Re: 3vl 2vl and NULL

David Cressey wrote:
> "dawn" <dawnwolthuis_at_gmail.com> wrote in message
> news:1134096079.471233.306790_at_g49g2000cwa.googlegroups.com...
> >
> > David Cressey wrote:
> > > "dawn" <dawnwolthuis_at_gmail.com> wrote in message
> > > news:1134055293.865826.300270_at_g47g2000cwa.googlegroups.com...
> > >
> > > > If you were to look at the dollars spent on U2 and the dollars spent
> on
> > > > DB2, ... I'm certain there are many more features in DB2, and I
> > > > respect the product (without having used it in any production
> > > > environment) but it is SQL-based which might explain why I suspect
> > > > suspect that if we could have a database-shootout for use in software
> > > > development, U2 just might win.
> > >
> > > You're begging the question.
> > > If there's something wrong with DB2, can you identify what's wrong?
> >
> > I have not spent enough time with DB2 to do anything more than lump it
> > with other SQL DBMS tools. Sorry.
> >
> > > If there's something wrong with SQL, can you identify what's wrong?
> >
> > I believe that over the past couple of years I have identified several
> > things that make it a less productive tool than what I might want.
> > I'll give just one example: You cannot back an arbitrary "screen" (UI,
> > e.g. web page) with a SQL VIEW even though views need not be
> > normalized.
> >
> > >
> > > What's a "database shootout", anyway?
> >
> > I can't believe i called it a shootout when I could have called it a
> > bake-off. It is a competition. Because most benchmarks in the dbms
> > area are based on SQL, there is standard way to give comparison data of
> > SQL DBMS tools with SQL-as-a-second-language tools (non-1NF DBMS's, for
> > example). There could be a competition where requirements are given
> > for an application and developers write the apps. A panel could judge
> > the solutions. This tests the entire development environment, rather
> > than strictly the DBMS, but a DBMS would be one of the tools in a
> > winning solution. I wish I could find a competition like this. Last
> > year Spectrum International help one strictly for MultiValue/Pick
> > vendors. --dawn
> >

>

There are surely many instances where superior technology is not noticed that way by the marketplace: OS/2 and Win NT might be one instance. Also, MV vendors rarely market directly to customers so it isn't easy to see what their pie of the pie is. They have traditionally used a VAR model. An exception is about to arrive on the scene as InterSystems Cache' will now have an MV/Pick implementation and they do market to customers directly. I know industry media rarely mentions Pick, but I don't know if that relates to a small market.

Sometimes MV/Pick is treated like an embedded database. When IBM acquired U2 (UniData and UniVerse) through the Informix acquisition, they were surprised to find that they were an end-customer of U2. The end-customer often pays the VAR and might not need to think about the database.

I would like to see a pie chart that shows just two categories -- SQL-DBMS (or RDBMS) and "other DBMS" (which might have SQL as one language, but would not be classified as relational) with a percentage of transactions (the def of which might be difficult) occuring in each category. I honestly don't know which slice of that pie would be larger - SQL-DBMS or Other, do you? The second category would include IMS, Berkeley DB, MultiValue/PICK, MUMPS/Cache' and many more, both old and new offerings. --dawn

>
> Not that it couldn't happen. But why?
Received on Fri Dec 09 2005 - 15:36:26 CET