Wed, 1 Dec 1999 02:08:38 -0600
Geoff Bull wrote:
> I am surprised by your defeatist attitude.
For better or worse, Richard Conn is right about
trying to compete with Microsoft on its main turf.
Believing that Ada can have an impact on mainstream
PC development these days is like believing that the
Pricipality of Liechtenstein can mount a successful
land invasion of the United States. MS will continue
to be unassailable on the PC for at least another decade,
whatever the final outcome of the 'monopoly' trial and
in spite of the zooming stock price of RedHat.
On the "server" side of the internet, the situation is
not as bleak, but almost. Entrenched forces and billion
dollar investments there ensure that the development of
that software will be done for the most part using the
Unlike Richard Conn, however, I don't think I'm
whistling Dixie when I say that Ada can be a player
in the *real-time* market -- the larger bubble in
the Venn diagram that includes the 'safety-critical'
circle. The reasons are that the real-time market is
not so well-defined as the one for PC applications,
and that Ada has at least a sliver of mind-share in
"Grand schemes" of developing Ada replacements for
Windows and/or Linux should be discarded. The Ada
Dream -- all software in the world well-designed,
coded and commented in pure Ada -- is long dead.
"Success" for Ada now means (1) making sure that it
stays in use for current projects (a considerable
task in itself); (2) seeing to it that it continues
to be chosen for future projects in its own current
domain -- safety-critical systems; and (3) expanding
into some other subsectors of the real-time market.
The last of these is crucial because it has the
most significant potential for *growth*. I don't
think the 'safety-critical' market has the growth
curve Ada needs to thrive.
Any other here-and-there usages of Ada are pure
serendipity. Meaning that there will be no
sustainable *economy* in those areas.
This does not mean that if you want to create a
great new application for the internet that you can't
or shouldn't do it in Ada. It just means that if it's
a programming product -- one that requires its users
code in Ada -- it is doomed. Even Liechtenstein knows
better than to mount its canons at the US border.
mailto:[log in to unmask]