GSL collaboration 
Author Message
 GSL collaboration

Aaron's final comment seems to have got lost in the general discussion:

Quote:
>Is anyone interested in forming a collaborative project to provide
>the gsl extension to pop-11???

Well I certainly am!  Any other volunteers?

--
Steve



Sun, 09 Nov 2003 18:26:15 GMT  
 GSL collaboration
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



Quote:
>Aaron's final comment seems to have got lost in the general discussion:
>>Is anyone interested in forming a collaborative project to provide
>>the gsl extension to pop-11???

>Well I certainly am!  Any other volunteers?

I am willing to contribute.  

I am also interested in integrating a Poplog subsystem into applications
written in other languages. In particular, a MOD_POPLOG for Apache would
be useful, as would a mechanism to use Pop-11 from within C/C++, similar
to the mechanism for embedding a TCL interpreter within a C/C++ program.
I would also like to build CORBA and Active-X objects which embed
Poplog. Is anyone else interested in these projects?

Regards,
- --

-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.7.1

iQA/AwUBOwvT0/Hj+enJbeYqEQKLhQCeKizpKCmEpDL6SLGmnhds0S/LNakAoNyg
ddWlU6Uz42jqEVGlvvwchmrW
=Q+F2
-----END PGP SIGNATURE-----



Mon, 10 Nov 2003 07:12:50 GMT  
 GSL collaboration
Hi,

Quote:

>  I am also interested in integrating a Poplog subsystem into applications
>  written in other languages. In particular, a MOD_POPLOG for Apache would
>  be useful,

Yes, I'm up for this one too!  In fact I'm willing to devote a _lot_
of time to this project.

I had a go at this a little while back.  POPC claims that it is
possible to create shared libraries - which it is - but it appears to
be less than straightforward to include the compiler undamaged.  It
_must_ be possible since this is the same mechanism which builds
Poplog itself - but I ran out of time.

Quote:
>  I would also like to build CORBA and Active-X objects which embed
>  Poplog. Is anyone else interested in these projects?

Not sure what "embed Poplog" means.

--
Steve



Mon, 10 Nov 2003 17:18:51 GMT  
 GSL collaboration
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



Quote:
>Hi,


>>  I am also interested in integrating a Poplog subsystem into applications
>>  written in other languages. In particular, a MOD_POPLOG for Apache would
>>  be useful,

>Yes, I'm up for this one too!  In fact I'm willing to devote a _lot_
>of time to this project.

I am a month into a 6-month R&D sabbatical in Canada, so I have a chunk
of time with no clients pressing for my attention, and some spare hours
every week for non-core projects. Sadly, although I shipped a spare PC
and an 82GB Linux disk, the two don't like each other. For the time-
being, I only have access to the Windows 2000 and Windows 95 volumes. I
have trial installations of U/WIN and Cygwin under W2K, and I am trying
to get Apache, PHP and mySQL working also. This limits the test
environments I can use, unless I grab a few GB from one of the spare
drives to install another SuSE instance.

What environments do you have access to for testing? How do you want to
handle collaboration?

Quote:
>I had a go at this a little while back.  POPC claims that it is
>possible to create shared libraries - which it is - but it appears to
>be less than straightforward to include the compiler undamaged.  It
>_must_ be possible since this is the same mechanism which builds
>Poplog itself - but I ran out of time.

My understanding is that POPC doesn't work under Windows. I'd like to
change this, if it is possible.

Quote:
>>  I would also like to build CORBA and Active-X objects which embed
>>  Poplog. Is anyone else interested in these projects?

>Not sure what "embed Poplog" means.

In this case, a wrapper around Poplog that has either ActiveX or CORBA
interfaces. This could be as simple as an object with pipes to an
instance of a Poplog process, handling an input stream sent from the
object, but I'd hope we could achieve a more optimal solution.

Regards,
- --

VoIP# 446043000002

-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.7.1

iQA/AwUBOwzc0PHj+enJbeYqEQJmZQCg7VKi87IeZFxPbemIXYAZW/zRZMIAnipY
j+IRVnVR89dmMLh7267IpBy4
=ZrYl
-----END PGP SIGNATURE-----



Tue, 11 Nov 2003 02:03:02 GMT  
 GSL collaboration

Quote:
>Aaron's final comment seems to have got lost in the general discussion:
>>Is anyone interested in forming a collaborative project to provide
>>the gsl extension to pop-11???

My request for explanation/example of 'gsl extension to pop-11'
received no reply.

My initial guess that it refered to greater mathematical functionality,
is apparently wrong ?   A superficial look suggests that there's already
substantial maths capability.

Since gsl is merely a legal consideration, what technical extensions
was Aaron suggesting ?

Apparently over time, poplog users consider it to be their main
general-purpose language/tool ?
This contrasts to what drew me to poplog: by far the best available
(a must have) tool for AI.   I'm disapointed that the unique AI capabilities
that poplog offers are submerged by emphasis on 'low level' considerations.
I'm reminded of (decades ago) motor enthusiasts who were continually
'tuning their machine' and 'improving on the original designers' effort',
yet never got around to making a substantial journey.

I've observed the social/psycological aspect of the 'development' of
the OS/language oberon, which is also university based and has
outside collaboration:
   with a frequency of multiples of the stock market irrational cycles,
     enthusiasm arises to redesign the compiler with some bleeding-edge
     optimisation, which would save me 2 milliseconds a year.
Yet vast numbers of applications remain unusable for lack of documentation.

Poplog beginners are lucky that substantial documentation exists.

All poplog users benefit by making it easier for new entrants - who
become potential contributors.  Nobody can do the unpleasant task
of proof reading and detailed testing better than the ignorant (unbaised)
beginner.  It just needs a bit of 'management from above to harness the
man power'.

I guess this whole topic "dynamics of dispersed collaboration" will be a
recongised  'subject' soon  -- of growing importance due to inet.
Apparently linux was/is created by dispersed collaboration.

Chris Glur.

PS. lets have a formal definition/proposal for the gsl extensions.
I stick by my (disputed) view that poplog has taken the wrong route
in being informal:
   "AI pretends to imitate humans; hence welcomes ambiguity and
   redundancy".

Aviation wasn't advance by imitating the birds, with feathers and flapping.



Tue, 11 Nov 2003 15:11:08 GMT  
 GSL collaboration
[To reply replace "Aaron.Sloman.XX" with "A.Sloman"]

Quote:

> Date: 25 May 2001 07:11:08 GMT
> Organization: The South African Internet Exchange

> My request for explanation/example of 'gsl extension to pop-11'
> received no reply.

Apologies: I meant to reply earlier, but I am trying to do too many
things and cannot keep up with all of them as well as I'd like.

Quote:
> My initial guess that it refered to greater mathematical functionality,
> is apparently wrong ?   A superficial look suggests that there's already
> substantial maths capability.

Pop-11 has quite a lot of mathematical capabilities (including
support for integers, biginters, ratios, single floats, double floats)
with some standard mathematical and trigonometric functions defined over
all of those, as described in REF NUMBERS, browsable here without Ved:

    http://www.*-*-*.com/

But that's still a tiny subset of what users may require. Of course, any
required algorithms can be coded in pop-11, but that requires work, and
in some cases the resulting code might run 5 to 10 times slower than the
same thing implemented in C. (Has anyone done any speed comparisons
recently between Pop-11 and C?). People whose programs already run for
hours, or days or weeks (e.g. programs training neural nets, or doing
evolutionary exploration, etc.) could benefit a lot from the increased
speed.

What I was proposing can be compared with two things that already exist
in Pop-11:

(a) The X window system provides a large collection of utilities for
managing graphical displays along with mouse and keyboard events,
written in C. The poplog developers wrote lots of C code, and Pop-11
code to invoke the C code, along with a lot of documentation on these
facilities in $usepop/pop/x/pop/ref/ browsable without Ved in
    http://www.*-*-*.com/

as a result of which pop-11 programmers can conveniently invoke X
utilities.

These are used in the editor XVed and the facilities described in TEACH
RC_GRAPHIC, as well as my RCLIB package (rclib.tar.gz) available from
the poplog site.

(b) David Young at Sussex University produced a library with a
collection of utilities and documentation for using or learning about AI
vision techniques.

His popvision library, browsable here
    http://www.*-*-*.com/
or retrievable here:
    http://www.*-*-*.com/

includes a number of image manipulation programs written in C.

To make them accessible for users he has Pop-11 wrappers. The
popvision/lib/ subdirectory has a bunch of files some defining C
programs and some defining Pop-11 procedures to invoke the C programs.

E.g. if you compare these two files

http://www.*-*-*.com/
http://www.*-*-*.com/

The first one defines a lot of C procedures for doing things on arrays,
and the second one defines corresponding Pop-11 procedures which invoke
the C procedures. e.g.

in C:
void float_arraywtdav_mask(alpha1, alpha2, arr1, arr2, mask, arr3, n)

in Pop-11:
define float_arraywtdav_mask(alpha1, alpha2, arr1, arr2, mask, arr3) -> arr3;
    ;;; Averages arr1 and arr2 into arr3. Alpha1 is the weight attached
    ;;; to arr1 when mask is zero, and alpha2 is the weight when mask is
    ;;; non-zero.
    ;;; Rounds the results, and clips to 0 to 255

This means that people who want to use these image manipulating
procedures (a) have the algorithms already coded and (b) can get the
speed of C combined with the convenience of Pop-11.

There are also some neural net utilities available like that in
    http://www.*-*-*.com/
    http://www.*-*-*.com/

The point about the GSL library is that there is a HUGE collection of
generally useful utilities coded in C. Making them easily accessible
from Pop-11 would be very useful for many applications in AI and other
fields.

Quote:
> Since gsl is merely a legal consideration, what technical extensions
> was Aaron suggesting ?

I don't know what you mean by "merely a legal consideration".

Perhaps you thought I was proposing copying the algorithms into Pop-11 ?

Quote:
> Apparently over time, poplog users consider it to be their main
> general-purpose language/tool ?

Some people. E.g. all my programming, email work, news reading and
posting, use of{*filter*}to write papers, creation of html files, etc. is
done in Ved, since any missing facility I can provide by writing pop-11
utilities, which will then integrate with the rest of my environment.

More people do that with Emacs, using Emacs lisp, however.

Quote:
> This contrasts to what drew me to poplog: by far the best available
> (a must have) tool for AI.   I'm disapointed that the unique AI capabilities
> that poplog offers are submerged by emphasis on 'low level' considerations.

I don't understand. Where are they submerged?

AI is a huge field, using many different kinds of techniques: theorem
proving, sentence analysis, image analysis and interpretation, neural
nets, evolutionary computation, constraint propagation, symbolic problem
solving, robot-control mechanisms, etc. etc. For overviews of AI see

    http://www.*-*-*.com/ ~axs/courses/ai.html
    http://www.*-*-*.com/ ~axs/misc/aiforschools.html

Not all the GSL facilities would be equally useful for AI, but I would
expect a significant subset to be.

More importantly, if Pop-11 +GSL becomes available as a free, open
source, alternative to some expensive packages for convenient
interactive development using powerful mathematical facilities, then
perhaps a lot more people might use it and there could be a lot more
support for future developments that would benfit all users.

This would not rule out other kinds of developments. For instance work
is still continuing on the sim_agent toolkit which is also available
from the poplog site,

    http://www.*-*-*.com/ ~axs/cogaff/simagent.html

and is used for quite a lot of student AI projects as well as AI
research.

Incidentally, there is already quite a lot of language processing stuff
available in the contrib library, as documented in the books on Natural
Language processing by Gazdar and Mellish.

    http://www.*-*-*.com/

Making new general purpose facilities available through pop-11 is not
incompatible with parallel development work concerned with specialised
AI domains.

Quote:
> PS. lets have a formal definition/proposal for the gsl extensions.

The people doing the work would certainly have to do that. In fact I
would hope that a significant amount of the work of generating Pop-11
wrappers for GSL utilities could be automated, and that would be
impossible without formalisation.

Steve Leach and Jeff Best have expressed an interest in doing this and
both are experienced software engineers, with a lot of first hand
experience of real world development.

Quote:
> I stick by my (disputed) view that poplog has taken the wrong route
> in being informal:

Well feel free to make proposals for new developments to fix that, and
you may find collaborators!

Quote:
>    "AI pretends to imitate humans; hence welcomes ambiguity and
>    redundancy".

What exactly are you referring to?

Incidentally a redundant and therefore more readable syntax need not be
at all ambiguous.

Quote:
> Aviation wasn't advance by imitating the birds, with feathers and flapping.

A propeller is just a wing that rotates instead of flapping.

The principles of aerodynamics remain the same for aeroplanes and for
birds! (Not for jet-engines and rockets.)

cheers.

Aaron
====
Aaron Sloman, ( http://www.*-*-*.com/ ~axs/ )
School of Computer Science, The University of Birmingham, B15 2TT, UK

PAPERS: http://www.*-*-*.com/
FREE TOOLS: http://www.*-*-*.com/



Wed, 12 Nov 2003 18:17:45 GMT  
 
 [ 7 post ] 

 Relevant Pages 

1. GSL collaboration and (and GSL Manual contents)

2. GSL collaboration and formal methods.

3. GSL collaboration

4. filenames, argument list separators + gsl extension

5. Other gsl interfacing projects

6. GSL interface: another Pop-11 development task?

7. GSL and Lapack

8. ANN: ruby-gsl 0.1.6

9. ruby-gsl 0.1

10. ada bindings for gsl (GNU Scientific Library)

11. gfoftran and GSL subroutines

12. Fortran interface for GSL (alpha)

 

 
Powered by phpBB® Forum Software