Well, I am distinctly pleased that the message wasn't full of questions for
Dr. Carlisle, since the DoD is a big place and I'm not sure I could answer
them all. However, since I'd hate to be lumped in the category of
non-respondent DoD staff, I'll attempt as best as I can.
1) I am not aware of any funding for JGNAT by the DoD
2) Well, for one, I did. Don't know of anyone else in DoD that was
involved, but I certainly tried to assist as much as possible by doing early
testing, and providing a version of RAPID for the JGNAT compiler.
3) I think the failure of JGNAT was not technical, but demand-based. If ACT
had a customer base that wanted the technology, you can bet they'd proceed.
Follow the $$. Lately I've spent a fair amount of time in the JGNAT
sources, and they are quite technically sound.
4/5) I think you'd be hard-pressed to identify "parts of the DoD" that are
for or against Java. Within my own department there are different opinions
on Java technology. The JVM hasn't caught on nearly as well as Sun would
have liked, so it is understandable that a much smaller market (Ada
compilers for JVM) hasn't flourished. Personally, I found the speed of Java
implementations, the kludgeyness of some of its libraries, and the security
failures to be quite disappointing.
The DII COE is, in my opinion, an attempt by the Department of Defense to
provide some uniformity in its computing platforms, much as Ada was to
provide uniformity in programming language. I think that it will be as
successful as the Ada mandate (i.e. not very).
DISCLAIMER: At this point, I'm sure the Public Affairs office would want me
to point out that all opinions specified above are mine, and do not
necessarily reflect the opinions of the US Air Force Academy, the US Air
Force, or the US Department of Defense.
FYI, at SIGAda 2002, I will be presenting a very similar project (not
supported by ACT or any grand "part" of the DoD) which compiles Ada to .NET,
and is largely based on the JGNAT sources.
Martin C. Carlisle, PhD
Associate Professor and Director of Advising
Department of Computer Science
United States Air Force Academy
From: Craig Carey
To: [log in to unmask]
Sent: 8/18/02 11:21 AM
Subject: JGNAT 'Ada 95 to bytecode' compiling project ends
I ask for some more information on who wanted JGNAT, and when they
over to a view agreeing with ACT which was it had bugs.
At 31 Jan 2001 09:29 -0700 Wednesday, Carlisle Martin C Dr USAFA/DFCS
>You should certainly look into JGNAT, which allows an Ada programmer
>access all of the Java classes (including Swing, e.g.) and compile Ada
>programs into JVM byte code (so they have all of the same
>capabilities). See http://www.gnat.com (select products, then JGNAT)
>Download (free) from: ftp://ftp.cs.nyu.edu/pub/gnat/jgnat/jgnat-1.1p/
ACT pulled the plug on the JGNAT project, so said a messages at
It is quoted below. (This message is not full of questions for Mr
responses from the DoD staff have been brief.).
Here is a copy of the Usenet message:
>From: [log in to unmask] (Robert Dewar)
>Subject: Re: JGNAT status??? (was: Re: Help: Problem with JGnat)
>Date: 3 Aug 2002 07:36:06 -0700
>"Dr. Michael Paus" <[log in to unmask]> wrote in message
news:<[log in to unmask]>...
>> Yes, a new public version of the current JGNAT sources
>> could be great.
>We have no plans for any further work on JGNAT
>?> if it is true that a version exists which runs on Java
>It is false. There were indeed filed bugs in this area, but
>they were never solved.
>> Maybe we can convince ACT to consider
>What is "this"? Doing more work on JGNAT? That's not going
>to happen. The sources are out there, the first task would
>Ada Core Technologies
* Did the DoD fund the ACT JGNAT project?.
* Who was in the DoD that wanted the JGNAT system?.
* Will it be revived with such force and an excess of expenditure that
JGNAT is made to run properly. I presume ACT argued strongly before
JGNAT was started, that it would fail for sure, and that is what ACT
is now saying. So it sounds like the DoD wanted it. [I saw something
identifying the DoD and I can't find it now.]
* Which part of the DoD was pro-Java/JVM ?. I don't seem to have that
sort of detail here.
* As seems likely, the DoD opinions of Java and JVMs may have worsened
over time. Has anybody got comments on which parts of the DoD held
which views, and when, as Java and buggy software become less and
credible inside of the DoD ?.
I do not have quotable comments saying that the DoD was even interested
in JGNAT. I hope to not mislead if wrong on that point.
Is there anti-Java DoD that can be requested under US Freedom of
Information law ?.
This next message of 2001, indicated some serious interest in putting
JVMs into military machines. Or perhaps the DII COE made a mistake
At 19 June 2001 07:51 -0400 Tuesday, Marc A. Criley wrote:
>About a year ago I discovered that the DII COE Integration and
>Specification (I&RTS) 4.0 prohibited the use of tools that compiled
>languages other than Java into Java Byte Code (JBC). This obviously
>precluded the use of Ada products such as JGNAT and AppletMagic to
>"portable Web-based" applications. (The use of analogous compilers
>other languages, such as Eiffel and Python, were therefore barred as
>Here's the original prohibition:
>"Developers shall not use compilers designed to convert code developed
>other languages (e.g., Ada, C++) to create Java byte-codes. This
>restriction is important because such compilers may inadvertently
>intended Java security features." -- Section 8.2.3, bullet 3.
>As I mentioned, having never received any feedback, I have no idea
>role our submission played in the descision to remove the bullet.
>major, perhaps none, perhaps it was the Python advocates that got it
>:-) Whatever the reason, we did our part, and I'm pleased with the
>things turned out.
>Marc A. Criley
>Senior Staff Engineer
>P.S. A plaintext version of the submission is available in the Team
[ Subject: Recommended DII COE change to permit Ada (etc) to Java
Date: Mon, 10 Apr 2000 10:27:12 -0400 ]
It seems like the DoD is adapting its views about Java. At some point in
time they would attempt to discourage its use. Has that happened
Also, the GNU Java compiler project is not proceeding excellently too.
This if ACT won't do the coding it could be an Ada project with no
possibility ending up being used. With this clearly failed Ada 95
who were the people or agencies that wanted JGNAT ?.
I can't recall a webpage listing Ada 95 failure stories. I have not
looked at the http://www.adaic.org/ website for comments on this topic.
Ada mailing lists listed: http://www.ijs.co.nz/ada_95.htm