Hi Everyone

My name is Patrick, I have been posting here for years, just from different profiles.

We all know that Ada has declined since the 1980s but if you don't care what flavour of coolaid the other kids are drinking does it matter... gnat is maintained as of today.

However, let's just say the GCC project dropped support, aliens abducted all of the Adacore staff and you did not feel you would be able to patch GCC yourself, how long would you use an old GCC version that had Ada support?

I ask this because my son is 15 and my daughter is 11 and I am wondering about a project for them that would need to be viable for a long time.

Thanks for reading-Pat

Pat Van Canada <google@spellingbeewinnars.org> writes:

> We all know that Ada has declined since the 1980s

I don't; what is your evidence?

AdaCore appears to be growing. I get more trouble tickets for Emacs
ada-mode than I used to.

> However, let's just say the GCC project dropped support, aliens
> abducted all of the Adacore staff and you did not feel you would be
> able to patch GCC yourself, how long would you use an old GCC version
> that had Ada support?

You could pose the same improbable "what if" for any other language.

My answer; as long as it still works for my current projects.

A similar example; I was using monotone for a long time, because it is a
better structured version control system (particularly for managing
merge conflicts). I'm now using git, because monotone finally failed to
build on a Debian system.

> I ask this because my son is 15 and my daughter is 11 and I am
> wondering about a project for them that would need to be viable for a
> long time.

Ada is a better language because it encourages good programming style.
Even if they go on to use other languages, learning Ada first will make
them better programmers.

If we are allowed to consider the available commercial support, AdaCore
provided _much_ better support than the other compiler/OS companies I
dealt with when I worked for NASA (several years ago now; I hope their
quality is still high). And they have a college support program; do
other compiler companies have such things (seems like they should)?

AdaCore serves a niche market compared to RedHat or Oracle, but that
market is essential, growing, and will be around for the foreseeable
future. So I would bet on AdaCore and Ada.

--
-- Stephe

On 2021-05-31 19:20, Stephen Leake wrote:

> AdaCore serves a niche market compared to RedHat or Oracle, but that
> market is essential, growing, and will be around for the foreseeable
> future.

True. However growing software complexity will inevitably change the
picture. Things cannot continue as they do infinitely. E.g. this
announce is showing:

https://www.adacore.com/company/partners/nvidia

--
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

>...
>
> A similar example; I was using monotone for a long time, because it is a
> better structured version control system (particularly for managing
> merge conflicts). I'm now using git, because monotone finally failed to
> build on a Debian system.
>
Have you tried Mercurial ?

>...

On Monday, May 31, 2021 at 10:10:33 AM UTC-6, spellingbeewinnars.org wrote:
>
> However, let's just say the GCC project dropped support, aliens abducted all of the Adacore staff and you did not feel you would be able to patch GCC yourself, how long would you use an old GCC version that had Ada support?
GCC is kind of overly-convoluted, and GNAT has to cater to this.
I think that if this happened right now, we (as a community) could use this as impetus to create a MIT-licensed SPARK-proved (inasmuch as possible) free compiler; thus we could be in a far better position, in the end.

Don't get me wrong, GNAT isn't fundamentally bad, nor is AdaCore -- despite disagreements on design or [in AdaCore's case] marketing -- but it's easy for newcomers to mistake Ada as belonging to AdaCore.

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:

> On 2021-05-31 19:20, Stephen Leake wrote:
>
>> AdaCore serves a niche market compared to RedHat or Oracle, but that
>> market is essential, growing, and will be around for the foreseeable
>> future.
>
> True. However growing software complexity will inevitably change the
> picture. Things cannot continue as they do infinitely. E.g. this
> announce is showing:
>
> https://www.adacore.com/company/partners/nvidia

I think that increases the probability that AdaCore will last; they are
keeping up with current hardware trends, not just resting on past work.

--
-- Stephe

DrPi <314@drpi.fr> writes:

>>...
>> A similar example; I was using monotone for a long time, because it
>> is a
>> better structured version control system (particularly for managing
>> merge conflicts). I'm now using git, because monotone finally failed to
>> build on a Debian system.
>>
> Have you tried Mercurial ?

I looked at it. In my opinion, it's not sufficiently better than git to
merit bucking the trend. Using git, I can get help on any topic related
to git just by reading old stackoverflow posts.

--
-- Stephe

Le 07/06/2021 à 19:33, Stephen Leake a écrit :
> DrPi <314@drpi.fr> writes:
>
>>> ...
>>> A similar example; I was using monotone for a long time, because it
>>> is a
>>> better structured version control system (particularly for managing
>>> merge conflicts). I'm now using git, because monotone finally failed to
>>> build on a Debian system.
>>>
>> Have you tried Mercurial ?
>
> I looked at it. In my opinion, it's not sufficiently better than git to
> merit bucking the trend. Using git, I can get help on any topic related
> to git just by reading old stackoverflow posts.
>
Mercurial is not better, it is different. What I like is it's
simplicity. There is a good mailing list for support.

DrPi <314@drpi.fr> writes:

> Mercurial is not better, it is different. What I like is it's
> simplicity. There is a good mailing list for support.

Git is a lot more capable than Hg. I admit that the command line is
annoyingly inconsistent (e.g. git tag -l/--list vs git stash list).