[Ocaml-biz] Let's choose a market

[Brian Hurt]

On Thu, 9 Sep 2004, Brandon J. Van Every wrote:

> > 1) Unix-centric C developers.
> 
> I think this is too vague to count as a market.  Unix and C are used for
> an awful lot of things.  Applications?  Data warehousing?  Web servers?
> Embedded?  Realtime OS?  Security, firewalls?  Rendering farms?
> Cryptographic challenges?  Academic research in conjunction with
> supercomputers?

Appplications- hell yes! (emphasis on things that make large scale code 
easier to write, debug, and maintain)

Data warehousing- maybe

Web servers- yes (emphasis on the extra security gaurentees ocaml gives)

Embedded/Realtime- maybe later.  I have concerns with GC & realtime (of 
course, swappable memory and realtime don't play well together).  Also, 
they tend to have too much hardware interface.

Security/firewalls- maybe, see webservers above wrt security.

Rendering farms, crypto- probably not.  These are two places where even a 
10% performance hit isn't acceptable.

Numeric computation/supercomputers- yes, because this is one place where 
advanced data structures can overcome small performance hits, and that 
ugly bits can get hidden away in libraries.

> 4) CAD

Definately yes.  This is a combination of applications work and numeric 
work- except the numeric work tends to really reward advanced algoritms 
and data structures.  Things Ocaml really rocks at.

> 5) simulation, engineering, FEM
> 6) scientific visualization

Yes- see above.

I'm really starting to think something Ocaml needs is a kick ass, take no 
prisoners, best of breed linear algebra library.

> 7) financial visualization
> 8) modeling, rendering, animation for film and TV
> 9) courtroom visualization
> 
> These are the other kinds of 3D graphics markets.  They happen on all
> major platforms.  Developments here would aid game development.  Parts
> of these markets don't require 3D graphics per se, but rather
> mathematics.  One could spend time converting Fortran developers.

One critical difference- the CAD/numeric people don't generally care all 
that much about single precision support, so that's less of an issue.

> 
> 9) embedded
> 
> This is probably more than one market.  I know little about such
> markets.

It splits into two broad categories- realtime and non-realtime.  I don't 
think we have much hope for realtime, ever.  Non-realtime we can probably 
make inroads into.  But I can't think of anything we can do to help across 
the board getting into embedded, other than get ocaml to cross-compile.

> 
> 10) enterprise computing
> 
> Java and C#'s meat and potatoes.  I doubt OCaml is ready to take them on
> in that space.

Without the marketing muscle of a major corporation (or several major 
corporations), I wouldn't expect a lot of movement here.  There is still a 
lot of Cobol in this space.

One thing we do need is something like tomcat for Ocaml to get anywhere in 
this market.


-- 
"Usenet is like a herd of performing elephants with diarrhea -- massive,
difficult to redirect, awe-inspiring, entertaining, and a source of
mind-boggling amounts of excrement when you least expect it."
                                - Gene Spafford 
Brian