Porting Ada to C (Stealth development)

Porting Ada to C (Stealth development)

[Dale Pontius]

I am taking up some new work, and would like to do it in Ada. However
there is a strong 'Do it in C so others can pick it up' bias, which I
guess makes some sense. So I'd like to do my development in Ada, and
have the occasional conversion to C to play well with others. This
probably needs to be free, as well, since it wouldn't cost money if
if I just started the project in C. (Can't justify) I did a quick scan
at www.adahome.org without luck. I'll also need to link in C libs, but
that should be straightforward and in the documentation.

At the moment, I'm not going to fight any sort of C vs Ada battles,
and I'd just as soon keep it completely under the table. Iff I start
to show spectacular developmental speed and track record for bugs,
then I might surface it. I may have managed to win a minor VHDL vs
Verilog fight for part (but not all) of the project, because Verilog
couldn't describe everything that needed done as simply as VHDL. I'm
not going to push my luck.

So for the moment, I need an Ada to C converter, and comments on how
human-readable the resulting code is.

Thanks,
Dale Pontius

Re: Porting Ada to C (Stealth development)

[Dan Andreatta]

pontius@btv.MBI.com.invalid (Dale Pontius) wrote in 
news:a26o04$jfc$1@news.btv.ibm.com:

> I am taking up some new work, and would like to do it in Ada. However
> there is a strong 'Do it in C so others can pick it up' bias, which I
> guess makes some sense. So I'd like to do my development in Ada, and
> have the occasional conversion to C to play well with others. This
> probably needs to be free, as well, since it wouldn't cost money if
> if I just started the project in C. (Can't justify) I did a quick scan
> at www.adahome.org without luck. I'll also need to link in C libs, but
> that should be straightforward and in the documentation.
> 
> At the moment, I'm not going to fight any sort of C vs Ada battles,
> and I'd just as soon keep it completely under the table. Iff I start
> to show spectacular developmental speed and track record for bugs,
> then I might surface it. I may have managed to win a minor VHDL vs
> Verilog fight for part (but not all) of the project, because Verilog
> couldn't describe everything that needed done as simply as VHDL. I'm
> not going to push my luck.
> 
> So for the moment, I need an Ada to C converter, and comments on how
> human-readable the resulting code is.
> 
> Thanks,
> Dale Pontius


If you can live with C++, you could take a look at www.ada2cc.com.

Re: Porting Ada to C (Stealth development)

[Jim Rogers]

Dale Pontius wrote:

> I am taking up some new work, and would like to do it in Ada. However
> there is a strong 'Do it in C so others can pick it up' bias, which I
> guess makes some sense. So I'd like to do my development in Ada, and


Actually, it makes no sense, but opening that can or worms begins the
language war. If a person is a competent C programmer he or she should
be able to learn a new language with minimal effort. My own experience
is that it was easier to learn Ada, knowing C first, than it was to
read and understand a lot of C programs.

I am now in the business of training people in computer languages.
I have little patience for the argument that learning a new language
is a significant barrier to productivity.


> have the occasional conversion to C to play well with others. This
> probably needs to be free, as well, since it wouldn't cost money if
> if I just started the project in C. (Can't justify) I did a quick scan
> at www.adahome.org without luck. I'll also need to link in C libs, but
> that should be straightforward and in the documentation.
> 


When you say "play well with others", do you mean that others will
take over the maintenance of the C pieces, or simply that others will
want to call the Ada libraries from C? These are really two different
issues, and should be handled separately.


> At the moment, I'm not going to fight any sort of C vs Ada battles,
> and I'd just as soon keep it completely under the table. Iff I start
> to show spectacular developmental speed and track record for bugs,
> then I might surface it. I may have managed to win a minor VHDL vs
> Verilog fight for part (but not all) of the project, because Verilog
> couldn't describe everything that needed done as simply as VHDL. I'm
> not going to push my luck.
> 


Is there already any history of mixed-language development for your
team? This includes such pieces as shell scripts, perl scripts,
C++ libraries, or assembler routines. My guess is that your product
already uses some mixture of C and other languages. This situation
is not currently a political problem. Try to deal with Ada in this
light.


> So for the moment, I need an Ada to C converter, and comments on how
> human-readable the resulting code is.


Good luck.

Jim Rogers
Colorado Springs, Colorado USA

Re: Porting Ada to C (Stealth development)

[Dale Pontius]

In article <Xns9199960B37554andreattamailchemsce@12.253.140.251>,
        Dan Andreatta <andreatta@mail.chem.sc.edu.REMOVEME> writes:
> pontius@btv.MBI.com.invalid (Dale Pontius) wrote in
> news:a26o04$jfc$1@news.btv.ibm.com:
>
>> I am taking up some new work, and would like to do it in Ada. However
>> there is a strong 'Do it in C so others can pick it up' bias, which I
>> guess makes some sense. So I'd like to do my development in Ada, and
>> have the occasional conversion to C to play well with others. This
>> probably needs to be free, as well, since it wouldn't cost money if
>> if I just started the project in C. (Can't justify) I did a quick scan
>> at www.adahome.org without luck. I'll also need to link in C libs, but
>> that should be straightforward and in the documentation.
>>
>> At the moment, I'm not going to fight any sort of C vs Ada battles,
>> and I'd just as soon keep it completely under the table. Iff I start
>> to show spectacular developmental speed and track record for bugs,
>> then I might surface it. I may have managed to win a minor VHDL vs
>> Verilog fight for part (but not all) of the project, because Verilog
>> couldn't describe everything that needed done as simply as VHDL. I'm
>> not going to push my luck.
>>
>> So for the moment, I need an Ada to C converter, and comments on how
>> human-readable the resulting code is.
>>
>> Thanks,
>> Dale Pontius
>
>
> If you can live with C++, you could take a look at www.ada2cc.com.

I can live with conversion to C++, but I can't live with commercial.
Using Ada will be a stealth issue. Anything above-board would have to
be C, C++, or perhaps Java.

Dale Pontius
(NOT speaking for IBM)

Re: Porting Ada to C (Stealth development)

[Dale Pontius]

In article <3C47375B.8060604@worldnet.att.net>,
        Jim Rogers <jimmaureenrogers@worldnet.att.net> writes:
> Dale Pontius wrote:
>
>> I am taking up some new work, and would like to do it in Ada. However
>> there is a strong 'Do it in C so others can pick it up' bias, which I
>> guess makes some sense. So I'd like to do my development in Ada, and
>
>
> Actually, it makes no sense, but opening that can or worms begins the
> language war. If a person is a competent C programmer he or she should
> be able to learn a new language with minimal effort. My own experience
> is that it was easier to learn Ada, knowing C first, than it was to
> read and understand a lot of C programs.
>
Doesn't matter. This is a rather conservative area, and "C is the way
to program, though perhaps C++ or Java would be OK." Note that this is
also not a programming area, it's silicon design. Sometimes you have
to turn part-time programmer to get the silicon out. This is one of
those.

...
>> have the occasional conversion to C to play well with others. This
>> probably needs to be free, as well, since it wouldn't cost money if
>> if I just started the project in C. (Can't justify) I did a quick scan
>> at www.adahome.org without luck. I'll also need to link in C libs, but
>> that should be straightforward and in the documentation.
>
> When you say "play well with others", do you mean that others will
> take over the maintenance of the C pieces, or simply that others will
> want to call the Ada libraries from C? These are really two different
> issues, and should be handled separately.
>
One perpetual concern is, "How does this stuff get maintained when
XXXX moves on?" For that reason, they want a mainstream language,
especially considering (as mentioned above) the programmers here do
it on the side, not as a profession.
>
...
> Is there already any history of mixed-language development for your
> team? This includes such pieces as shell scripts, perl scripts,
> C++ libraries, or assembler routines. My guess is that your product
> already uses some mixture of C and other languages. This situation
> is not currently a political problem. Try to deal with Ada in this
> light.
>
Obviously yet. Mix of C, C++, ksh, perl, rexx, and I've thrown Python
into the mix. But I suspect Ada is too far out for acceptance. I had
a tough enough time getting VHDL in where needed, and avoiding a
Verilog-only mandate, but only because Verilog couldn't have easily
done that part of the job. The bulk of the HDL is still Verilog, and
the VHDL will get translated after doing its special part.

As I mentioned on the other subthread, Ada to C, C++, or Java would
probably be equally acceptable.

Dale Pontius
NOT speaking for IBM

Re: Porting Ada to C (Stealth development)

[Larry Hazel]

Dale Pontius wrote:
> 
> In article <3C47375B.8060604@worldnet.att.net>,
>         Jim Rogers <jimmaureenrogers@worldnet.att.net> writes:
> > Dale Pontius wrote:
> >
> >> I am taking up some new work, and would like to do it in Ada. However
> >> there is a strong 'Do it in C so others can pick it up' bias, which I
> >> guess makes some sense. So I'd like to do my development in Ada, and
> >
> >
> > Actually, it makes no sense, but opening that can or worms begins the
> > language war. If a person is a competent C programmer he or she should
> > be able to learn a new language with minimal effort. My own experience
> > is that it was easier to learn Ada, knowing C first, than it was to
> > read and understand a lot of C programs.
> >
> Doesn't matter. This is a rather conservative area, and "C is the way
> to program, though perhaps C++ or Java would be OK." Note that this is
> also not a programming area, it's silicon design. Sometimes you have
> to turn part-time programmer to get the silicon out. This is one of
> those.
> 
> ...
> >> have the occasional conversion to C to play well with others. This
> >> probably needs to be free, as well, since it wouldn't cost money if
> >> if I just started the project in C. (Can't justify) I did a quick scan
> >> at www.adahome.org without luck. I'll also need to link in C libs, but
> >> that should be straightforward and in the documentation.
> >
> > When you say "play well with others", do you mean that others will
> > take over the maintenance of the C pieces, or simply that others will
> > want to call the Ada libraries from C? These are really two different
> > issues, and should be handled separately.
> >
> One perpetual concern is, "How does this stuff get maintained when
> XXXX moves on?" For that reason, they want a mainstream language,
> especially considering (as mentioned above) the programmers here do
> it on the side, not as a profession.
> >
> ...
> > Is there already any history of mixed-language development for your
> > team? This includes such pieces as shell scripts, perl scripts,
> > C++ libraries, or assembler routines. My guess is that your product
> > already uses some mixture of C and other languages. This situation
> > is not currently a political problem. Try to deal with Ada in this
> > light.
> >
> Obviously yet. Mix of C, C++, ksh, perl, rexx, and I've thrown Python
> into the mix. But I suspect Ada is too far out for acceptance. I had
> a tough enough time getting VHDL in where needed, and avoiding a
> Verilog-only mandate, but only because Verilog couldn't have easily
> done that part of the job. The bulk of the HDL is still Verilog, and
> the VHDL will get translated after doing its special part.
> 
> As I mentioned on the other subthread, Ada to C, C++, or Java would
> probably be equally acceptable.
> 
> Dale Pontius
> NOT speaking for IBM

The few times I have been required to write in C, I wrote Ada anyway.  Then
commented out the ada and hand translated to C.  The only comments were the Ada
code and comments that were originally in the Ada code.

Larry

Re: Porting Ada to C (Stealth development)

[Ray Blaak]

Dan Andreatta <andreatta@mail.chem.sc.edu.REMOVEME> writes:
> pontius@btv.MBI.com.invalid (Dale Pontius) wrote in 
> news:a26o04$jfc$1@news.btv.ibm.com:
> 
> > I am taking up some new work, and would like to do it in Ada. However
> > there is a strong 'Do it in C so others can pick it up' bias, which I
> > guess makes some sense. So I'd like to do my development in Ada, and
> > have the occasional conversion to C to play well with others. This
> > probably needs to be free, as well, since it wouldn't cost money if
> > if I just started the project in C. (Can't justify) I did a quick scan
> > at www.adahome.org without luck. I'll also need to link in C libs, but
> > that should be straightforward and in the documentation.

If it has to be in C so that others can potentially develop and maintain it,
then it makes no sense to convert to C from Ada at all.

Conversion raises issues that otherwise would not exist, simply complicating
things, creating unnecessary problems.  Think about it: any C warts and issues
will exist anyway, and with conversion you would further have to deal Ada->C
incompatibilities, possible unusal link libraries, potentially unnatural
machine generated C code, etc.

Write it in C from the beginning. Just use your "Ada sensibilities" to ensure
that your C code is the best on your team.

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
blaak@telus.net                                The Rhythm has my soul.

Re: Porting Ada to C (Stealth development)

[Stephen Leake]

Just offering sympathy; I have no solution to your problem.

Although you might try writing Ada and seeing if anybody actually
objects to "picking it up"; it is easier to get forgiveness than
permission :). You could manually translate it to C at that time, if
necessary. 

pontius@btv.MBI.com.invalid (Dale Pontius) writes:

> One perpetual concern is, "How does this stuff get maintained when
> XXXX moves on?" For that reason, they want a mainstream language,
> especially considering (as mentioned above) the programmers here do
> it on the side, not as a profession.

I get the same response, even when we _are_ writing software as a
profession. 

I recently had occasion to confront my boss on why UML was "ok", but
Ada was "bad" (both are being proposed for new projects; we have no
history of using either). She was honest enough to admit it was
because "UML is a buzzword". Maybe I'll get her to consider real
issues yet!

-- 
-- Stephe

Re: Porting Ada to C (Stealth development)

[Ray Blaak]

Larry Hazel <lhhazel@otelco.net> writes:
> The few times I have been required to write in C, I wrote Ada anyway.  Then
> commented out the ada and hand translated to C.  The only comments were the
> Ada code and comments that were originally in the Ada code.

I guess I am fundamentally a lazy programmer.

Why implement it twice? This increases the chance for errors and is more work.

When (not if!) you fix a bug, do you fix only the C code, or the Ada as well?
How do you test the Ada fixes?

If you have to write in a particular programming language (I mean just *have*
to), it is fundamentally more efficient to write directly in that language,
even if that language is deficient, using that language's mode of thought
(influenced, of course, by the wisdom that Ada experience provides).

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
blaak@telus.net                                The Rhythm has my soul.

Re: Porting Ada to C (Stealth development)

[Larry Hazel]

Ray Blaak wrote:
> 
> Larry Hazel <lhhazel@otelco.net> writes:
> > The few times I have been required to write in C, I wrote Ada anyway.  Then
> > commented out the ada and hand translated to C.  The only comments were the
> > Ada code and comments that were originally in the Ada code.
> 
> I guess I am fundamentally a lazy programmer.
> 
> Why implement it twice? This increases the chance for errors and is more work.
> 
> When (not if!) you fix a bug, do you fix only the C code, or the Ada as well?
> How do you test the Ada fixes?
> 
Well, you should document your code with comments anyway, and you should change
the comments when you change the code for whatever reason.  Actually the Ada
code I wrote probably wasn't compilable anyway (more like pseudo code), but much
more readable than the C code could ever be.  It's the only way I could ever
write anything in C.  Never have been able to read C and figure out what it's
supposed to be doing.

Larry

Re: Porting Ada to C (Stealth development)

[Dan Andreatta]

Using Google, searching for ada2d, I found a couple of sites. Unfortunately 
in one they just say they did it, and in the other the links are broken. 
However, this is what I found:
http://www.idi.ntnu.no/~zoran/
http://www.ceng.metu.edu.tr/~e106170/aysecv.html

HTH
                 Dan

-- 
D. Andreatta
Univ. of S. Carolina
Chemistry Dept.

Bruce I:   ... American beer is a little like making love in a canoe!
Bruce III: Making love in a canoe? 
Bruce I:   It's f***ing close to water!

--- Monty Python, "The Bruces", Live at the Holliwood Bowl

Re: Porting Ada to C (Stealth development)

[Ted Dennison]

pontius@btv.MBI.com.invalid (Dale Pontius) wrote in message news:<a26o04$jfc$1@news.btv.ibm.com>...
> guess makes some sense. So I'd like to do my development in Ada, and
> have the occasional conversion to C to play well with others. This
> probably needs to be free, as well, since it wouldn't cost money if
> if I just started the project in C. (Can't justify) I did a quick scan
> at www.adahome.org without luck. I'll also need to link in C libs, but
> that should be straightforward and in the documentation.

Converters are notoriously horrible. Intermetrics claims to have one
that makes readable code, but it isn't free (and I have too many
relatives from Missouri to believe that without seeing it).

I have heard of people successfully using Ada as their PDL, and then
using C for the actual code. Since Ada is more human-readable anyway,
many people might not even realise your PDL is Ada. :-)

Another possible approach, if its a one-man project and they are as
open to alternate languages as you seem to imply they are, would be to
just write it in Ada without raising it as an issue in the first
place. You can't exactly get in trouble for not following orders if
you were never given any.

Re: Porting Ada to C (Stealth development)

[Wes Groleau]

> I have heard of people successfully using Ada as their PDL, and then
> using C for the actual code. Since Ada is more human-readable anyway,
> many people might not even realise your PDL is Ada. :-)

But the first guy that recognizes it may arouse
the heresy hunters and set up the stake.  :-)

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: Porting Ada to C (Stealth development)

[tmoran]

> > > The few times I have been required to write in C, I wrote Ada anyway.
> > Why implement it twice? This increases the chance for errors and is more work.
> Actually the Ada code I wrote probably wasn't compilable anyway
> (more like pseudo code),
  I've written and debugged in Ada, then transliterated to a C
deliverable.  It was faster than trying to get it running in C in the
first place.  Once it was translated, of course, any further changes
were strictly in the C version.

Re: Porting Ada to C (Stealth development)

[Zoran]

Fixed: 
 http://www.idi.ntnu.no/~zoran/ada2c.arj

It was just a small term project, so don't expect too much... 


--zoran

Dan Andreatta <andreatta@mail.chem.sc.edu> wrote in message news:<Xns919A90899DA2andreattamailchemsce@12.253.140.251>...
> Using Google, searching for ada2d, I found a couple of sites. Unfortunately 
> in one they just say they did it, and in the other the links are broken. 
> However, this is what I found:
> http://www.idi.ntnu.no/~zoran/
> http://www.ceng.metu.edu.tr/~e106170/aysecv.html
> 
>

Re: Porting Ada to C (Stealth development)

[Brian Rogoff]

On 18 Jan 2002, Dale Pontius wrote:
> In article <3C47375B.8060604@worldnet.att.net>,
>         Jim Rogers <jimmaureenrogers@worldnet.att.net> writes:
> > Dale Pontius wrote:
> >
> >> I am taking up some new work, and would like to do it in Ada. However
> >> there is a strong 'Do it in C so others can pick it up' bias, which I
> >> guess makes some sense. So I'd like to do my development in Ada, and

Fuggedaboutit.

Better to just use some decent C libraries

	http://www.cs.princeton.edu/software/cii

and a tool like splint http://www.splint.org/ to make your C as clean
and type safe as possible.

> > Actually, it makes no sense, but opening that can or worms begins the
> > language war. If a person is a competent C programmer he or she should
> > be able to learn a new language with minimal effort. My own experience
> > is that it was easier to learn Ada, knowing C first, than it was to
> > read and understand a lot of C programs.

I agree completely with Jim Rogers, but getting an organization to adopt
a language because one programmer likes it is nearly impossible.

> Doesn't matter. This is a rather conservative area, and "C is the way
> to program, though perhaps C++ or Java would be OK." Note that this is
> also not a programming area, it's silicon design. Sometimes you have
> to turn part-time programmer to get the silicon out. This is one of
> those.

There are lots of C (and C++) as HDL tools floating about too. None for
Ada.

-- Brian

Re: Porting Ada to C (Stealth development)

[Frode Tennebø]

On Friday 18 January 2002 15:12 Dale Pontius wrote:

> I can live with conversion to C++, but I can't live with commercial.
> Using Ada will be a stealth issue. Anything above-board would have to
> be C, C++, or perhaps Java.

Just a crazy idea: Write it in Ada, compile it into Java bytecode 
(JGNAT) and use any of the Java decompilers out there to get Java (if 
you absolutely have to). :)

 -Frode
-- 
^ Frode Tenneb� | email: frode@tennebo.com | Frode@IRC ^
|  with Standard.Disclaimer; use Standard.Disclaimer;  |

Re: Porting Ada to C (Stealth development)

[Simon Wright]

tmoran@acm.org writes:

> > > > The few times I have been required to write in C, I wrote Ada anyway.
> > > Why implement it twice? This increases the chance for errors and is more work.
> > Actually the Ada code I wrote probably wasn't compilable anyway
> > (more like pseudo code),
>   I've written and debugged in Ada, then transliterated to a C
> deliverable.  It was faster than trying to get it running in C in the
> first place.  Once it was translated, of course, any further changes
> were strictly in the C version.

Ada to assembler, in my case, in about 1984. The Ada was the
documentation, so of course it had to change (don't know what the
maintainers are doing now, don't think it's been touched for years ..)

I don't remember how many defects there were in the assembler
as-coded, but there was only one in the assembler as-delivered. I was
very proud of that.

Re: Porting Ada to C (Stealth development)

[Simon Wright]

Brian Rogoff <bpr@bpr.best.vwh.net> writes:

> I agree completely with Jim Rogers, but getting an organization to
> adopt a language because one programmer likes it is nearly
> impossible.

But someone has to start the ball rolling!

Re: Porting Ada to C (Stealth development)

[Ray Blaak]

Larry Hazel <lhhazel@otelco.net> writes:
> Ray Blaak wrote:
> > When (not if!) you fix a bug, do you fix only the C code, or the Ada as 
> > well? How do you test the Ada fixes?
>
> Well, you should document your code with comments anyway, and you should
> change the comments when you change the code for whatever reason.  Actually
> the Ada code I wrote probably wasn't compilable anyway (more like pseudo
> code), but much more readable than the C code could ever be.  It's the only
> way I could ever write anything in C.  Never have been able to read C and
> figure out what it's supposed to be doing.

True, comments should be kept in synch. However, pseudo-code comments are a
major pain to maintain. If one eliminates them, the amount of work is that much
less.

There is no reason in the standard languages why actual code cannot be its own
pseudo-code. The use of appropriate identifiers and the appropriate
decomposition of logic into suitable routines lets the high-level logic of the
code be easily seen.

C/C++'s basic control flow constructs are not really that much different from
Ada's. Even if some arcane bit twiddling is involved, one can always parcel
that up behind some suitably named routine. 

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
blaak@telus.net                                The Rhythm has my soul.

Re: Porting Ada to C (Stealth development)

[Brian Rogoff]

On 19 Jan 2002, Simon Wright wrote:
> Brian Rogoff <bpr@bpr.best.vwh.net> writes:
> > I agree completely with Jim Rogers, but getting an organization to
> > adopt a language because one programmer likes it is nearly
> > impossible.
>
> But someone has to start the ball rolling!

Well, Sisyphus, errr, Simon, if you happen to work in a place where you
are the lone programmer maintaining some piece of code, or where you have
lots of power, you may be able to effectively start that ball rolling,
but Dale indicated that that is not the situation. If you're one of many
equals, each with their own pet language, the chance that you'll rally
everone behind Ada is vanishingly small.

VHDL is a tough sell against Verilog in many places in the US, and VHDL is
far, far, far more popular as an HDL than Ada is as a general purpose
programming language.

-- Brian

Re: Porting Ada to C (Stealth development)

[Nick Roberts]

"Ray Blaak" <blaak@telus.net> wrote in message
news:u8zavy44g.fsf@telus.net...

> Write it in C from the beginning. Just use your "Ada sensibilities" to
ensure
> that your C code is the best on your team.

100% endorsed.

--
Best wishes,
Nick Roberts

Re: Porting Ada to C (Stealth development)

[Peter]

In article <Xns9199960B37554andreattamailchemsce@12.253.140.251>, Dan says...
>
 
>
>
>If you can live with C++, you could take a look at www.ada2cc.com.

wow! I had no idea such a thing existed! 

If the above is really good, this could be the best thing ever to
happen to Ada. Why? Becuase now many of us can code in Ada, then 
translate the code to C++ to hand it over to the bean counters who
want to see only C/C++ code.

This way, Ada will live on and prosper. 

Peter