not
> just one. In an object oriented world, there are objects, and there are
> messages. The messaging scheme of languages ranging from Smalltalk to
Java
> is woefully inadequate.
>There has been almost no fundamental advance here
> in 30 years.
I'm not so sure about that.
> In order to build on the successes that OOP has acheived, the messaging
> scheme is going to have to go through a profound shift. When people get
> around to building a better messaging scheme, they will discover that the
> fundamental question is: how can objects share data coherently?
By quantum communication. By telepathy. Who knows.
> This turns out to be the same question that database theory began working
> on, back in 1970, when Codd published. It's in a different guise, but
it's
> the same question.
I've read somewhere that OOP is about isolation (sharing ?).
dawn
> "Marshall" <marshal...@gmail.com> wrote in message
> news:1147886349.7...@38g2000cwa.googlegroups.com...
>
> > To achieve the big wins, though, we need a programming language that
> > uses the RM at its core, and that has support for physical
> > independence.
> > I am afraid that at this time this is just a wish.
>
> I have a different take on this, slightly.
>
> We need a programming language that has a data model as its core, and has
> support for physical data independence.
> The desired data model would incorporate all the benefits of the present
> RDM, at the least. such a data model would probably have to incorporate the
> RDM as a sub model.
I'm not sure it would "have to" (depending on dbms's used) but it
certainly could have such a sub model. If such a programming language
has RDM as only one of several ways of modeling and working with data
in large shared data banks, then the RDM loses the exclusivity it seems
to demand, however. So, is the RDM sub model you mention one that is
still the only way to view persisted data, or simply one way? For
example, could XQuery (or similar) and SQL (or a better implementation
of the RDM) function side by side in a language that works with large
shared data banks without violating the RDM? (recognizing this could
be considered as not supporting physical data independence, a term that
perhaps we could find specific functional requirements for and ensure
they are met)
> The programming language also needs a highly developed process model at its
> core. The object oriented process model provides a good starting place.
Agreed.
> Here's where I would start:
>
> Since the time OOP became popular, the design and construction of objects
> has been revolutionized. But OOP depends on two fundamental concepts, not
> just one. In an object oriented world, there are objects, and there are
> messages. The messaging scheme of languages ranging from Smalltalk to Java
> is woefully inadequate. There has been almost no fundamental advance here
> in 30 years.
I have no doubt you are right about this, but am curious what, other
than generally-employed efficient standards, you find to be amiss in
this area with languages such as Java.
> In order to build on the successes that OOP has acheived, the messaging
> scheme is going to have to go through a profound shift. When people get
> around to building a better messaging scheme, they will discover that the
> fundamental question is: how can objects share data coherently?
Will that question be addressed separately from how we can share
procedures/functions? With Java Remote Method Invocation (RMI) or JERI
from the Jini libraries you ask questions of a service using the
language (methods) of the service (e.g. getFirstName()), however you
must first agree on Java as the programming language (a rather
significant downside). With SOA, you also agree on a means of invoking
functions (along with schema for xml data exchange, e.g.).
Don't languages typically have an adequate means of handling messaging,
even for high availability, high performance in a distributed
environment (e.g. using Jini) until you are mixing languages? In the
case of .NET (where I am woefully ignorant), the barrier might be in
the requirement for a specific run-time vm rather than a specific
language.
> This turns out to be the same question that database theory began working
> on, back in 1970, when Codd published. It's in a different guise, but it's
> the same question.
In the 1970 ACM paper he says
"The simplicity of the array representation which becomes feasible when
all relations are cast in normal form is not only an advantage for
storage purposes but also for communication of bulk data between
systems which use widely different representations of data."
XML and JSON are indications that we might have moved beyond the
simplest of forms for communication of bulk data between systems,
however, as there is no need to put data in what-was-once-called-1NF
for data exchange (and no one suggests otherwise, as best I can tell).
With more richness in the current RDM, are there any folks using it for
data exchange, employing set processing commands for data and
constraints, for example?
Cheers! --dawn
Marshall
> "Marshall" <marshal...@gmail.com> wrote in message
> news:1147886349.7...@38g2000cwa.googlegroups.com...
>
> > To achieve the big wins, though, we need a programming language that
> > uses the RM at its core, and that has support for physical
> > independence.
> > I am afraid that at this time this is just a wish.
>
> I have a different take on this, slightly.
>
> We need a programming language that has a data model as its core, and has
> support for physical data independence.
> The desired data model would incorporate all the benefits of the present
> RDM, at the least. such a data model would probably have to incorporate the
> RDM as a sub model.
I think I agree, but I'm not 100% sure I know what you mean. I observe
that a programming language will need things the ability to define and
invoke functions, have some mechanism for concurrency, I/O, etc.
> The programming language also needs a highly developed process model at its
> core. The object oriented process model provides a good starting place.
Not sure what you mean by "process model." That sounds like a
term for concurrency and interprocess communication, but I think
maybe you mean an inter-object communication model?
> Here's where I would start:
>
> Since the time OOP became popular, the design and construction of objects
> has been revolutionized. But OOP depends on two fundamental concepts, not
> just one. In an object oriented world, there are objects, and there are
> messages. The messaging scheme of languages ranging from Smalltalk to Java
> is woefully inadequate. There has been almost no fundamental advance here
> in 30 years.
Hmmm.
First off, I must note that I dislike the Smalltalk-y term "message
passing" to denote function invocation. As a term, I think it obscures,
rather than illuminates. Also, since function invocation is necessarily
of the synchronous-request-reply kind, it pushes out other forms
of (actual) message passing from the nomenclature. And it is my
belief that actual messaging is something that belongs in the language,
rather than in a library.
Secondly, can you expand on what you mean by "woefully inadequate?"
I'm not sure I know what you mean--do you have some specific
application (ha ha) in mind?
Thirdly, I'm not sure if you're aware of what some of the more advanced
functional programming languages can do. I note: closures, passing
fuctions as parameters, returning functions as results, and partial
application. These all seem desirable to me, and are all missing from
popular OOPLs. (Unless you count, say, function pointers in C, or
inner classes as closures.)
> In order to build on the successes that OOP has acheived, the messaging
> scheme is going to have to go through a profound shift. When people get
> around to building a better messaging scheme, they will discover that the
> fundamental question is: how can objects share data coherently?
Hmmm. If you mean read-only data, then this is something of an
implementation detail. If you mean read/write data (often called
"state") then I think the right answer is externalizing them from
the object. The separately-manageable database is the right
answer.
The problem with OOP is that it sprinkles state all through your
program, like sprinkling sand into fine machinery. That state
can't be managed or inspected, except with huge difficulty.
> This turns out to be the same question that database theory began working
> on, back in 1970, when Codd published. It's in a different guise, but it's
> the same question.
Is it possible it has the same answer now that it did then?
Marshall
x
> David Cressey wrote:
> > "Marshall" <marshal...@gmail.com> wrote in message
> > news:1147886349.7...@38g2000cwa.googlegroups.com...
> > In order to build on the successes that OOP has acheived, the messaging
> > scheme is going to have to go through a profound shift. When people get
> > around to building a better messaging scheme, they will discover that
the
> > fundamental question is: how can objects share data coherently?
> Hmmm. If you mean read-only data, then this is something of an
> implementation detail. If you mean read/write data (often called
> "state") then I think the right answer is externalizing them from
> the object. The separately-manageable database is the right
> answer.
I mean read-write code.
> The problem with OOP is that it sprinkles state all through your
> program, like sprinkling sand into fine machinery. That state
> can't be managed or inspected, except with huge difficulty.
It's like sprinkling oil.
Bob Badour
> David Cressey wrote:
>
>>"Marshall" <marshal...@gmail.com> wrote in message
>>news:1147886349.7...@38g2000cwa.googlegroups.com...
>>
>>
>>In order to build on the successes that OOP has acheived, the messaging
>>scheme is going to have to go through a profound shift. When people get
>>around to building a better messaging scheme, they will discover that the
>>fundamental question is: how can objects share data coherently?
>
>
> Hmmm. If you mean read-only data, then this is something of an
> implementation detail. If you mean read/write data (often called
> "state") then I think the right answer is externalizing them from
> the object. The separately-manageable database is the right
> answer.
>
> The problem with OOP is that it sprinkles state all through your
> program, like sprinkling sand into fine machinery. That state
> can't be managed or inspected, except with huge difficulty.
I think now is a good time to revisit Keith's statement about
simulation. One should first note that there are different types of
simulation.
Simula was invented for the sort of simulation that builts large
unpredictable state machines by combining lots of small predictable
state machines arranged in complex patterns.
What each simulation will look like will depend largely on what one
wants to simulate.
A very simple simulation of digital logic circuits might involve:
A relvar of nodes { node, voltage }
A relvar of events { time, node, new_voltage }
A relvar of devices { device }
Each node has a specific voltage.
A possible representation of a device is a device name and a relation of
connections: { pin, node }
Each event specifies a time when a device will drive a node to a
threshold voltage.
The nodes and devices define the circuit and initial voltages.
Initially, the events relvar is empty.
The device supertype would declare a virtual state transition function
that given the device, the connected nodes, and the time will calculate
the relation of events on the pins driven by the device:
state_transitions(device,nodes,time): SAME TYPE AS events
Specific subtypes of device would define the state_transitions function
differently. For instance, an AND gate type and an OR gate type might
drive their output pins to different values with the same propagation
delay while two different AND gate types might drive the output pins to
the same value with a different propagation delay.
The simulation itself might appear something like:
t = time(0);
changed_nodes = nodes;
WHILE changed_nodes{} AND t <= time(simulation_end) BEGIN
/* Determine the future events based on the current state */
WITH EXTEND devices ADD
state_transitions(
device
,nodes JOIN ( connections(device) {node} )
,t
) AS transitions
{transitions}
AS new_transitions,
WITH UNGROUP new_transitions (transitions)
AS new_events,
events = events UNION new_events;
/* Advance time to the soonest next event */
WITH events {time} [time > t] AS future_time
t = MIN(future_time,time);
/* Trim the fat and remove no-op transitions */
DELETE FROM events
WHERE time = t
AND (
TUPLE { node node, voltage new_voltage } =
TUPLE FROM nodes[nodes.node = events.node]
)
;
/* Reflect the state transitions in the nodes
and record which nodes changed */
WITH events[time = t] AS current_events,
changed_nodes = ( current_events
RENAME new_voltage AS voltage)
{node, voltage},
UPDATE nodes JOIN current_events
SET voltage = new_voltage;
END;
All that remains to make a fully functional simulation is to create the
device subtypes, flesh out the type specifications and initialize the
relvars. Each device subtype requires only a name and a state transition
function.
erk
> If such a programming language
> has RDM as only one of several ways of modeling and working with data
> in large shared data banks, then the RDM loses the exclusivity it seems
> to demand, however. So, is the RDM sub model you mention one that is
> still the only way to view persisted data, or simply one way? For
> example, could XQuery (or similar) and SQL (or a better implementation
> of the RDM) function side by side in a language that works with large
> shared data banks without violating the RDM?
I don't know about XQuery (which is meant for XML values), but
certainly most UIs are a tree view of data, of which a relational
database can support many (different restricted views of the same
data). However, updates to data need the same constraints as the
database itself, so from that standpoint, the application is a node of
a distributed database.
But for presentation, different languages make a lot of sense.
> > The programming language also needs a highly developed process model at its
> > core. The object oriented process model provides a good starting place.
>
> Agreed.
What is an "object oriented process model"?
> XML and JSON are indications that we might have moved beyond the
> simplest of forms for communication of bulk data between systems,
> however, as there is no need to put data in what-was-once-called-1NF
> for data exchange (and no one suggests otherwise, as best I can tell).
I'm not sure what would be so different now, and I don't really see
that XML is an advance. I still see plenty of cases where
comma-separated values are more than sufficient, and properties files
(e.g. in Java apps) usually easier to read and write than XML
configuration files. S-expressions are fine, rich structures, and date
from long before the relational model.
> With more richness in the current RDM, are there any folks using it for
> data exchange, employing set processing commands for data and
> constraints, for example?
I don't think the RDM has changed radically; the richness derives from
user-defined types which Codd identified very early, and from proper
use of relation-valued attributes (RVAs). Beyond that, the tree
structure of an XML document doesn't capture anything additional; it
is, perhaps, better suited to naturally hierarchical data. But most of
the XML I see in practice communicates (badly) either a graph, or
relations (via ID/IDREF and cobbled-together versions of these). Date
identified a simple XML encoding of relations.
Constraints and commands are different, but constraints should be
carried with the data. However, this would require a standard language,
for complex constraints not capable of being encoded with simple tags.
- Eric
Bob Badour
> dawn wrote:
>
>>If such a programming language
>>has RDM as only one of several ways of modeling and working with data
>>in large shared data banks, then the RDM loses the exclusivity it seems
>>to demand, however. So, is the RDM sub model you mention one that is
>>still the only way to view persisted data, or simply one way? For
>>example, could XQuery (or similar) and SQL (or a better implementation
>>of the RDM) function side by side in a language that works with large
>>shared data banks without violating the RDM?
Dawn is a self-aggrandizing ignorant who lacks the ability to reason
simple things. Due to the proscription of subversion, it would be
pointless to implement anything else 'side by side'. If one wants to use
a different syntax, it would make much more sense to use a translator.
However, none of the examples of alternate syntaxes she gives could
possibly offer any useful benefit over predicate logic.
> I don't know about XQuery (which is meant for XML values), but
> certainly most UIs are a tree view of data, of which a relational
> database can support many (different restricted views of the same
> data). However, updates to data need the same constraints as the
> database itself, so from that standpoint, the application is a node of
> a distributed database.
>
> But for presentation, different languages make a lot of sense.
>
>
>>>The programming language also needs a highly developed process model at its
>>>core. The object oriented process model provides a good starting place.
>>
>>Agreed.
>
> What is an "object oriented process model"?
See what I mean about her self-aggrandizing ignorance? "Process model"
is meaningless gibberish, and she is either too ignorant or too stupid
to recognize that fact.
Even if one goes the extra mile and tries to make sense of the gibberish
by substituting "computational model" or "model of concurrency", the
object oriented folks do not have a consensus on either so using the
definite article and suggesting that provides a unique starting place is
lunacy.
erk
> If one wants to use
> a different syntax, it would make much more sense to use a translator.
> However, none of the examples of alternate syntaxes she gives could
> possibly offer any useful benefit over predicate logic.
For data representation, I agree completely. But I think of UIs and
reports as forms of data extracts: they're restricted and (typically,
but not always) tree-structured "views" of a subset of a database. The
power of relational for shared databases is in part its ability to
support arbitrary extracts of the data in an egalitarian fashion - no
application bias. Any given UI or report includes a set of (typically
key-related) restrictions and projections across different relations.
It appears to me (and I could be wrong, as this is only a pet
hypothesis at this point) that expression languages to derive these
would be tree-structured, or at least would be different than a
relational algebra or calculus.
> Even if one goes the extra mile and tries to make sense of the gibberish
> by substituting "computational model" or "model of concurrency", the
> object oriented folks do not have a consensus on either so using the
> definite article and suggesting that provides a unique starting place is
> lunacy.
In fairness, it wasn't Dawn who introduced the term. I thought maybe it
had a meaning, but suspect it doesn't. Agreed that the O-O folks don't
have a consensus on this, nor is it critical in the discussion of data.
- Eric