Re: Ada IDE for Win 95

[dewar@merv.cs.nyu.edu (Robert Dewar)]

Larry said

<<Some are free and some cost money.  But since we have you here, which
are the free IDEs for your chosen non-Ada languages and how would you
rate their quality?  My own experience with non-Ada IDEs has been with
those that cost money, and knowing about the others would be useful.>>

One problem is that the term IDE is used rather casually. To me it should
refer to something like the Apex environment, where you have a relatively
closed system that contains a complete set of tools aimed at effective
configuration and control and devlopment of large scale systems. An IDE
might or might not have a visual interface of some kind (these are really
two quite orthogonal issues -- integrated development environment, and
visual interfaces).

A lot of people use the phrase IDE to refer to any system, no matter how
simple and incomplete, that allows some kind of simple point and click
interface for compiling, linking etc.

TO me, this is a confusing use of the term, but increasingly that is what
it comes to mean. For example, something like the Brighton "IDE" for
Windows-95 is a handy little visual interface which may well be of use
for beginning students (at which it is aimed), but is not intended, or
usable, for serious large scale development (e.g. it does not begin
to be a full development environment, and at least in the version I
saw, has nothing at all to say about configuration management, which
seems to me to be at the heart of any real IDE.

On the other hand, one can certainly have an integrated development
environmnent that is not point-and-click oriented (indeed I can say
for my self that the *only* IDE that would possibly be of interest to
me is one which did NOT rely on such an interface, since I consider
them useless to me for serious development wor).

Now of course there are people who really like visual interfaces and
find them effective, and so visual IDE's certainly make a lot of sense
for this kind of person.

For me, I am always doing six things at once, and I find it clearer to avoid
any one of these things fragmenting itself into windows ... for example, I
find pmgdb, the PM version of GDB under OS/2 to be very impressive, and
you can use it to give very nice demoes of windows popping up with useful
stuff, clicking on variables etc, but I personally prefer to use gdb in
normal command line mode. 

As usual, different tools suit different folks. One of the advantages of
an open environment like GNAT is that you can tailor the tools you use,
and the way you use them, to match your own requirements. Some people in
the GNAT group are real EMACS fans, and in particular like the GDB
integration that it provides, and the Ada mode stuff, but it's not to my
taste -- which is fine, I choose different sets of tools (and indeed no
two people in Ada Core Technologies work with exactly the same set of tools,
or even on the same machine -- almost everyone is using different machines,
everyone uses what they feel most comfortable and productive with).

I do worry that we mis train students if we *only* present them with vey
simple ways of preparing programs that do not scale up. I think it is fine
to have simple systems when you are starting out (and for example, Mike
Feldman's simple GNAT interface for DOS seems very nice for starting
students), but I don't think this is all they should be introduced to!