[Ocaml-biz] IDEs

[Brandon J. Van Every]

Brian Hurt wrote:
> Brandon J. Van Every wrote:
>
> > Let's say, for sake of argument, I wanted to actively
> > prevent you from
> > using Vi to program in OCaml.  How would I do that?  I
> > don't see a way
> > for me to do that.  So on this point, what's to discuss?
>
> 1) Build the editor, build tool, and compiler into one huge integrated
> application (scheme).  Trying to combine the build tool and
> the compiler (Java) wins you negative points.

Eclipse has a plugin architecture, that's how the OCaml support is being
provided, so not our concern.

> 2) Require the source code to be something other than plaintext (APL)

Doesn't apply to OCaml, so not our concern.  Could apply to a visual
programming language, like for instance Lava.
http://lavape.sourceforge.net/index.htm

> 3) Selling my manager on the idea that everyone in
> development should be
> required to use the same editor, to prevent "compatibility problems".

That's not a real issue.  Your manager could jolly well mandate just
about anything.  If you don't like your manager, you either pick your
battles, change your job, become the manager, or become self-employed.
In the latter case, you may still have to use editors that various
clients mandate, for their own sanity.

There's no way to prevent you, as a UNIXen who likes a large variety of
tools, from doing what you want to do.  You might be politically against
the idea of generating critical mass towards a narrower field of tools,
and unwiling to support such an effort, but such efforts can't actually
stop you from doing anything.  They can only convert people who want
one, easy to use toolset.  In short, I'm saying the strategy should be
to convert the abundance of Java and Windows guys who prefer this kind
of tool package.

I think to convert UNIXen who like lotsa different tools, a different
strategy is required.  I don't know what that strategy would be.  Open
source UNIXen seem to oppose 'strategy'.

> You can be hostile enough to other development
> environments that almost no one will use them.

Well, I'm not suggesting a misinformation campaigns, like "GNU Emacs,
XEmacs, and Vim are bad."  I think it would be perfectly reasonable,
however, to pitch Eclipse as better integrated for OCaml than any of
those, if indeed it is true or can be made true.  Some UNIXen will
scream it's all not true, a bunch of lies, beat their chests about it,
etc. but people with strong knee-jerk reactions like that don't matter.
Most UNIXen will simply ignore the glossies and do what they want, as
you said.

What you call 'hostility', I'd call "focusing on supporting one IDE
well, and acknowledging the limited resources of ocaml-biz."  Now of
course, in open source everyone does what they want.  You could choose
to champion a slick Vim installer or whatever.  But I'm going to work on
what I think is strategically most likely to get a large volume of
programmers doing OCaml.  I suspect that's Eclipse.

> Sigh.  Actually, I haven't even been addressing the issue of
> how to market to Linus Torvals and Richard Stallman.

Richard Stallman cannot be marketed to.  Do you think Linus Torvaldis
needs something that OCaml can provide?  I'm serious, I mean Linus
himself.

> Dumb managers have gotten quite enchanted with the idea of
> shipping programming jobs overseas.  They're the sorts of
> managers I was saying we have no chance with anyways.

Well, I don't agree.  I have a terribly optimistic view about our
ability to manipulate dumb people, and I'm not afraid to try.  One might
note the nature a Shaper.  http://www.teams.org.uk/shaper.htm  I'm not
interested in pursuing OCaml goals solely according to 'utility' or
'morality' or 'technocracy'.  I live in a town where most people
understand the merits of steamrollering the stupid or lazy with a good
marketing campaign.  I'd press every available marketing method into
service to reach my goals, short of lying to people.  That's where I do
draw the line.  But as far as I'm concerned, if a crowd is smart, work
with that.  If a crowd is stupid, work with *that*.  You tailor the
message for the different crowds.  I have no problem whatsoever with
using lotsa Java and C# buzzwords for the consumption of PHBs, so that
they start believing they're "missing The Next Big Thing" or some crap
like that.  That's what the Marketing tool is there for.

> Because they aren't deciding based on the technical facts.

Marketing is a toolbox.  Sometimes, you get people to decide things
based on the technical facts.  Other times, you get people to decide
things based on how snazzy the logo looks, or something equally stupid
from a techie's narrow world view.  I say, if it works, use it.  Just
decide where your ethical boundaries are.

> They're deciding on things we, the Ocaml
> community, have no control over what so ever.

Not true.  We have plenty of opportunity to take control over all kinds
of marketing perceptions and technological realities.  It is, of course,
hard work.  We get what we pay for.

> Linux didn't start with that level of labor either.  It
> started with one kid in a college dorm, and grew from there.

Unfortunately for us, Linus tapped into a totally unpenetrated space.
There are many competing open source languages and 3D engines out there.
If you can think of a totally unpenetrated space, that would be way
cool.  I can't offhand.


Cheers,                     www.indiegamedesign.com
Brandon Van Every           Seattle, WA

Taking risk where others will not.