Startup Poplog Inquiries 
Author Message
 Startup Poplog Inquiries

I was just trying to download free Poplog from

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

using the Linux with Motif option but
nothing is downloaded, and when I try to connect to

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

I get connection refused.  Perhaps it is a temporary
problem.

I am thinking about Poplog because it includes common
AI languages: Pop-11, Lisp, Prolog, ML, (and Scheme?),
has a system command subsystem (hopefully for bash too
or can be made so), and has socket capability.

The idea, and please let me know how far wrong I am, is
to use Poplog to build an AI that would control all or
all significant software applications on my computer
inluding Internet access. E.g., C and C++ are very primary
languages and it seems as if there must be a way of, say,
building a C program in Poplog using AI techniques, and
then compiling and executing the program via some
instruction in Poplog. And then we should be able to do
the same for Java, fortran, Octave, etc. And then there
should be a method of passing parameters or information
between these various language subsystems.

Thank you,

Neil Nelson



Sat, 25 Oct 2003 08:42:03 GMT  
 Startup Poplog Inquiries
I was just looking the Sneps package that is looking
for CLISP or CMU LISP or some others.  It does not
appear to be directly mentioning Lisp under Pop-11.
Would it be the case that a package running under
these other very common implementations of Lisp will
run without problem under Pop-11 Lisp?

Regards,

Neil Nelson



Sat, 25 Oct 2003 11:06:04 GMT  
 Startup Poplog Inquiries

Quote:

> I was just trying to download free Poplog from

> http://www.poplog.org/resources/freepoplog.html#linux

> using the Linux with Motif option but
> nothing is downloaded, and when I try to connect to

> http://www.cs.bham.ac.uk/research/poplog/freepoplog.html

> I get connection refused.  Perhaps it is a temporary
> problem.

It was a temporary problem.  I am downloading now from
http://www.cs.bham.ac.uk/research/poplog/new/linuxmotif1553.tar.gz

Regards,

Neil Nelson



Sat, 25 Oct 2003 21:03:05 GMT  
 Startup Poplog Inquiries
[To reply replace "Aaron.Sloman.XX" with "A.Sloman"]


Quote:
> article: 2739 in comp.lang.pop
> Date: Mon, 07 May 2001 17:42:03 -0700
> Organization: SBC Internet Services

> ...
> I get connection refused.  Perhaps it is a temporary
> problem.

Yes, from early afternoon yesterday (07 May) till this morning
there was a problem on the MidMan network (which includes
several universities in the West Midlands area.
As you have found, this is now working again.

You should be able to connect OK  now.

Quote:
> I am thinking about Poplog because it includes common
> AI languages: Pop-11, Lisp, Prolog, ML, (and Scheme?),
> has a system command subsystem (hopefully for bash too
> or can be made so), and has socket capability.

The Pop-11 sysexecute procedure can be used (with sys_fork or sys_vfork)
to run arbitrary unix programs, including bash, sh, etc. The sysobey
procedure already accepts syntax to run commands in a user-specified
shell.

Both are described in the REF sysutil file,
    $usepop/pop/ref/sysutil

Also browsable here
    http://www.cs.bham.ac.uk/research/poplog/doc/popref/sysutil

Unfortunately most of the online poplog documentation uses a "private"
format that makes the files look nice in the Poplog editor Ved (or
XVed), by screws up in most editors e.g. emacs and nedit, though you can
read them in a unix xterm window using "more". The doc/ directory at the
poplog web site has "stripped" versions of all the files, which should
be readable in any text editor.

If you are an emacs user you may find it useful to fetch a packaged
designed to improve the interaction between emacs and poplog. It's
here:
    http://www.cs.bham.ac.uk/research/poplog/emacs.tar.gz

There are several other packaged options for invoking unix programs from
pop-11 including this
    run_unix_program
described in its own help file or

    http://www.cs.bham.ac.uk/research/poplog/doc/pophelp/run_unix_program

It covers many cases, but unfortunately not all!

Quote:
> The idea, and please let me know how far wrong I am, is
> to use Poplog to build an AI that would control all or
> all significant software applications on my computer
> inluding Internet access.

Sounds very ambitious.

But if you can do it in C or C++ you should be able to do it in pop-11.

E.g. I do all my work in the poplog editor Ved, which is implemented in
Pop-11, and has been extended to run arbitary unix programs and read
results in a ved buffer, to read and post mail, read and post news,
browse directories, invoke a web browser spawn X utilities, etc.

Quote:
> E.g., C and C++ are very primary
> languages and it seems as if there must be a way of, say,
> building a C program in Poplog using AI techniques, and
> then compiling and executing the program via some
> instruction in Poplog. And then we should be able to do
> the same for Java, Fortran, Octave, etc. And then there
> should be a method of passing parameters or information
> between these various language subsystems.

That is not so easy. You can in principle use pop-11 to assemble program
source code and then hand a file to the relevant compiler. Running the
resulting program from inside pop-11 may be easy or difficult or
impossible (e.g. Java, C++ probably can't be done though you can always
run programs as separate unix processes and communicate with them via
pipes or sockets).

However, automatic program synthesis is a very old and difficult area of
AI. E.g. see

    Charles Rich and Richard C. Waters. The Programmers Apprentice.
       Frontier Series. ACM Press, 1990.

There are probably lots more things to read, but I don't have a
bibliography. You could try searching in google, etc. for
    program synthesis
    automatic programming
    programmer's apprentice

etc.

Second message

Quote:
> Date: Mon, 07 May 2001 20:06:04 -0700
> Organization: SBC Internet Services

> I was just looking the Sneps package that is looking
> for CLISP or CMU LISP or some others.  It does not
> appear to be directly mentioning Lisp under Pop-11.
> Would it be the case that a package running under
> these other very common implementations of Lisp will
> run without problem under Pop-11 Lisp?

Usually a small amount of work is required to port programs from one
lisp implementation to another, depending on how complex the program is.

Poplog lisp implements most of common lisp as defined in the second
edition of Common Lisp the Language (G.Steele) but not all. For
details look in
    $usepop/pop/lisp/help/bugs

also accessible as
    http://www.cs.bham.ac.uk/research/poplog/doc/lisphelp/bugs

I hope that helps.

Aaron
==
====
Aaron Sloman, ( http://www.cs.bham.ac.uk/~axs/ )
School of Computer Science, The University of Birmingham, B15 2TT, UK

PAPERS: http://www.cs.bham.ac.uk/research/cogaff/
FREE TOOLS: http://www.cs.bham.ac.uk/research/poplog/freepoplog.html



Sat, 25 Oct 2003 23:11:28 GMT  
 Startup Poplog Inquiries

 > However, automatic program synthesis is a very old
 > and difficult area of AI.

Yes you are quite correct and of course for the reason
that a search of more than a trivially small algorithm
space is computationally unfeasible. Several strategies
may be employed to reduce the search space:

(1) If you already have the desired algorithm on file,
you only need to provide an efficient index to the algo-
rithm. That is not particularly simple in a comprehen-
sive way, but it is much easier than trying to find a
complex algorithm through a brute force search, which is
generally impossible (by a very significant margin). But
the primary point here is that the algorithm store be as
large as possible that then requires a multi-language
approach because of the many languages in which algo-
rithms are now and will be available. Any reasonably
large available solution set not included could
seriously degrade subsequent performance because of
the extraordinary difficulty of trying to find those
missing solutions using a direct search.

(2) An analysis of existing solutions should identify
common structures that would reduce direct searches in
the algorithm space. This is the rationale of common
education.

(3) Goal paths must be divided to a sufficiently fine
degree using existing solutions or close approximations
to allow direct searches when they cannot be avoided.

(4) Goals must be prioritized according to their likeli-
hood of success and expected effort required.

(5) Searches should be conducted using all the available
human and machine resources but under the requirements
at (4) and tailored to those resources expected best
able to solve that search/problem.

But I need to do a lot of study and work and do not
want to waste much of your time on a very early pie-
in-the-sky idea at this stage.

I will certainly be looking at your very comprehensive
tutorial and posts here to get up to speed.

Regards,

Neil nelson



Sun, 26 Oct 2003 00:49:01 GMT  
 
 [ 5 post ] 

 Relevant Pages 

1. a startup file for MSWLogo - startup.lgo (0/1)

2. a startup file for MSWLogo - startup.lgo (0/1)

3. Inquiry: Desktop publishing, APL fonts

4. Brouwer inquiry

5. CinCom sales inquiries go unanswered

6. INQUIRY: PPS2.0 Help systems

7. Inquiry about cancelling autoincremental key at runtime.

8. Inquiry about cancelling autoincremental key at runtime.

9. Client - Server platform support inquiry

10. Inquiries about Clarion4

11. Employment Inquiry

12. copy protection inquiry

 

 
Powered by phpBB® Forum Software