Re: Why are [Database] Mathematicians Crippled ?

Jan
Norbert
Tegiri

> On Friday, 30 January 2015 19:21:16 UTC+11, Jan Hidders wrote:
> On Saturday, 31 January 2015 07:01:04 UTC+11, Norbert_Paul wrote:
> On Saturday, 31 January 2015 08:00:56 UTC+11, Tegiri Nenashi wrote:
> On Saturday, 31 January 2015 11:01:42 UTC+11, Jan Hidders wrote:

> No, I think you misunderstand what he is trying to do and what he is asking. He is proposing an extension of the relational model with new concepts that can deal with data that represents toplogies over other data. And his question is if practitioners could make sense of these concepts and use them.

Well, the RM can't be "extended". End of story. So if that is what he is trying to do, his project is dead in the water.

(There may be a communication problem; by relational model", you may be meaning one of the 42 "relational model" that methematicians use, all of which have little bearing to The Relational Model, and which are unknown to practitioners [except for people like me, who know a little about what you guys do in the void]. So much for the "exactitude of mathematics", where they use terms out of context and with little specific meaning. In that case, you need to specify which "relational model" you are using; what the base relations are; etc. )

> > The answer seems to be "no".
>
> Yes, obviously.

Er, the evidence is, if you bother to read just the intro to my first response in the Topology thread, I do understand it. What you are trying to do is completely understandable, they way you communicate it is gibberish. I repeat, there are thousands of such systems, in terms of products, millions of such systems, in terms of licences. It is staggering that you do not know this.

I can't speak the way you guys do, I don't have your training or jargon, it is silly to expect that from a practitioner. Since you are trying to communicate with the other side of the chasm, you cannot hold onto your demands, you need to communicate differently. Otherwise you will fail.

And in case it is not obvious from the evidence, I have spent a fair amount of time and energy (a) reading your papers (b) typing responses (c) erecting data models, trying to make sure that you do *not* fail. So I am trying to communicate, from this side of the chasm.

It applies the other way around as well, because you guys live in such abject isolation from the real world. You have to be able to read and understand the models and pictures that implementers have been using since 1985. If you have never seen them before, fine, start now. I am happy to help in that, but you haven't asked a question, I don't know what it is that you do not understand.

> There is a reason for that. As Todd J. Green eloquently put it, database practitioners are buried in a soul crushing routine work (select * from employees and some such). There is not a lot of insight to gain there.

Now that is the most pathetic attempt to assert superiority that I have seen in a long time. Let me assure you that I spend no time doing any such nonsense. All that is, is myths that you have created about us, in order to concoct and maintain a sense of superiority. Manufactured, false. You degrade yourselves when you do that. Transparently childish methods of shifting the responsibility from yourselves, the cause, to the other person, who did nothing, and therefore cannot be blamed. Pathetic. Pharissaic. Unscientific. Babies with big heads.

If not for the scientists, unknown to you because they do not publish their papers, and the practitioners at the high end, the RDBMS market would have not moved at all since Codd's work. You guys have produced nothing since 1970, nothing useable, nothing installed anywhere, except wind, and you have the hide to position yourself as superior. Give me a break.

Further, I repeat, I have no such idiotic problems when I deal with scientists in the manufacturing or banking industry. We have no problems at all, communicating with each other. Differences, yes, problems, no. We learn each others diagramming and standards, and we are off. This problem (thread title and this interaction) is peculiar to the theorists in the RDB field.

So get off your childish high horses, get back on the ground, and start doing some honest work. Find out what real theoreticians in other industries do, and start emulating them.

> > The answer seems to be "no".
>
> Yes, obviously.

Ok. Do something worthwhile with the lifeforce, the energy, that God has given you. Get a goal in life. Don't just sit there at "no, Derek doesn't understand". Bridge the gap. Make it your goal to get Derek to understand it. If you want to succeed, you have to be able to communicate your product to the intended client.

Don't forget, the law we use on this side of the chasm, outside the void: if it isn't documented, it does not exist. Documentation forces you to think in terms of the reader, the recipient. When I produced the data model, I was trying, in my mediocre, stupid, innocent, dumb-implementer way, to start that process for you. Pardon me if I failed.

> It is not a secret that [SQL] database applications in science are mediocre, at best. Still, I would suggest looking up there (rather than in "more practical" fields. For example, chemical databases do some nontrivial matching of substances. Likewise, in protein chemistry topology might be very important.

By all means, do some research and find out what is out there. Start with google, find the standards bodies, chase them down, find the universities, chase them down.

Be careful with that. I have some experience helping a couple of American universities. Under-funded. They used "models" for studying DNA sequences and alleles. Post-grad research level. They acquired the "models" from some non-standard "standard" site, and then extended them. It turned out to be very poor, freeware/shareware/vapourware/nowhere. The point is, there are a lot of mediocre theoreticians out there, and their SQL implementations are thus mediocre. I rewrote their entire database in a month, it replaced hundreds of months "work". Their queries that took overnight to run now take seconds. Their SPSS analyses (export to SPSS; cubify; grind; produce a graph) are now halved, because the other half is in the RDB, in useable form, available in seconds. I did the work free, because their work is a benefit to humanity.

> Derek seems to think it was already already implemented in commercial DBMSs 20 years ago.

Exactly. Forget the "seems to think". Forget "Derek". Do some work. Find out for yourself. Otherwise, you are reinforcing the vacuum that you live in (I have already identified that to you, and you are once again proving it).

Type "GIS" or "GEAC GIS" into a google search window. Type <return>. Examine the result window. With your eyes open. Type "mining industry", "architectural spaces", and repeat. Now try all those searches, again, after adding the word "images" to the search window. Repeat. Look at the pretty pictures. Are there any images that look like what you are trying to do ? Good. We had them THIRTY YEARS AGO. There is nothing new there, except for range-of-colour; ease of capture, storage; multilayered componentry. The database tables are the same as twenty years ago, plus a few new attributes.

Here's a system that I worked on in Canada TWENTY years ago. They have come a long way in that thirty years, but they are, being a govt dept, well behind the commercial market http://proceedings.esri.com/library/userconf/proc02/pap1302/p1302.htm Notice how small their GIS component is.

> > There is a reason for that. As Todd J. Green eloquently put it, database practitioners are buried in a soul crushing routine work (select * from employees and some such). There is not a lot of insight to gain there.
>
> Well, yes and no, the concepts should of course be intuitive for the intended users and it is always easy to fool yourself into thinking that it is, but, yes, it might be a particular class of users that we don't find here. FWIW, I do think there is indeed something interesting going on in Norbert's ideas,

I have stated that, perhaps in different words. I may be "novel" as pertains to research, I wouldn't know.

As it pertains to the world of Relational Databases, your work is about twenty years behind the commercial products, about ten years behind what I saw in Germany in 2007 (that means 17 years behind). Please check for yourslef. Its only value is, as already stated, if you package it as a library plus SQL scripts, that any app can use, if I could get that, I would use it, instead of the awk:gnuplot that I use now. All detailed in my posts.

Yeah, sure, it might be new in the isolated world of "RDB" theoreticians, or to one of the forty two "relational models" that they talk about, but that has no relevance to the real universe of RDB.

And the question that begs, the title of this thread, is WHY is there such a chasm betwen the real RM and the "relational models" that you use, why is there a chasm re communication ? Deal with that, not Norbert's paper, not the Topology thread. I made that clear from the OP. Or, focus on the topology thread and prove, yet again, loud and clear, without need for a mathematical poof, that you cannot deal with the chasm, the cause of it, and the best you can do is focus on a tangent and make immature comments.

Cheers
Derek Received on Sun Feb 01 2015 - 09:58:31 CET