PDA Ada

PDA Ada

[Richard Riehle]

In a conversation with a colleague yesterday,  the topic
of PDA's came up, yet again.   We have a lot of projects
funded for using wireless technology and PDA's such as
the Palm and IpaQ.   He would be willing to let peole
use Ada if there were compilers available.

Does anyone know if there is any on-going effort to
create an Ada compiler for PDA's.   We are going to
miss a huge market opportunity if we fail to be ready
for this burgeoning part of the embedded marketplace.

At present, no one in the PDA community takes Ada
seriously because there are no compilers.  I fear no
one in the compiler publishing community takes
PDA's seriously because there are no customers.

The Ada compiler publishers have always, with a
few exceptions now and then, operated on a risk
averse model with regard to products.   I suppose
that is what we can expect again.  Unless there is
customer demand, there will be no compiler. Meanwhile,
there are no customers demanding Ada for PDA's because
they have other alternatives.

Richard Riehle

Re: PDA Ada

[Martin Dowie]

> Does anyone know if there is any on-going effort to
> create an Ada compiler for PDA's.   We are going to
> miss a huge market opportunity if we fail to be ready
> for this burgeoning part of the embedded marketplace.

I was wondering if anyone had looked into porting GNAT
to Symbian OS* (formaly EPOC32). The SDK available from
their website is a port of gcc (C/C++ front ends).

Is there such a thing 'Dummy's guide to porting GNAT'???

*Symbian are a joint venture company made up from -
Ericsson, Nokia, Matsushita (Panasonic), Motorola,
Psion, Siemens and Sony-Ericsson. So a fairly heavy
weight crowd!

Re: PDA Ada

[Marin David Condic]

That's always a dillema for the vendors and I can't blame them too much.
They've been burned in the past trying to build for targets where the end
clientelle really weren't willing to consider Ada in the first place. The
excuse was always "Ada is a non-starter because there's no compiler for my
target..." when in reality there was an underlying attitude of "I don't like
Ada because its not C so I won't use it no matter what you do..."

If the vendors did want to go after the PDA market, it would be a major
mistake to go after it with anything less than a massive effort. People have
already got development kits for PDAs and they aren't going to drop them for
Ada if all they are offered is just another compiler for a different
langauge. In particular, Ada's selling points of reliability,
comprehensibility, maintainability, etc., are simply not compelling in a
market where the apps are going to live for a few months and get tossed out
in favor of something "New and Improved" that will make the marketing
department happy. A vendor would need to go into the PDA market with a
development kit that provided *more* than what is currently available, or
nobody is going to find any advantage to switch.

I agree that Ada could benefit dramatically by getting in on the ground
floor of some emerging technology and PDA's, cellular phones, digital TV and
a variety of other things are still new enough to not be hide-bound in their
ways. But whatever you bring to the table with these folks, it *must* give
them leverage in the time to market area or they have no incentive to
switch.

That's a lot to ask of vendors with limited resources and no clear mandate
to build something.

MDC
--
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jast.mil/

Send Replies To: m c o n d i c @ a c m . o r g

    "I'd trade it all for just a little more"
        --  Charles Montgomery Burns, [4F10]
======================================================================

Richard Riehle <richard@adaworks.com> wrote in message
news:3DA4E4A8.5C1DEAE0@adaworks.com...
>
> The Ada compiler publishers have always, with a
> few exceptions now and then, operated on a risk
> averse model with regard to products.   I suppose
> that is what we can expect again.  Unless there is
> customer demand, there will be no compiler. Meanwhile,
> there are no customers demanding Ada for PDA's because
> they have other alternatives.
>

Re: PDA Ada

[Ted Dennison]

Richard Riehle <richard@adaworks.com> wrote in message news:<3DA4E4A8.5C1DEAE0@adaworks.com>...
> Does anyone know if there is any on-going effort to
> create an Ada compiler for PDA's.

I've never heard of one.

> We are going to
> miss a huge market opportunity if we fail to be ready
> for this burgeoning part of the embedded marketplace.

PDA's may sell a lot of units, but that doesn't mean their compiler
market is huge.

> I fear no one in the compiler publishing community takes
> PDA's seriously because there are no customers.

Possibly. 

A collegue of mine came in here a few years back with a similar
argument about VenturCOM's RTX (realtime extensions for Windows). My
personal opinion is that we should have siezed the oppertunity to port
Gnat to it ourselves. We had no compunction about doing that with
Fortran (which we couldn't do without, due to large amounts of legacy
code). C++ in RTX is turning out to be 10 times the nightmare I had
feared.

Realisticly, its up to compiler companies to decide for themselves
that there's a profitable niche open or them in a marginal market like
PDAs, DSPs, RTX, etc. If they don't, but motivated users ask for it,
one would think it quite reasonable for the vendor to expect said
motivated users to pony up at least some of the deveolpment costs.

Or of course the users could just do it themselves. It looks like it
would be doable. There appears to be a Palm version of GCC. The
processor looks like a 68K variant, which has been known to support
full Ada compilers in the past.

Re: PDA Ada

[Adrian Hoe]

Richard Riehle wrote:

> In a conversation with a colleague yesterday,  the topic
> of PDA's came up, yet again.   We have a lot of projects
> funded for using wireless technology and PDA's such as
> the Palm and IpaQ.   He would be willing to let peole
> use Ada if there were compilers available.
> 
> Does anyone know if there is any on-going effort to
> create an Ada compiler for PDA's.   We are going to
> miss a huge market opportunity if we fail to be ready
> for this burgeoning part of the embedded marketplace.
> 
> At present, no one in the PDA community takes Ada
> seriously because there are no compilers.  I fear no
> one in the compiler publishing community takes
> PDA's seriously because there are no customers.
> 
> The Ada compiler publishers have always, with a
> few exceptions now and then, operated on a risk
> averse model with regard to products.   I suppose
> that is what we can expect again.  Unless there is
> customer demand, there will be no compiler. Meanwhile,
> there are no customers demanding Ada for PDA's because
> they have other alternatives.
> 
> Richard Riehle
> 
> 


My colleagues and I are longing for an Ada compiler for many years 
now. Sadly, we have to develop PDA application using C and Java. Yucks!

We want Ada for PDAs!
-- 
type Dmitry is new Adrian;           -- Adrian Hoe
                                      -- http://adrianhoe.com
                                      -- Remove *nospam* to email

Re: PDA Ada

[Hyman Rosen]

Ted Dennison wrote:
 > C++ in RTX is turning out to be 10 times the nightmare I had feared.

Just out of curiosity, what problems are you encountering?

Re: PDA Ada

[Michael Bode]

dennison@telepath.com (Ted Dennison) writes:

> Or of course the users could just do it themselves. It looks like it
> would be doable. There appears to be a Palm version of GCC. The
> processor looks like a 68K variant, which has been known to support
> full Ada compilers in the past.

So the problem is how to port the Ada runtime to target X. Has anyone
besides ACT ever ported the gnat runtime to any target?

Re: PDA Ada

[John Kern]

If you are already prepared to use Java for the Palm computer, you might
want to check out the
JGNAT to Palmtm Demo
available on the ACT website at the bottom of this page:
http://www.gnat.com/texts/products/prod_java.htm
I downloaded it a while back, but haven't yet had the time to get it
working.

Adrian Hoe wrote:

> Richard Riehle wrote:
>
> > In a conversation with a colleague yesterday,  the topic
> > of PDA's came up, yet again.   We have a lot of projects
> > funded for using wireless technology and PDA's such as
> > the Palm and IpaQ.   He would be willing to let peole
> > use Ada if there were compilers available.
> >
> > Does anyone know if there is any on-going effort to
> > create an Ada compiler for PDA's.   We are going to
> > miss a huge market opportunity if we fail to be ready
> > for this burgeoning part of the embedded marketplace.
> >
> > At present, no one in the PDA community takes Ada
> > seriously because there are no compilers.  I fear no
> > one in the compiler publishing community takes
> > PDA's seriously because there are no customers.
> >
> > The Ada compiler publishers have always, with a
> > few exceptions now and then, operated on a risk
> > averse model with regard to products.   I suppose
> > that is what we can expect again.  Unless there is
> > customer demand, there will be no compiler. Meanwhile,
> > there are no customers demanding Ada for PDA's because
> > they have other alternatives.
> >
> > Richard Riehle
> >
> >
>
> My colleagues and I are longing for an Ada compiler for many years
> now. Sadly, we have to develop PDA application using C and Java. Yucks!
>
> We want Ada for PDAs!
> --
> type Dmitry is new Adrian;           -- Adrian Hoe
>                                       -- http://adrianhoe.com
>                                       -- Remove *nospam* to email

Re: PDA Ada

[Adrian Hoe]

Michael Bode wrote:

> dennison@telepath.com (Ted Dennison) writes:
> 
> 
>>Or of course the users could just do it themselves. It looks like it
>>would be doable. There appears to be a Palm version of GCC. The
>>processor looks like a 68K variant, which has been known to support
>>full Ada compilers in the past.
>>
> 
> So the problem is how to port the Ada runtime to target X. Has anyone
> besides ACT ever ported the gnat runtime to any target?
> 


We tried to port gnat to 68K Dragonball (Palm) two years ago. It was 
very tedious and lack of documentation and the port was dropped.

What about GCC3.x? I heard it supports more targets now but I haven't 
got my hand on it yet.
-- 
type Dmitry is new Adrian;           -- Adrian Hoe
                                      -- http://adrianhoe.com
                                      -- Remove *nospam* to email

Re: PDA Ada

[Vadim Godunko]

**** Post for FREE via your newsreader at post.usenet.com ****

> So the problem is how to port the Ada runtime to target X. Has anyone
> besides ACT ever ported the gnat runtime to any target?
Some time ago I port gnat to x86/FreeBSD. If you system is POSIX compatible
this is simple process.

Vadim Godunko
North-Caucasian Area Air Traffic Control Center



-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 *** Usenet.com - The #1 Usenet Newsgroup Service on The Planet! ***
                      http://www.usenet.com
Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Re: PDA Ada

[Steffen Huber]

Michael Bode wrote:
> dennison@telepath.com (Ted Dennison) writes:
> > Or of course the users could just do it themselves. It looks like it
> > would be doable. There appears to be a Palm version of GCC. The
> > processor looks like a 68K variant, which has been known to support
> > full Ada compilers in the past.
> 
> So the problem is how to port the Ada runtime to target X. Has anyone
> besides ACT ever ported the gnat runtime to any target?

I know of one person who ported GNAT 3.03 and its runtime to RISC OS.
Apart from not supporting tasking (back then, RISC OS did not have
a working POSIX-like threading possibility), it was complete. And I
am sure that he was not employed by ACT ;-)

Steffen

-- 
steffen.huber@gmx.de               steffen@huber-net.de
GCC for RISC OS  - http://www.arcsite.de/hp/gcc/
Private homepage - http://www.huber-net.de/

Re: PDA Ada

[Ted Dennison]

Michael Bode <m.g.bode@web.de> wrote in message news:<m3u1juyrbc.fsf@jupiter.solar.system>...
> dennison@telepath.com (Ted Dennison) writes:
> 
> > Or of course the users could just do it themselves. It looks like it
> > would be doable. There appears to be a Palm version of GCC. The
> > processor looks like a 68K variant, which has been known to support
> > full Ada compilers in the past.
> 
> So the problem is how to port the Ada runtime to target X. Has anyone
> besides ACT ever ported the gnat runtime to any target?

I believe the MacOSX port is completely volunteer. It may have
leveraged a bit off the existing PowerMac port, which was initially
done by ACT if I remember correctly. But that port's been maintained
by volunteers for years now as well. I think their website is
http://www.adapower.net/macos/

Re: PDA Ada

[Rob Veenker]

The JGNAT solution also applies to Pocket PC devices such as the iPAQ.
We are about to finish a military application that runs on a ruggedised
iPAQ. It uses a military radio for data communications and this protocol is
100 % Ada. Only the GUI part is Java.
Because the protocol was implemented in Ada, We could use the same code on a
Windows NT box :-)

"John Kern" <jkern3@corecomm.net> wrote in message
news:3da62ce4$0$1428$272ea4a1@news.execpc.com...
> If you are already prepared to use Java for the Palm computer, you might
> want to check out the
> JGNAT to Palmtm Demo
> available on the ACT website at the bottom of this page:
> http://www.gnat.com/texts/products/prod_java.htm
> I downloaded it a while back, but haven't yet had the time to get it
> working.
>
> Adrian Hoe wrote:
>
> > Richard Riehle wrote:
> >
> > > In a conversation with a colleague yesterday,  the topic
> > > of PDA's came up, yet again.   We have a lot of projects
> > > funded for using wireless technology and PDA's such as
> > > the Palm and IpaQ.   He would be willing to let peole
> > > use Ada if there were compilers available.
> > >
> > > Does anyone know if there is any on-going effort to
> > > create an Ada compiler for PDA's.   We are going to
> > > miss a huge market opportunity if we fail to be ready
> > > for this burgeoning part of the embedded marketplace.
> > >
> > > At present, no one in the PDA community takes Ada
> > > seriously because there are no compilers.  I fear no
> > > one in the compiler publishing community takes
> > > PDA's seriously because there are no customers.
> > >
> > > The Ada compiler publishers have always, with a
> > > few exceptions now and then, operated on a risk
> > > averse model with regard to products.   I suppose
> > > that is what we can expect again.  Unless there is
> > > customer demand, there will be no compiler. Meanwhile,
> > > there are no customers demanding Ada for PDA's because
> > > they have other alternatives.
> > >
> > > Richard Riehle
> > >
> > >
> >
> > My colleagues and I are longing for an Ada compiler for many years
> > now. Sadly, we have to develop PDA application using C and Java. Yucks!
> >
> > We want Ada for PDAs!
> > --
> > type Dmitry is new Adrian;           -- Adrian Hoe
> >                                       -- http://adrianhoe.com
> >                                       -- Remove *nospam* to email
>

Re: PDA Ada

[Adrian Hoe]

Rob Veenker wrote:

> The JGNAT solution also applies to Pocket PC devices such as the iPAQ.
> We are about to finish a military application that runs on a ruggedised
> iPAQ. It uses a military radio for data communications and this protocol is
> 100 % Ada. Only the GUI part is Java.
> Because the protocol was implemented in Ada, We could use the same code on a
> Windows NT box :-)
> 


We haven't tried JGNAT yet. Do you mean the data communication 
protocol is written in JGNAT? Is the GUI part also written in JGNAT? I 
am confused.
-- 
type Dmitry is new Adrian;           -- Adrian Hoe
                                      -- http://adrianhoe.com
                                      -- Remove *nospam* to email

Re: PDA Ada

[Dr. Michael Paus]

Rob Veenker wrote:
> The JGNAT solution also applies to Pocket PC devices such as the iPAQ.
> We are about to finish a military application that runs on a ruggedised
> iPAQ. It uses a military radio for data communications and this protocol is
> 100 % Ada. Only the GUI part is Java.

Just for curiousity. Is it possible to tell us which military radio and
which protocol (message standard ?) you are referring to?

> Because the protocol was implemented in Ada, We could use the same code on a
> Windows NT box :-)

Well, I think the same would have been true if the protocol were implemented
in Java :-(

Are you funding ACT for supporting JGNAT. The last I heard is that this product
is not supported anymore due to lack of customer interest. The current version
also does not run anymore on recent Java versions. So you will have to stick
with an outdated Java version which is probably not such a big problem for you
because the one delivered with the iPAQ is also quite old.

Michael

Porting GNAT to a new target, was: Re: PDA Ada

[Simon Clubley]

In article <3daa2c68_1@news.tm.net.my>, Adrian Hoe <mailbox@*nospam*adrianhoe.com> writes:
> 
> We tried to port gnat to 68K Dragonball (Palm) two years ago. It was 
> very tedious and lack of documentation and the port was dropped.
> 

I have wondered what kind of effort this involves, so I have a few questions:

What exactly is involved in porting the GNAT runtime to a new target ?

Is there any documentation at all about the overall process involved ?

What problems did you encounter attempting this port ?

Thanks for any information,

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       
Microsoft: The Lada of the computing world.

Re: PDA Ada

[Wes Groleau]

"Dr. Michael Paus" wrote:
> 
> Rob Veenker wrote:
> > The JGNAT solution also applies to Pocket PC devices such as the iPAQ.
> > We are about to finish a military application that runs on a ruggedised
> > iPAQ. It uses a military radio for data communications and this protocol is
> > 100 % Ada. Only the GUI part is Java.
> 
> Just for curiousity. Is it possible to tell us which military radio and
> which protocol (message standard ?) you are referring to?
> 
> > Because the protocol was implemented in Ada, We could use the same code on a
> > Windows NT box :-)
> 
> Well, I think the same would have been true if the protocol were implemented
> in Java :-(
> 
> Are you funding ACT for supporting JGNAT. The last I heard is that this product
> is not supported anymore due to lack of customer interest. The current version
> also does not run anymore on recent Java versions. So you will have to stick

However, there are other Ada to J-code compilers.
And the GMGPL would allow you to use the Java lib
bindings to the J2ME (IF for some reason the other
vendor's bindings were not adequate).

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

Re: Porting GNAT to a new target, was: Re: PDA Ada

[Adrian Hoe]

Simon Clubley wrote:

> In article <3daa2c68_1@news.tm.net.my>, Adrian Hoe <mailbox@*nospam*adrianhoe.com> writes:
> 
>>We tried to port gnat to 68K Dragonball (Palm) two years ago. It was 
>>very tedious and lack of documentation and the port was dropped.
>>
>>
> 
> I have wondered what kind of effort this involves, so I have a few questions:
> 
> What exactly is involved in porting the GNAT runtime to a new target ?
> 
> Is there any documentation at all about the overall process involved ?
> 
> What problems did you encounter attempting this port ?
> 
> Thanks for any information,
> 
> Simon.
> 
> 


It's actually a cross-compiler port to Dragon ball. We hit on many 
walls trying to build the cross-compiler. :(
-- 
type Dmitry is new Adrian;           -- Adrian Hoe
                                      -- http://adrianhoe.com
                                      -- Remove *nospam* to email

Re: PDA Ada

[Rob Veenker]

The radio concerned is of the PR4G family (aka FM9000) by thales. It uses a
protocol called SYCOMORE, used for remote control of the radio itself. This
protocol however also allows (limited) data communication.
On the iPAQ a SYCOMORE protocol stack written in Ada was implemented using
JGNAT.
To allow messaging, we also wrote message acknowlegdement, dispatching and
marshalling functionality (all in Ada as well). The result is a
communication 'jar' (Java archive) that is used by a Java GUI.

We are indeed using the supported version of JGNAT and are aware of the JVM
problems. The JVM's available on the iPAQ have not shown this problem
(Insignia/Jeode and NSI/CrEme).
Whether this becomes a problem is still unknown.
Also the Java standard for small devices (Personal Java) differs from the
JDK. The JVM on the iPAQ is not outdated, it's just different (sort of a
subset).

Since the protocol is implemented in Ada there are alternatives like going
native on the iPAQ using VxWorks/GNAT :-)


"Dr. Michael Paus" <paus@ib-paus.com> wrote in message
news:aoea6s$jq3$1@news.online.de...
> Rob Veenker wrote:
> > The JGNAT solution also applies to Pocket PC devices such as the iPAQ.
> > We are about to finish a military application that runs on a ruggedised
> > iPAQ. It uses a military radio for data communications and this protocol
is
> > 100 % Ada. Only the GUI part is Java.
>
> Just for curiousity. Is it possible to tell us which military radio and
> which protocol (message standard ?) you are referring to?
>
> > Because the protocol was implemented in Ada, We could use the same code
on a
> > Windows NT box :-)
>
> Well, I think the same would have been true if the protocol were
implemented
> in Java :-(
>
> Are you funding ACT for supporting JGNAT. The last I heard is that this
product
> is not supported anymore due to lack of customer interest. The current
version
> also does not run anymore on recent Java versions. So you will have to
stick
> with an outdated Java version which is probably not such a big problem for
you
> because the one delivered with the iPAQ is also quite old.
>
> Michael
>

Re: PDA Ada

[Dr. Michael Paus]

Rob Veenker wrote:
> The radio concerned is of the PR4G family (aka FM9000) by thales. It uses a
> protocol called SYCOMORE, used for remote control of the radio itself. This
> protocol however also allows (limited) data communication.
> On the iPAQ a SYCOMORE protocol stack written in Ada was implemented using
> JGNAT.
> To allow messaging, we also wrote message acknowlegdement, dispatching and
> marshalling functionality (all in Ada as well). The result is a
> communication 'jar' (Java archive) that is used by a Java GUI.

I was just wondering how you implemented the messaging in Ada so that it
can be compiled by JGNAT. The message standards I know (e.g. Link 16 J-Messages)
are heavily bit oriented and people tend to implement this via Ada records
with lots of representation clauses. But as far as I know JGNAT does not
support representation clauses (correct me if I am wrong). So how did you
do that?

Michael

Re: PDA Ada

[Rob Veenker]

"Dr. Michael Paus" <paus@ib-paus.com> wrote in message
news:aoovst$pf9$1@news.online.de...
> Rob Veenker wrote:
> > The radio concerned is of the PR4G family (aka FM9000) by thales. It
uses a
> > protocol called SYCOMORE, used for remote control of the radio itself.
This
> > protocol however also allows (limited) data communication.
> > On the iPAQ a SYCOMORE protocol stack written in Ada was implemented
using
> > JGNAT.
> > To allow messaging, we also wrote message acknowlegdement, dispatching
and
> > marshalling functionality (all in Ada as well). The result is a
> > communication 'jar' (Java archive) that is used by a Java GUI.
>
> I was just wondering how you implemented the messaging in Ada so that it
> can be compiled by JGNAT. The message standards I know (e.g. Link 16
J-Messages)
> are heavily bit oriented and people tend to implement this via Ada records
> with lots of representation clauses. But as far as I know JGNAT does not
> support representation clauses (correct me if I am wrong). So how did you
> do that?
>
> Michael
>

Most of the low-level stuff is done using byte arrays and logical AND's and
OR's :-)
The records are really not that complicated but I admit that representation
clauses would make code better readable and maintainable. At least Ada does
have hexadecimal notation and bitwise operations.

Rob