Re: Can Ada by popularized faster ?

[No Spam <Die_Spammer@flinet.com>]

Dave Wood wrote:
> 
> I find your letter interesting in that it provides an eclectic
> mix of accurate deductions, naive assumptions, positive
> spirit, and nascent defeatism.  It's probably a decent
> microcosm of the "Ada community"!  I have a few comments to
> contribute.
<lots of good stuff snipped>

Here's an idea which I have suggested before and would do myself were I
to have the available resources:

Think of why C got so popular: The language itself wasn't the issue - it
was all the other things that were bundled around it. The Unix OS, Lots
of platforms it was bundled with, availability on small systems for use
in Universities, Single Board Computer manufacturers targeting compilers
to their specialized little computers, etc. etc. Pretty soon it was the
volume of applications that got it entrenched.

I don't think anyone is going to go out and write the next Great
American Operating System in Ada any time soon, but if they did, Ada
would quickly get a toe hold on a large market. Developing some similar
important million-user-app would be as effective, but what would it be?

My personal interest would be to get a relatively inexpensive single
board computer that had Ada riding on a PC platform with the support
stuff needed to plug my code into the SBC and drive it around. Boards
like this regularly sell for a couple of hundred dollars and typically
have the development environment bundled in at some nominal additional
cost. (This means you get a C compiler, linking, loading and debugging
software, interface to the PC from a serial port and some other goodies)
So SBC developers have latched on big time to C just because thats what
they got when they bought into the board architecture.

What if the guys with the compilers were to put together an SBC - or
simply get into a partnership with an SBC manufacturer to bundle the
compiler with the development kit and collect some royalty? If the board
is popular *and* the development environment is really spiffy, providing
the stuff an embedded developer needs, maybe Ada gets to ride along with
somebody's industrial controller.

If you supply a run time kernel and maybe a bunch of hardware interface
code with the compiler and make it easy for some developer to include
all this software in the product, you could maybe get a small royalty
for every computer that drives around a Dodge Neon. Programmers may be
reluctant at first but if you go in the door with the pitch: "Let me
show you how much quicker you'll get to market using my compiler, my
RTK, my hardware interface code, my executive, my realtime monitor..." -
all that on top of the proven advantages of Ada for reliability, etc.
and you may win a bunch of converts.

If it isn't the SBC idea that excites you, then think of something along
a similar line. What if there were a spiffy data base out there which
was written in Ada and had all of it's programmer hooks specified in
Ada? Get enough people putting their address books into your runtime
database and developers of other software will want to build with
bindings to your runtime. Or a word/document processing package that had
a callable interface written in Ada so that end users find the product
to be hip and developers of other applications want to call your
software whenever they need to manipulate text?

I think there's some reasons Ada has an edge - it's use in college level
courses represents a big buy-in of future developers. All it needs is
some additional push because of an installed base of users of some
application other than the compiler itself.

Marin David Condic
CONDICMA@DONT.SEND.ME.SPAM.PWFL.COM