Re: Objective Heresy

Mark Baker (markb@iosphere.net)
Tue, 3 Feb 1998 11:28:27 -0500 (EST)


I suppose with Ron as an ex-FoRKer (for now), I'm one of the few left at the
altar of the Church of Objectology.

On Mon, 2 Feb 1998, David Crook wrote:
> Objects, CORBA, IIOP, Javabeans, ActiveX, DCOM -- Its all hype.

You take that back! 8-)

> something in this field. I know that this email probably should have
> been sent to dist-obj, but I've just started reading that list and
> besides, bringing up the really tech stuff on here seems to tick off
> Tim.

Oh, I doubt that.

> I hear lot about this stuff, some of it on here. I hear that its the
> next big thing, the wave of the future, etc. I think back to people
> saying the same thing about VRML and virtual reality. Heck I even
> talked to Worlds official VRML Evangelist (that was her official job
> title, no joke). It sounded really great at the time. They went under
> a few months back. It sounded really good, but no one was buying
> (actually, it wasn't that there was nobody, it was that there wasn't
> enough).

Hence the need for evangelists - *more* evangelists. I consider myself a
purveyor of the *right* tools and ideas, and hope that customers will
come. I've resigned myself to the fact that I'll never be a millionaire
simply because I can't bring myself to give people what they want - I
*need* to give them what they need. Anything less is selling out.
(anybody see the movie "Big Night" with Stanley Tucci?)

Slowly but surely though, the gap is narrowing. CORBA, DCOM, Java, and
the Web are quite mainstream nowadays. They may not be perfect, but
they're here and in our faces. Sure, 97% of the population doesn't know
what to do with them, but that'll change in time - once there's something
better for the 3% to think about. 8-)

> I've been doing this programming thing for about 6 years now. Basically
> it comes down to this, I write some code for people, they pay me money.
> Its a pretty simple system. I've watched C kinda fall by the wayside in
> favor of C++ for most application development. The object oriented
> programming method seems to be a real good fit for most app development,
> especially for windows stuff. Using C++ class libraries sure beat
> having a gigantic, ugly while loop in your code. So, don't get me
> wrong, I'm not one of those programmers that, for lack of a better word,
> are frightened about object oriented stuff. I just don't see the big
> deal.

Step back. Go meta.

Ask yourself what "app development" is. Why is it done? Why does
practically every new set of requirements that comes down the pipe
require a new "app"? And why are all these apps silos? Why don't they
interoperate? Isn't this Internet thing about sharing and
collaboration? Shouldn't we really take that to heart?

> work. Most of the jovial people stayed on the bitter end. It taught me
> something, keep up with the new stuff or practice saying "Do you want
> fries with that". Its the primary reason I went back to graduate
> school, I figured that my Unix/C background was reaching the far end of
> "Current" and was starting to slip into "Outdated".

I realized that fairly late in the game - 1994 - but I got the hell out
of Unix/C/RDBMS as soon as I could.

> So I like to keep up on the newer stuff. Now I'm doing a research
> class, which may eventually grow up into a thesis. So I figure its a
> good excuse to check out this object stuff. I don't believe in it now,
> but I just want to check it out to be sure. You got to keep an open
> mind about things. So I've checked out dist-obj a bit and I've
> downloaded a few papers. I'm ready to be convinced.

Good.

>Maybe some of you
> people on here can point me to some stuff that will convince me that
> this wasn't all made up by marketers.

I know it's really boring (and expensive), but I still consider Oliver Sims'
book to be the definitive work on why and how we need to move away from
"apps". Even his "far out" SDS (semantic data stream) stuff.

http://www.amazon.com/exec/obidos/ISBN=0077079574/forkrecommendedrA

He's got another book coming out next month;

http://www.amazon.com/exec/obidos/ISBN=0471191760/forkrecommendedrA

>I just don't see that much
> difference between having a object library and just having a regular
> code library. You call some routines that someone else wrote, wow.
> This talk about writing re-usable components, yawn, people have been
> saying that one for a few years now. Distributed objects, you mean like
> downloading other peoples code on the fly and running it, like Active
> X? Excuse me while I run away in horror.

You don't necessarily have to download code, though it obviously helps in
some cases (for all those agent-ish reasons). You might just want to
send a message.

> Am I just not understanding this stuff? Is there some major great new
> thing that I just am not seeing?

It sounds like it. Read the book.

I could also point you to Infospheres, but unless you really *get* that
business objects == agents == components == documents, and more
importantly that a set of these distributed objects comprise an entire
information system (no "apps" required), you probably won't appreciate
the relevance of its goals.

MB

--
Mark Baker, Ottawa Ontario CANADA.                Java, CORBA, XML, Beans
http://www.iosphere.net/~markb               distobj@acm.org  ICQ:5100069

Will distribute business objects for food.