Re: JGNAT users

["Rob Veenker" <veenker@xs4all.nl>]

"Marc A. Criley" <mcqada95@earthlink.net> wrote in message
news:3C10AB00.74E4DB85@earthlink.net...
> Torben Hoffmann wrote:
> >
> > Hi.
> >
> > I was wondering if anyone is actually using JGNAT and, if so, if they
have
> > made a web-page or some other documentation of their deeds.
> >
> > I am contemplating using JGNAT instead of Java since I am less than
> > satisfied with the software engineering level in Java compared to Ada -
> > especially maintainability and the ability to return to some code after
a
> > while and actually "get into to" fast are important issues for me.
>

I am using the supported version of JGNAT in order to be able to reuse code
that should run on Windows NT and on a PDA running a JVM. I have implemented
a communications application that connects ruggedised PDA's to our main
application. By using JGNAT I only needed to change a few Ada package bodies
specific to the JVM environment :-)
Nearly all of the problems I found are fixed already in the wavefront
version. My compliments to the ACT people !

> I used JGNAT 1.1p quite heavily for awhile to build the client,
> GUI-oriented portion of an application.  While the effort was
> successful--it runs and works well--I did have to work around some
> problems and idiosyncracies of JGNAT.
>
> Mind you, I did not have problems with approach taken by JGNAT to
> implement things like interfaces and inheriting from Java classes, but I
> did have some problems with the compiler.  A few months ago in this
> newsgroup I characterized JGNAT 1.1p as almost, but not quite,
> production quality.  There were certain constructs that JGNAT did not
> handle well, such as (as I recall from last spring) functions returning
> unconstrained arrays.
>
> There was nothing show-stopping, but you did have to code around such
> situations.  I was hoping that there would be a new public release
> addressing these problems (I did submit several error reports to ACT),
> but nothing has turned up yet.  (And for a one-person effort, commercial
> support isn't an option.)
>
> My take on using JGNAT is that if it's something relatively small, or if
> you're willing to be a technology explorer, go ahead with it.
>

I can't comment on the public version of JGNAT, but I don't feel the
supported version is that restrictive ;-)
I assume this technology will also become available in the public version.

> Aonix also has an Ada-to-JBC compiler, but I have no experience, or
> knowledge of other's experience, with it.
>
> Marc A. Criley
> Senior Staff Engineer
> Quadrus Corporation
> www.quadruscorp.com


Rob Veenker
Dutch Ministry of Defense
RNLA/DMKL/C3I/C2 Support Centre