ANN: Ada 2005 Math Extensions 20101223

ANN: Ada 2005 Math Extensions 20101223

[Simon Wright]

This release is now available at
http://sourceforge.net/projects/gnat-math-extn/files/20101223/

Changes:

The package is renamed to Ada_Numerics.Generic_Arrays.

An additional overloaded procedure Eigensystem returns the generalized
eigenvalues and eigenvectors of a pair of non-symmetric real matrices.
NB, this is only supported for unconstrained Float and Long_Float at
this time.

To do: 
   Add support for constrained floats to this Eigensystem.
   Add generalized solutions for non-hermitian complex matrices.
   ...

Enjoy (this, and the season's festivities!)

Re: ANN: Ada 2005 Math Extensions 20101223

[Ada novice]

On Dec 23, 11:21 pm, Simon Wright <si...@pushface.org> wrote:
> This release is now available athttp://sourceforge.net/projects/gnat-math-extn/files/20101223/
>

Thanks Simon. I was just wondering which file to compile. I opened
test_extensions.gpr as usual. I see that the file test_extensions.adb
is the same as in earlier version and this was the one that I used to
set as main. You have some file ending with _eigenvalues now and I
would like to know which one to set as the main file.
I am not familiar with AUnit.

Thanks

Best wishes for the end-of-year festivities
YC

Re: ANN: Ada 2005 Math Extensions 20101223

[Simon Wright]

Ada novice <ycalleecharan@gmx.com> writes:

> On Dec 23, 11:21 pm, Simon Wright <si...@pushface.org> wrote:
>> This release is now available at http://sourceforge.net/projects/gnat-math-extn/files/20101223/
>>
>
> Thanks Simon. I was just wondering which file to compile. I opened
> test_extensions.gpr as usual. I see that the file test_extensions.adb
> is the same as in earlier version and this was the one that I used to
> set as main. You have some file ending with _eigenvalues now and I
> would like to know which one to set as the main file.  I am not
> familiar with AUnit.

The new subprogram is in src/ada_numerics-generic_arrays.ad[bs] with the
others. Compiling test_extensions.adb using test_extensions.gpr will go
ahead and build it with the new library source.

Unfortunately I forgot to update test_extensions.adb to use the new
subprogram; so there are no hints about the calling sequence. You can
see how it works in test/real_generalized_eigenvalues.adb (eg, line
361).

As I think I've said before, test_extensions.adb should really be called
demonstrate_extensions.adb; a test program really ought to state the
expected results!

> Best wishes for the end-of-year festivities

Thanks very much! and the same to you ...

Re: ANN: Ada 2005 Math Extensions 20101223

[Ada novice]

On Dec 24, 7:30 pm, Simon Wright <si...@pushface.org> wrote:

> Unfortunately I forgot to update test_extensions.adb to use the new
> subprogram; so there are no hints about the calling sequence. You can
> see how it works in test/real_generalized_eigenvalues.adb (eg, line
> 361).

Thanks Simon. I shall study the above fie.

Cheers
YC

Paraffin: Parallelism generics for Ada 2005

[Brad Moore]

The initial release of paraffin is available at

http://paraffin.sourceforge.net/

It provides generics for adding parallelism to loops and recursive 
structures.

For iterative parallelism there are
work-sharing, work-seeking, and work-stealing forms.

For recursive parallelism there are
work-sharing and and work seeking forms.

Also, the recursive parallelism can provide stack-safe recursion that
avoids stack overflow.

The stack safe recursion generics provide this capability even if
used on a single processor.

Any feedback is greatly appreciated.

Brad Moore

Re: Paraffin: Parallelism generics for Ada 2005

[R. Tyler Croy]

Brad Moore <brad.moore@shaw.ca> wrote:
> The initial release of paraffin is available at
> 
> http://paraffin.sourceforge.net/
> 
> It provides generics for adding parallelism to loops and recursive 
> structures.
> 
> For iterative parallelism there are
> work-sharing, work-seeking, and work-stealing forms.
> 
> For recursive parallelism there are
> work-sharing and and work seeking forms.
> 
> Also, the recursive parallelism can provide stack-safe recursion that
> avoids stack overflow.
> 
> The stack safe recursion generics provide this capability even if
> used on a single processor.
> 
> Any feedback is greatly appreciated.

The site has a sourceforge placeholder page in place, and the SVN repo is
empty:
    http://paraffin.svn.sourceforge.net/viewvc/paraffin/

Where's the beef? :)

-- 
- R. Tyler Croy
--------------------------------------
    Code: http://github.com/rtyler

Re: Paraffin: Parallelism generics for Ada 2005

[Georg Bauhaus]

On 1/25/11 10:06 PM, R. Tyler Croy wrote:

>
> The site has a sourceforge placeholder page in place, and the SVN repo is
> empty:
>      http://paraffin.svn.sourceforge.net/viewvc/paraffin/
> Where's the beef? :)
>

My browsing software was directed to here:
http://sourceforge.net/projects/paraffin/
and here:
http://sourceforge.net/projects/paraffin/files/

All files available in one single compressed archive,
very effective, and inclusive ;-)

Re: Paraffin: Parallelism generics for Ada 2005

[R. Tyler Croy]

Georg Bauhaus <rm-host.bauhaus@maps.futureapps.de> wrote:
> On 1/25/11 10:06 PM, R. Tyler Croy wrote:
> 
>>
>> The site has a sourceforge placeholder page in place, and the SVN repo is
>> empty:
>>      http://paraffin.svn.sourceforge.net/viewvc/paraffin/
>> Where's the beef? :)
>>
> 
> My browsing software was directed to here:
> http://sourceforge.net/projects/paraffin/
> and here:
> http://sourceforge.net/projects/paraffin/files/
> 
> All files available in one single compressed archive,
> very effective, and inclusive ;-)

Yes, I saw that, I was more curious why the source is only distributed via
compressed archives instead of imported into a public version control system.


-- 
- R. Tyler Croy
--------------------------------------
    Code: http://github.com/rtyler

Re: Paraffin: Parallelism generics for Ada 2005

[Yannick Duchêne (Hibou57)]

Le Tue, 25 Jan 2011 17:04:59 +0100, Brad Moore <brad.moore@shaw.ca> a  
écrit:

> The initial release of paraffin is available at
>
> http://paraffin.sourceforge.net/
Thanks, this must be fun. Hope I will have some comments to send to you,  
as you expressed your express enjoyment about it :)

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.

“I am fluent in ASCII” [Warren 2010]

Re: Paraffin: Parallelism generics for Ada 2005

[Yannick Duchêne (Hibou57)]

Le Tue, 25 Jan 2011 17:04:59 +0100, Brad Moore <brad.moore@shaw.ca> a  
écrit:

> The initial release of paraffin is available at
>
> http://paraffin.sourceforge.net/
Wanted to link to that thread, but surprisingly, this one does not appears  
in Google Group (strange).

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.

“I am fluent in ASCII” [Warren 2010]

Re: Paraffin: Parallelism generics for Ada 2005

[Brad Moore]

On 25/01/2011 2:53 PM, R. Tyler Croy wrote:
> Georg Bauhaus<rm-host.bauhaus@maps.futureapps.de>  wrote:
>> On 1/25/11 10:06 PM, R. Tyler Croy wrote:
>>
>>>
>>> The site has a sourceforge placeholder page in place, and the SVN repo is
>>> empty:
>>>       http://paraffin.svn.sourceforge.net/viewvc/paraffin/
>>> Where's the beef? :)
>>>
>>
>> My browsing software was directed to here:
>> http://sourceforge.net/projects/paraffin/
>> and here:
>> http://sourceforge.net/projects/paraffin/files/
>>
>> All files available in one single compressed archive,
>> very effective, and inclusive ;-)
>
> Yes, I saw that, I was more curious why the source is only distributed via
> compressed archives instead of imported into a public version control system.
>
>


No particular reason, I suppose. This is my first release of a project 
on source forge. I checked out several other Ada open source projects, 
and that seemed to be the approach taken elsewhere. If it makes sense to
release in another manner, I am open to considering that. One issue I 
think with releasing in a public version control system, is choosing 
which one. There are a number of them out there.

Re: Paraffin: Parallelism generics for Ada 2005

[Yannick Duchêne (Hibou57)]

Le Wed, 26 Jan 2011 02:53:39 +0100, Brad Moore <brad.moore@shaw.ca> a  
écrit:
> No particular reason, I suppose. This is my first release of a project  
> on source forge. I checked out several other Ada open source projects,  
> and that seemed to be the approach taken elsewhere. If it makes sense to
> release in another manner, I am open to considering that. One issue I  
> think with releasing in a public version control system, is choosing  
> which one. There are a number of them out there.
You can safely keep going with a single archive: that's handy for every  
one (no need for a repository war the same there was a browser war). Also  
an archive is better suited for personal storage (at the user side) than a  
versioning system is. Unless one plan to be involved in the development, a  
versioning system tree is more uselessly-heavy and encumbered than useful.

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.

“I am fluent in ASCII” [Warren 2010]

Re: Paraffin: Parallelism generics for Ada 2005

[Brad Moore]

On 25/01/2011 6:59 PM, Yannick Duchêne (Hibou57) wrote:
> Le Wed, 26 Jan 2011 02:53:39 +0100, Brad Moore <brad.moore@shaw.ca> a
> écrit:
>> No particular reason, I suppose. This is my first release of a project
>> on source forge. I checked out several other Ada open source projects,
>> and that seemed to be the approach taken elsewhere. If it makes sense to
>> release in another manner, I am open to considering that. One issue I
>> think with releasing in a public version control system, is choosing
>> which one. There are a number of them out there.
> You can safely keep going with a single archive: that's handy for every
> one (no need for a repository war the same there was a browser war).
> Also an archive is better suited for personal storage (at the user side)
> than a versioning system is. Unless one plan to be involved in the
> development, a versioning system tree is more uselessly-heavy and
> encumbered than useful.
>

Good points. The single archive approach seemed easiest to me,
and it still sounds like the way to go.

Re: Paraffin: Parallelism generics for Ada 2005

[R. Tyler Croy]

Yannick Duchêne (Hibou57) <yannick_duchene@yahoo.fr> wrote:
> Le Wed, 26 Jan 2011 02:53:39 +0100, Brad Moore <brad.moore@shaw.ca> a  
> écrit:
>> No particular reason, I suppose. This is my first release of a project  
>> on source forge. I checked out several other Ada open source projects,  
>> and that seemed to be the approach taken elsewhere. If it makes sense to
>> release in another manner, I am open to considering that. One issue I  
>> think with releasing in a public version control system, is choosing  
>> which one. There are a number of them out there.
> You can safely keep going with a single archive: that's handy for every  
> one (no need for a repository war the same there was a browser war). Also  
> an archive is better suited for personal storage (at the user side) than a  
> versioning system is. Unless one plan to be involved in the development, a  
> versioning system tree is more uselessly-heavy and encumbered than useful.

O_o

I've heard a lot of things, but I don't think I've ever heard an argument
against version control before!

I personally don't care which kind of repository somebody uses, but I've found
myself after very weary of the "just download this tarball" approach after
years of finding interesting looking projects which have stagnated and become
completely out of date that have no publicly visible source history (student or
university projects are notoriously bad on this).

I don't mean to get too far off-topic, to each their own, I just find it odd
that one would go through the trouble of creating a SourceForge project just to
distribute tarballs.

-- 
- R. Tyler Croy
--------------------------------------
    Code: http://github.com/rtyler

Re: Paraffin: Parallelism generics for Ada 2005

[Simon Wright]

"R. Tyler Croy" <tyler@linux.com> writes:

> I've heard a lot of things, but I don't think I've ever heard an
> argument against version control before!
>
> I personally don't care which kind of repository somebody uses, but
> I've found myself after very weary of the "just download this tarball"
> approach after years of finding interesting looking projects which
> have stagnated and become completely out of date that have no publicly
> visible source history (student or university projects are notoriously
> bad on this).
>
> I don't mean to get too far off-topic, to each their own, I just find
> it odd that one would go through the trouble of creating a SourceForge
> project just to distribute tarballs.

I completely agree with this.

Some sort of VCS is vital in any development, even if on one machine
only; and nowadays there are distributed VCSs which give even a
single-handed development the benefit of VC both locally (eg, while
disconnected from the network) and on a network host. [Not so say there
aren't other advantages.]

SourceForge supports Mercurial (which I like) and Git; others include
Monotone and Bazaar.

Re: Paraffin: Parallelism generics for Ada 2005

[Brad Moore]

On 26/01/2011 1:27 PM, Simon Wright wrote:
> "R. Tyler Croy"<tyler@linux.com>  writes:
>
>> I've heard a lot of things, but I don't think I've ever heard an
>> argument against version control before!
>>
>> I personally don't care which kind of repository somebody uses, but
>> I've found myself after very weary of the "just download this tarball"
>> approach after years of finding interesting looking projects which
>> have stagnated and become completely out of date that have no publicly
>> visible source history (student or university projects are notoriously
>> bad on this).
>>
>> I don't mean to get too far off-topic, to each their own, I just find
>> it odd that one would go through the trouble of creating a SourceForge
>> project just to distribute tarballs.
>
> I completely agree with this.
>
> Some sort of VCS is vital in any development, even if on one machine
> only; and nowadays there are distributed VCSs which give even a
> single-handed development the benefit of VC both locally (eg, while
> disconnected from the network) and on a network host. [Not so say there
> aren't other advantages.]
>
> SourceForge supports Mercurial (which I like) and Git; others include
> Monotone and Bazaar.

I also think version control is important, so I am not arguing against 
it. In fact, the tar-ball approach itself can be a form of version 
control, so long as the older tar balls are still accessible, and the 
newer ones are numbered so that its easy to see the sequence.

I am currently using git for version control on all my stuff, so at 
least I'm able to see the history of each file.

Whether the real version control should be publicly visible is another 
question.
I think the benefits for that depends more on the nature of the project.
If its a larger project being developed and maintained by an active 
online community then I would think a public version control would be 
beneficial.

If it is a smaller project however, then it may not be worth the trouble
setting the version control up, if the only person maintaining the 
software is the author.

For example if my project were to somehow pickup momentum and people
were expressing interest in getting involved then I would consider 
setting up the version control for that project to facilitate the
development process. Until I see a need, I would likely opt to go with 
the simpler approach.

Re: Paraffin: Parallelism generics for Ada 2005

[Yannick Duchêne (Hibou57)]

Le Wed, 26 Jan 2011 05:50:10 +0100, R. Tyler Croy <tyler@linux.com> a  
écrit:
>> You can safely keep going with a single archive: that's handy for every
>> one (no need for a repository war the same there was a browser war).  
>> Also
>> an archive is better suited for personal storage (at the user side)  
>> than a
>> versioning system is. Unless one plan to be involved in the  
>> development, a
>> versioning system tree is more uselessly-heavy and encumbered than  
>> useful.
>
> O_o
>
> I've heard a lot of things, but I don't think I've ever heard an argument
> against version control before!
I talked with some precise rationals given for precise usages and context.  
This was not advocating against CVS in a fundamental way.

-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.

“I am fluent in ASCII” [Warren 2010]