Re: An OS in Ada, why not RTEMS ?

[James Ross <rem.jr@rem.webross.com>]

On Fri, 26 Apr 2002 15:58:15 -0500, sk <noname@myob.com> wrote:

>A brief look at the RTEMS specs ... 

Their README for the current version erroneously states: "There are
two implementations of RTEMS in this source tree in Ada and in C".

There is no current Ada tree. RTEMS is written in C and implements the
POSIX API. OAR provides the means, by a patched GNAT, to create
real-time apps for it in Ada. 

>I am wondering why the contributors to the said 
>threads do not regard RTEMS as a suitable OS to 
>adopt as a platform for playing with ?

However, I did take another look at it and found out that at one time
it was written in Ada ( I'm pretty sure Ada83 ).  File dates are from
1995. That was version 3.2.1 and even though it is no longer
maintained you can still download it!  Now, as for taking this and
hacking away at it, hmmm maybe, I don't know? (It appears the License
would not preclude one from doing so).  Perhaps MarteOS might be a
better platform to "borrow" a code base?

I have been pondering / tinkering and researching the different
possibilities / problems / dilemmas of such a project (OS in Ada).
Here are some of my personal observations on the subject:

- As for the ultimate goal of an OS written in Ada -- I agree with the
AdaOS folks; To end up with a "complete OS".  Just to have a real-time
kernel written in Ada,  even though might be appealing to the embedded
programmers, is by itself of little use to me personally.  However, I
agree if there was one, it might generate enough interest to turn into
more than just that.

- I can't help from constantly thinking that any work from scratch on
a project of this scope is an exercise of re-inventing the wheel.  

- To begin with the boot process and progress through all the stages
of OS design starting with absolutely nothing is a monumental task!
Hence I believe the reason there has been no *apparent* progress on
the AdaOS project. 

- Then there is the GNAT raw binary executable images / linker
problems / issues.  Even though not insurmountable, it's a pretty big
obstacle.  Perhaps the GNAT Professional product has direct support
for creating raw binary executables that support the run-time and all
the Ada constructs without any dependence on an OS (I.e. an embedded
RT target).  But this does not seem to be the case with the free GNAT
version. It seems to use free GNAT, you would need to support either
ELF, or PE right of the bat in your new OS and be willing to do some
hacking on the GNAT run time sources as well.

I have come full circle in my thinking and believe it would be simpler
to start with say either FreeBSD or Linux and gradually "craft" it
into an "Ada OS".  You would not need to hack the free GNAT toolset.
Of the many advantages to this approach, one of them would be that
even if the goal is never reached, you might end up with some really
cool add-ons and/or extensions to an existing OS. Isn't this basically
what Apple has done with OS X?

At this point, I have about decided to chose a different project to
learn Ada 95 on.  To build an OS from scratch might be just a bit over
my head :)
JR