Re: Views for demoralizing

"Jonathan Leffler" <jleffler_at_earthlink.net> wrote in message news:420C558E.8000501_at_earthlink.net...
> Dawn M. Wolthuis wrote:
>> "Alfredo Novoa" <alfredo_novoa_at_hotmail.com> wrote:
>>> "Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote:
>>>>>I don't see a lot of usefulness in creating such a term.
>
> [JL: the prior posting doesn't provide the attribution for the quadruply
> indented statements such as "I don't see...". I suspect it is Alfredo.]
>
> [...major snippage...]
>
>>>>> IBM knows very little about the Relational Model. You should
>>>>> ignore all that.
>>>>
>>>>Wasn't Codd an IBM employee when he wrote his early papers?
>>>
>>> Yes but it does not change anything. IBM never understood the RM
>>> and they created the SQL aberration. Codd, Date and now Darwen
>>> abandoned IBM.
>>
>> When did Darwen leave IBM? I'm not in the inner circle and am apparently
>> not keeping current on such matters (perhaps the news
>> didn't make People magazine?). Nevertheless, I'm sure there are
>> some at IBM who would spout the same doctrine as you.
>>
>>>> I'm thinking there are people at IBM who know quite a bit about
>>>> relational theory, whether they opt to completely buy into it
>>>> or not.
>>>
>>>If you know a bit about relational theory you always opt to buy into
>>>it.
>>
>> Oh really? I know a bit about relational theory. I once bought into it.
>> I've learned a bit more. I'm not as gullible now.
>
> This is an intriguing comment, so I lobbed it to Fabian Pascal, and he
> asked me to follow up for him...
>
> FP: Do me a favor and post a message in that exchange and say I am asking
> her to post the link to our exchange from years ago and that I am still
> waiting for her mathematical proof that MVDBs are relational, which she
> promised then. :)

Say 'hi' to Mr. Pascal from me. One thing I said I would provide is the answers to Date's questions in the paper on 1NF. I e-mailed those about a year ago to two different e-mail addresses I had from him and both bounced. I was bummed. I figured he had put me on the "don't call me, I'll call you" list. I suspect that is what he meant by that. I don't know if I would use the words "proof that MVDBs are relational" because unlike others he chats with on proof of whether a particular product or theory is relational or not, I don't really care if they are called relational -- I call them "functional" which has the double meaning that they are based on mathematical functions (which are relations) and, well, they work.

> And, also:
>
> FP: You can also say that I said that Dawn is another Celko, throwing all
> sorts of fuzzy comments about mathematics around to scare the uninformed
> around and impress, without anything to back her up.

First of all, Celko does some good work and has helped many practitioners with their SQL work and I have no reason to saddle him with any comparison to me. Secondly, unlike Celko, I do not consider myself an expert in this field as yet. I have not (yet ;-) written nor published anything. That seems to me to be a very big difference.

I use this forum to learn -- I have other forums for teaching. Even when I'm doing the soapbox thing, it is to find out what those with other experiences think so I can be better informed. If what I'm saying doesn't resonate, then I want to understand why and either change my opinion or say it differently. But when it comes to scholarship, I am very interested in precision. You will see that I contributed to the glossary for this forum as that is something I do care about -- if we are sloppy with the terms we use to communicate with each other, we will talk around in circles wihout getting anywhere. You will also note that I've brought up how concerned I am about simply swapping out definitions under various key terms data modeling theory. Mathematical "relations" do not prohibit nested relations or embedded ordered lists, for example. This was true even before Date said it was. But there is certainly confusion on this among practitioners with a lot of that due to sloppy terminology and changes in such within the industry. When I use the term "relation" I am very careful to be precise about it and not toss everything but the kitchen sink into the defintiion. But I am a practitioner, interested in advancing the practice of data modeling for more flexible designs and I'm not (currently) interested in being the person who lays out all of the mathematics of functional dependencies or of the model underlying XML. I do read some of those papers, however.

> FP: Known technique and I am not gullible anymore either.

More fluff without content, my friend. Tell me precisely where I am wrong in my thinking, starting with something that is perhaps easy for a relational theorist like ordered lists as possible Type for an attribute in a relation. What is mathematically wrong with that approach? Where precisely does 1st order predicate logic fail to permit queries on such data? I have received no precise responses to any content from Mr. Pascal -- just agreement with me that I am truly ignorant. Perhaps someday he will say that, indeed, I have learned something and now am one step up the ladder from being totally ignorant. I won't hold my breath, however, and I suspect it is not related to whether or not I am ignorant (and I say that with a smile).

>>>If still there are people at IBM who know about relational theory it
>>>is clear that they don't have any decision power.
>
> Yes. There are a number of people who know and care, but it is hard to do
> much when the main systems are very much based on SQL. Backwards
> compatibility is an incredibly heavy shackle, forever constraining that
> which one would like to do. People like RequiredTech (of reputed TRDBMS
> fame) who can start from scratch are much better off - right up until they
> release version 2.0. Thereafter, they are about as hobbled as everyone
> else. :-(

Yes, and one of my points is that as the industry moves forward we might just have to leave SQL behind and NOT have all products be backwards compatible. It's a problem, undoubtedly.

>> Or perhaps they know about it, but the wealth of experience from
>> employees and customers of IBM might have lead them to some other
>> conclusions.
>
> As has been intimated elsewhere (not necessarily in this thread), there
> are few IBM customers who've had a chance to work with a non-SQL RDBMS.
> The same goes for the customers of pretty much any other supplier of SQL
> DBMS.
>
> --
> Jonathan Leffler #include <disclaimer.h>
> Email: jleffler_at_earthlink.net, jleffler_at_us.ibm.com
> Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/

Thanks Jonathan, and let me know if there is an e-mail address I can use for F. Pascal. I plan to start doing some writing in May and can send him something then if not before. --dawn Received on Fri Feb 11 2005 - 17:00:56 CET