Re: Relational and MV (response to "foundations of relational theory")

"Eric Kaun" <ekaun_at_yahoo.com> wrote in message news:DLzYb.23126$ft2.1420_at_newssvr16.news.prodigy.com...
> This is in reply to the “foundations of relational theory?” thread which
> began on September 26, 2003 and accumulated over 500 responses before
> petering out on November 7, 2003. Yes, I read them all, and was
> fascinated. Apologies - I found them only after the thread had run cold.
>
> * Dawn wrote: “[quoting Codd] 'Why focus on just one type of compound
> data? The main reason is that any additional types of compound data add
> complexity without adding power.' The last statement of these has no
> mathematical backing to it.”
>
> Perhaps

I disagree. Codd provided appropriate mathematical backing to all of his claims. The onus lies on Dawn to prove some deficiency requiring an additional compound data type.

> * Dawn wrote: “...why would mathematicians or scientists ever add
> complexity? Perhaps to have a better model?”
>
> Mathematicians and scientists have different definitions of “better.”

Dawn is ignorant and stupid. If complexity has any advantage, it is solely psychological. Any claim of improved psychology requires a carefully controlled empirical result to have any validity. Where is the empiricism backing the claim?

> * Dawn wrote: “...many PICK/MultiValue applications have survived over
> 20 years – NOT JUST THE DATABASE, BUT THE ENTIRE APPLICATION!”

Dawn is too ignorant and too stupid to recognize that Pick's flaws render desired changes prohibitively expensive. She then concludes the astounding cost and rigid inflexibility improves producitivity. Her idiotic nonsense does not warrant reply.

> * Dawn wrote: “...constraints mean nothing outside of the context of an
> application anyway – what one DOES with the data is part of what defines
> it.”
>
> The first part of this is completely wrong

It suffices to note that Dawn is profoundly ignorant of the most basic data management principles.

> * Albert Kallal wrote: “...the fact is that most applications don't need
> much constraints anyway. The popular success of dBase years ago, and the
> popularity of MySql tday proves this...”
>
> Popular success proves nothing except success (i.e. A => A).

You will find plenty of vociferous ignorami who are only too eager to prove their axioms.

> * Albert Kallal wrote: “...existing data AND CODE DID not have to be
> modified when a 'child data set' is added.”
>

I long ago proved Albert's fatuous assertion false. Search for "red blue car"

> * Anthony Youngman wrote: “...I'm not saying that SQL's abstraction is a
> bad thing (it does make ad-hoc queries easier), but by hiding the
> realities of implementation from programmers you are basically
> handicapping them such that they will NEVER be able to even compete
> successfully (let alone win) if one of the requirements is to stretch
> the hardware to its limits.”
>
> I've never seen a requirement phrased like that, but why don't we still
> use assembly language, and raw file system manipulation?

WOL is another profoundly stupid ignorant. Without physical independence, the programmers will never compete successfully. Period.

> * Anthony Youngman wrote: “The reason MV is so powerful is that it's
> easy for us to mimic relational – it's a *subset* of what we have
> available to us.”
>
> No, it's not. Since I can generate many “containment” hierarchies from a
> 3rd normal form database, I would argue the opposite.

Wol is too stupid to realize that less is more. In this particular case, the added complexity of Pick results in less expressibility, less functionality and incorrect functionality. I already proved all of the above. Wol is too stupid and too obstinate to comprehend that fact.

> * Dawn wrote: “MV/Pick originated by starting with the concept of
> querying data using spoken/written language. So, both XML and MV/Pick
> are language-based data models.”
>
> Which language? English? I'd wager that a human language has different
> requirements, and a completely different development process, than a
> computer language.

Dawn is a vociferous ignoramous who is too stupid to realize that people don't want dbmses with the information retrieval reliability of the human brain. They want dbmses with much greater reliability.

> * Dawn wrote: “...there is NO PROOF AT ALL that an RDBMS yields higher
> developer productivity...”

I disagree. All available empirical evidence indicates that they do yield higher productivity even with the flawed products we have to work with today. Again, I note that the Pick proponents invariably measure "riding horses" against "pushing cars" to show the higher productivity of horses relative to cars.

Horses still exist, and some people still use them. The longevity of the technology does nothing to demonstrate its relative worth.

> * Delusion wrote: “only value of the multivalued model is it's
> multivaluedness / the fact that it's so explicit / within the
> multivalued model data has more meaning”
>
> Relations are explicit, and types and constraints make the meaning of
> your data far more explicit. MultiValues express, at best, one kind of
> hierarchy.

Multivalues are ill-defined and sloppy, which make them rather more implicit than explicit.

> * Anthony Youngman wrote: “Sound mathematics is worthless.

Does anyone need any further proof of Wol's profound ignorance and stupidity?

> * Mike Preece wrote: “It [Pick/MV] works when you work at it!”
>
> So does practically anything.

Mike's statement hardly argues for high productivity. Does it?

> * Tony Gravagno wrote: “This is why Relational is being described as a
> religion in this thread, and relational theory as dogma. Yes, dogma is
> the correct word because the authority that hands down this
> misinformation is the educational system that creates misguided teachers
> out of misguided students.”
>
> Strange, then, that the educational system is becoming more of a trade
> school, teaching not relational but SQL, and even specific
> implementations of SQL.

Gravagno is likewise profoundly stupid and a blatant sophist. He is no different from a Christian Creationist who argues that natural selection is religious dogma in an effort to confuse the ignorant.

> * Cmurthi wrote: “Aside: Knuth's 'fundamental algorithms' note from '68
> or so, where he says proudly (imsr) 'none of these algorithms have
> actually been tested on a computer,' as if that made 'em superior to
> those that were programmed... does show the mind-set, what?”
>
> Yes, it shows that logic and mathematical analysis have some place in
> designing logical machines (programs). Knuth's pride might just be based
> on the quality of his thoughts, and the logical strengths of the
> algorithms.

I fail to see what Knuth's statements have to do with any of Codd's results, though.

> * Anthony Youngman wrote: “Don't forget, Pick doesn't have an optimizer
> because it doesn't need one. The stats I've seen – given a known key you
> can retrieve ANY record from the disk with on average 1.05 head
> movements. Bearing in mind the typical Pick record is equal to several
> relational rows, that's a hell of an advantage if your system is i/o
bound.”
>
> Don't forget that Donald Knuth (whose specialization was algorithmic
> optimization) quoted C.A.R. Hoare (one of the pioneers of axiomatization
> in software) in writing: “Premature optimization is the root of all evil
> [in programming].”

Wol's ignorant assertion is worthless sophistry. Pick doesn't have an optimizer because it forces this task onto all users--some of whom may have no interest in the arcana of physical representations and structures.

> Again, why don't you develop in assembly language?

Exactly! However, even assemblers can optimize code with peephole optimizers or with instruction reordering for superscalar pipelined architectures.

> * Tony Gravagno wrote: “In a relational database all columns _MUST_ be
> pre-defined by type. In a Pick database, dictionaries are used to
> describe views of the data but the application programs which update the
> data maintain the “definition” of how the data is used, referential
> integrity, etc. ... though a well documented database will have a
> definition for every field in the dictionary, or documentation to
> explain how the app is using the fields.”
>
> Though apparently you believe this lack of typing to be an advantage,
> the potentially catastrophic disadvantages of this are one of the
> problems relational was designed to address.

Again, it suffices to note Tony's profound ignorance and stupidity. How do you propose to reach a person who delivers direct evidence for your arguments while thinking they support the contrary argument?

> * Mark Brown wrote: “There are only three data file structures:
> sequential, indexed sequential, and random. Anything else is just
> packaging.”

Brown's assertion is so stupid and nonsensical it does not warrant a reply.

> * Mark Brown wrote: “These tools are ALL the same: they are only as good
> as their users. Pick one that works for you and go with it.”

Brown's idiotic statement amounts to an abdication of critical thought.

> * Bobj wrote: “You can implement a lot of things using a simple ISAM and
> a few subroutines (classes if you prefer) without carrying the baggage
> of a DB. And that's why we find the Pick model to work so well. We can
> use as much or as little of the DB as we care to.”
>
> Chris Rock (a comedian) said (paraphrasing): “You can drive a truck with
> your feet if you want to – that doesn't make it a good idea.”

One can also loosen the steering and disable the brakes before driving the truck--that doesn't make it a good idea either. (Especially in mountainous terrain.)

> * Dawn wrote: “...persisting data based on language – such as modeling
> entire propositions together rather then piecing them apart to the
> extent done in an RDBMS – makes sense because we are not trying to
> persist mathematical relations ultimately – we are trying to persist
> propositions.”
>
> I don't understand why persisting a “language proposition” (as if human
> language were a good model!) is better than persisting a logical one.
> Human language is a poor basis for mechanizing reasoning (unless maybe
> you're talking about Esperanto or Vulcan).

One might ask why we developed prepositional logic in the first place if human language is such a good way to express logic. Consider what this says about Dawn's alleged credentials as a mathematician. Have you met any other mathematician who is as profoundly ignorant of the nature and of the purpose of mathematics?

> * Mike Preece wrote: “[regarding cars and owners] There's an implied
> relationship and it's a parent-child, one-to-many relationship. [...]
> The database is required to record information about people, primarily,
> and the children they have, cars they own, and which car they usually
> drive.”

Cars can have only one owner? That's news to me! Preece is another vociferous ignoramus.

> More to come, in case you're having trouble sleeping. :-)

Are you a masochist? Received on Wed Feb 18 2004 - 20:21:42 CET