Newbie switching from TCL

[Alex Martelli]

"Grant Edwards" <nobody at nowhere.nohow> wrote in message
news:slrn8q8ukt.nri.nobody at isis.visi.com...
> In article <8o1li30e3r at news2.newsguy.com>, Alex Martelli wrote:
>
> >> It's very strange, but every time I'm in a university computing
> >> department they laugh at and make jokes about Bill, when I meet
> >> commercial computer people they don't have the same attitude. It's
> >> probably a money thing.
> >
> >Funny, I've heard just as much laughter and jokes in commercial
> >development shops.
>
> The laghter and jokes are from the people using Unix.  The swearing comes
> from people using Windows in a commercial environment.

I've carefully observed development of our (commercial) software
under several environments over the years, and I assure you the
jokes (often bitter) and swearing has not changed much.  There
_was_ a maximum of swearing back when we had to release under a
dozen different Unix versions, not that many years ago, particularly
since we ALSO had to release on VMS, Apollo Domain, DOS, Windows,
OS/2, and for a slice of time also Hewlett Packard's old HP/1000
mini (I forget how that OS was named -- it was on its way out when
I joined this firm).  The fact that our main chosen language was
Fortran, with different incompatible dialects on each of these
platforms, didn't particularly help, either.

When people at the time came to me for consulting on how to do
something on all platforms, my response (being a mostly-Unix
guy, although expert on each of these platforms save HP/1000
and Domain) was almost invariably, "Now, let's see, under Unix
that would be easy, but..." -- actually, "Unix" was something
of an abstraction, as finding a solution that would work on
SCO's may-it-rot-in-Hades-forever so-called-Unix as well as
on pristinely-pure-BSD-4.3 (not to forget AIX, HPUX, SunOs,
etc etc ad nauseam) was often harder than covering theoretically
incompatible platforms like, say, VMS and DOS.  We ended up
with a big, thick, hard-to-maintain virtualization/compatibility
layer as well as a 'rulebook' two inches thick of things not to
even try because they'd inevitably break on SOME platform.

Compared to that state, developing and releasing for Win32 is
far less stressful (even though I'd just LOVE dropping kludgy
Win98 & friends in favour of just-NT and/or -2000, but, oh
well, marketing insists).  Some day, no doubt, the market will
once again force our hand and make us support some Unix variant
(I pray it will be just one, rather than a host of them:-), as
Linux in particular seems to be growing very nicely (maybe I'll
even be able to reinstall Linux at home, as I did back when it
was at 0.99 -- well, one can *hope*!-).  So, we keep (in theory,
at least:-) all system-specific stuff still in its own layer.

The singlest hardest issue today would be moving to a platform
without good componentization support, as more and more our stuff
gets deeply component-based.  Oh well, worst case, we can crib
XPCOM from mozilla.org, I hope:-)...


Alex