Re: Object-oriented SQL statements

On Jul 25, 4:59 am, JOG <j..._at_cs.nott.ac.uk> wrote:
> On Jul 24, 5:33 pm, Zorro <zor..._at_tx.rr.com> wrote:
>
>
>
>
>
> > On Jul 23, 2:59 pm, JOG <j..._at_cs.nott.ac.uk> wrote:
>
> > > On Jul 23, 8:45 pm, paul c <toledobythe..._at_oohay.ac> wrote:
>
> > > > Marshall wrote:
> > > > > On Jul 23, 7:43 am, Nis Jørgensen <n..._at_superlativ.dk> wrote:
>
> > > > >>Bob Badour skrev:
>
> > > > >>>Zorro wrote:
>
> > > > >>>>Please comment on the approach illustrated in :
>
> > > > >>>>http://www.zhmicro.com/Database.pdf
>
> > > > >>>>On July 19 this was posted at comp.object.
>
> > > > >>>Typical nonsense from comp.object on how to cripple a DBMS by forcing it
> > > > >>>through a low-level procedural language and a profoundly limited interface.
>
> > > > >>That would be my guess as well, without following the link :-)
>
> > > > > LOL
>
> > > > > Skimmed over it. It doesn't support joins. You can only query one
> > > > > table at a time.
>
> > > > > This has become the first question I ask when I see a new
> > > > > approach: what does join support look like?
>
> > > > > Marshall
>
> > > > Good points, it looks like one has to re-invent the relops for oneself.
> > > > Successful consultants usually see the financial advantage in that.
> > > > Yet they say they have a happy marriage which confines them to the SQL
> > > > standard. I predict divorce at some point. It also says they have a
> > > > verifiable conceptual model. What does it mean to verify the conceptual
> > > > model of SQL? Looks like they are just throwing words around, like most
> > > > vendors, they might change all the words tomorrow and still sell the
> > > > same product. The selling point might be that it would be just as
> > > > coherent as it was yesterday, namely not at all. There seems to be no
> > > > lack of buyers who consider that an advantage.
>
> > > > Does the OP somehow qualify as SPAM, eg., the database equivalent of
> > > > scientology?
>
> > > > p
>
> > > I just wrote off the original post as spam as soon as I saw it. More
> > > retro-grade rubbish. No wonder people like pascal just give up.- Hide quoted text -
>
> > > - Show quoted text -
>
> > Well, I was looking for comments from experts who would also take a
> > few minutes to evaluate the model. If you do not have the time, there
> > is no need to say or do anything.
>
> > Thanks just the same.
> > Dr. Z.
>
> I did give you comments. When I read:
>
> "Generally, the purpose of executing the select statement is to
> receive a set of records from a database. An object-oriented approach
> to collecting a set of objects is to instantiate a container with the
> type of such records."
>
> I decided it was retro-grade nonsense. If you don't like my assessment
> this is hardly my fault. I then offered you the further comment that
> it is this sort of thing that makes real experts such as f. pascal
> give up.
>
> I would have found these very useful points for further research if it
> had been my work, but then i don't have anything to sell.- Hide quoted text -
>
> - Show quoted text -

I did not find time to check on this earlier. Sorry about that. Your opinion is just fine. At least it tells me to rewrite the article in greater detail so it will be within the grasp of more readers. There is nothing wrong with the phrase you are quoting. It is just too brief, expecting the reader to think and analyze it by reading it in the context in which it is presented.
The point is that, we want to use well-established relational databases in an OO setting (the language). We do not want to use imbedded SQL, nor we want class libraries such as ODBC. Instead, we want to be able to use statements that are almost like SQL, but that allow the use of OO facilities in the language. And, we want to involve the compiler directly (as opposed to class libraries). The only way for compiler to catch our errors at compile time, is to have linguistic constructs (like those illustrated in the article) with clear semantics. That way, we can understand what is it that we want to do, and the compiler can check on the validity of our way of saying it.
I hope this helps. There is no need to continue on this topic. Thanks,
Z. Received on Fri Jul 27 2007 - 18:15:23 CEST