RE: Ada The Best Language?

["Beard, Frank" <beardf@spawar.navy.mil>]

-----Original Message-----
From: Warren W. Gay VE3WWG [mailto:ve3wwg@home.com]

> I'm not certain the "itegration" issue is that big a deal. Granted, it
makes
> purchasing/acquisition simpler. I think developers on Windows systems
think
> more along these lines, only because they are used to Microsoft/Borland
etc.
> bundling their products along these lines.

As a former VAX and Unix developer, I think along those lines.  Maybe
it's because I'm currently a Windows developer, but I could have sworn
I was thinking that way before I moved over. ;-)

> In the UNIX world however, I don't think this thinking is quite so
mainstream.
> After all, you want a compiler that fits your
cost/reliability/conformance/
> validation requirements. You choose your source control tools on the basis
> of standards/company mandate/personal preference etc. You choose your
editor
> often on religious grounds. You choose your debugger on cost/productivity
> basis maybe.

Yes, but a complete bundled package sure would be nice, especially for 
students, novices, and beginners.

> In short, you choose the set of tools that work best for you. Many use vi
> in the UNIX world, for whatever reasons. Others use GNU emacs, or elvis.
> I've personally always used a heavily customized version of MicroEMACS.
> But within a group of UNIX developers, you are likely to find just as
> many preferences ;-)

Well, maybe if your choice of tools could be easily substituted it would
be more appealing to you.  But if the bundle was complete and cost 
effective enough, it might not even be an issue.

> Having said all that, I do know that a large portion of these developers
> like the IDE approach (if given the choice). I myself do not like them, 
> but then, maybe I'm strange that way.

Well, since you use Ada, I'll still talk to you. ;-)

> I find that the combination of my own modified editor,
> custom tools, command line editing (emacs mode of course) etc., allow
> me to be much more effective than any IDE has allowed me to be. Make files
> do the rest.

I just don't like the functionality being so disjoint.  I prefer an
all-in-one tool that automates many things for me in a default kind
of way.  If I need something special, there is usually a way to change
the default.

I started out in the DEC VAX environment and using their LSE editor,
eventually.  Then I got pulled into the Unix environment using vi
(kicking and screaming - about vi).  Then I found Emacs, which was
a huge improvement over vi.  Then I worked on various flavors of
Unix.  Inevitably, I would end up on a Unix that didn't have an
available Emacs.  Here we go back into vi (kicking and screaming),
or try a native GUI editor.  Usually the native editor wasn't bad
but certainly didn't exist anywhere else.  Fortunately, they were
usually very similar to the GUI editors on other flavors.  Then I
ended up on Windows.  Moving between a word processor, spreadsheet,
IDE development environment, or just about any other GUI app,
regardless of vendor, is nearly seamless.  One reason we chose
Aonix ObjectAda was because of it's ease of use and similarity to
VC++ and Delphi.  Another think I like about Aonix is the ability
to create a project (similar to most of the other Windows tools)
and specify which files belong to the project.  They can be in 
various directories, or I can pick out which ones I want in a 
particular directory, or click "Add All".  And I don't have to
worry about the external name versus the internal name.  By that
I mean you don't care what the file extension is.  It can be
My_File.Shazbot, and the IDE won't care so long as the source
code in the file compiles.  Out of shear sanity, you want to
keep the filename the same as the internal source code, though
the IDE doesn't require it.  Anyway, enough of that.  I don't 
want to start and IDE versus Emacs flame again.  I've been on
Emacs.  Yes, it's powerful, but I don't want to go back.  No
other tools look like it (for bettor or worse, and no matter how
inferior they are to Emacs).

> Yet I grant that others do like IDEs, and perhaps perform better that
> way.

I seem to.

> I personally don't see this "integration matter" as the issue. Management 
> doesn't argue against it at this level -- they site the cost of finding 
> Ada developers, the fact that it is "unusual" or "not popular". They also 
> site that we cannot train people on "Ada" because they themselves may not 
> want to be in that area as a career choice. These are the types of issues 
> I see and hear.  Rarely is the resistance based upon technical/packaging
> details.

While I understand what you're saying, having come from a similar
background, I think the lack of a complete/integrated environment
is exactly the reason you have the shortage of Ada developers.  I
like Marin's idea of the "Red Hat" approach to bundling a complete
and powerful environment.  If someone had the time and inclination,
they could bundle a package for Unix and one for Windows that
contained something like (forgive me if I get the tools confused):

- GtkAda
- GLIDE
- GLADE
- ODBC bindings
- a database
- COM/DCOM
- CORBA
- etc (whatever I'm forgetting)
- documentation and examples

I think it would make an incredible difference.  Being one of those
that like the IDE approach, I see it all the time in the Windows
environment.  You can't tell me so many migrated to Windows because
they like Bill Gates and just want to make him rich.  I have no desire
to start an OS flame, but Windows is much simpler for the end user and
for the developer for general applications.  Just look at the suite of
tools available from Micro$oft & Borland (Inprise).  Every time our
Delphi people want to implement something there is already and API or
interface available in the tool.  All they have to do is call it.  I'm
just amazed how fast those guys can crank things out.  I want to be
able to do that in Ada, but it would take me weeks if it weren't part
of my environment.

Just me 0.02, as well. :-)

Frank