Re: A database theory resource - ideas

On Mar 17, 8:50 pm, "Bruce C. Baker" <bcbake...@xxxxxxx> wrote:
"Tony D" <tonyisyour...@xxxxxxxxxxxx> wrote in message

news:1174173103.063176.163970@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

On Mar 17, 2:24 am, "Bruce C. Baker" <bcbake...@xxxxxxx> wrote:

Dawn, perhaps you could follow JOG's lead and set up your own MV theory
resourece?

I'm only half-kidding ...

She has. I'll let Dawn promote it for herself. Suffice to say some of
the comments on her site put her contributions here in fairly sharp
relief.

Are you referring to this:

http://www.tincat-group.com/index.html?

Looks more like a blog and some occasional pieces than an actual collection
of MV resources.

Yes, I have largely my own writings on my site and I have not written
very much. I do hope you enjoy my blogs, else the MV trilogy might be
of interest. I did a lot of research for the MV family tree poster.

I have interviewed many who have been in the MultiValue industry for
more than a quarter of a century, and I have several of the published
books on Pick in my library, right there next to Codd and Date. But
this is a group that has been heads down solving business problems
with their tools (when not coked up as I'll grant that the stories in
the history of Pick have included plenty of drugs, drinking, and
debauchery in addition to technology). There seems to be little
interest in the ivory tower.

Henry Eggers might be the closest the MV world has come to having
someone who has reflected over the years on theory related to Pick or
on Pick vs Relational. He says "Pick gives you enough rope to hang
yourself." This is a virtue in the PIck world. He does not do e-
mail, unfortunately, and I'm lousy at snail mail, so I have not had as
many conversations with him as I would like. If anyone in the LA
(Newport Beach) area is interested in talking to him, I do have
contact information.

Most old timers are aware of *** Pick's disparaging remarks about SQL
and many agree with him ("SQL is brain-dead" is one I was told). Many
have moved on to working with SQL DBMS's, however, picking up whatever
tools are appropriate for the work their company is doing. Those I
have talked to who know both do not tend to care about theory either.
In general, they like the tools available with SQL-based DBMS's but
the ease of development they had in the MV world. Few folks I have
talked to prefer 1NF (the form formerly known as 1NF) or 3VL.

When it comes to resources, the ACM portal has as many reputable
sources as anywhere that are related to MV, even if not specific to
it, with topics about functional databases (Pick BASIC is procedural,
but Pick files are handily modeled as functions, and some of what is
talked about as "functional databases" is relevent), di-graphs and
other graph theory-related models for data, objects, NF2 (not sure if
such papers are in that portal) etc.

If we look at the front lines of modern software development and also
at the pre-relational models that have survived through the relational
years, there is a lot of production work being done today using not-
exactly-relational data models, persisting data in something other
than the form formerly known as 1NF, for starters and using 2VL rather
than 3VL, thank goodness. But we also see set-based logic and less
iteration than was used in the 70's. So, the RM has had an impact
even if the ability to model data in multiple ways is more likely to
be the wave of the future (tossing aside the Information Principle).

It was very surprising to me to find out that Intersystems, with their
object implementation in addition to their MUMPs BASIC, SQL and other
interfaces, a company that advertises in Java & .NET cutting edge s/w
dev circles, was writing MV for Cache. Yes, they just wrote a new
Pick implementation for Cache data, can you believe it? You can write
data as objects and query it as Pick or SQL, for example. [If anyone
else is headed to their DEVCON, let me know.] I'm guessing that these
folks, too, are focussed on solving the problems of businesses much
more than on related theory, but I'll see if I can find any mumps
theorists at the conference at the end of the month.

At this point, I still have theory that doesn't translate into best
practices from what I can tell (relational) and data models and tools
for "agile" or just plain fast and maintainable software development
that doesn't have an adequate theory backing it (pick, mumps,
object). I'm selecting tools for a real project that I am funding and
I'm gonna have to go with the tools that yield the best value in
software development, whether backed by impressive theory or not. --
dawn

P.S. Thanks, Bob, I think I have misspelled "empirical" many times,
but I'll try to remember it now.

.