Guile questions
Janos Blazi
newbie questions and hope that they are not off topic in this news group.
I could not make Guile work from a .cpp source file, I had to use a .c file.
Now I am going to use wxWindows in my application and so I should need c++.
Is there a way of solving this problem?
Is there a manual for Guile in some easily readable form (html,pdf,...)?
TIA,
Janos Blazi
-----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------
http://www.newsfeed.com The #1 Newsgroup Service in the World!
-----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----
Siegfried Gonzi
> I have installed Guile on Cygwin and it works fine. I should like to ask two
> newbie questions and hope that they are not off topic in this news group.
Hi:
It is not my intention to convince you; I am sure you are aware of your
tasks. But wouldn't it be better to install Linux and use a multi-boot
system Windows/Linux? I mean the Cygwin solution is insane. If you can
cope with Cygwin, it will be easy for you to cope with Linux.
You are from Germany and obtaining SUSE 8.1 Linux should not be that
hard. The installation is a matter of 20 minutes (even for a beginner).
If you have a Gericom laptop: Gericom laptops are very common in Germany
but they suffer from the problem that most of their product lines just
feature a no-name winmodem. But even there I was lucky and found a site
which said that most of the no-name winmodems are from SMARTLINK. The
SMARTLINK guys emailed me a Linux driver for my Gericom winmoden in the
following. There is no need to use Windows on my laptop if I want to use
my email account, even.
I mean programming languages and Linux/Unix ties better than programming
languages and Windows.
Regards,
S. Gonzi
Janos Blazi
> tasks. But wouldn't it be better to install Linux and use a multi-boot
> system Windows/Linux? I mean the Cygwin solution is insane. If you can
> cope with Cygwin, it will be easy for you to cope with Linux.
Thx. I use Linux. But I teach and at school we use Windows. There is
absolutely no chance of changing this. So I have to write an Windows
standalone application.
Besides, using Linux would not help, the problems were the same.
Jens Axel Søgaard
>>It is not my intention to convince you; I am sure you are aware of your
>>tasks. But wouldn't it be better to install Linux and use a multi-boot
>>system Windows/Linux? I mean the Cygwin solution is insane. If you can
>>cope with Cygwin, it will be easy for you to cope with Linux.
>
>
> Thx. I use Linux. But I teach and at school we use Windows. There is
> absolutely no chance of changing this. So I have to write an Windows
> standalone application.
> Besides, using Linux would not help, the problems were the same.
Why don't you choose a Scheme implementation that runs on both Windows
and Linux?
--
Jens Axel Søgaard
JB
Here is what I should like to do:
I should write a c++ program, that uses wxWindows (which is
a GUI toolkit), Core (which is an esoteric MP arithmetic
library) and then the scripting language Scheme.
I am having tecnical problems with Guile and I should prefer
a Scheme that offers macros anyway. (I do not know if Guile
has macro support.) What I should be be able to
do, are calls like
(define g (line '(3 4) '(4 5)))
(define h (line '(-1 2) '(3 0)))
(define p (section g h))
(draw g h p)
Then the Scheme functions 'line' and 'section' and 'draw'
would call C++ functions that would do the job.
MzScheme look interesting and I shall take a closer look at
it. Didi I understand the user manual correctly: I should
need embedding and extension at the same time?
--
JB
Jens Axel Søgaard
get a copy in e-mail.]
JB wrote:
> Here is what I should like to do:
> I should write a c++ program, that uses wxWindows (which is
> a GUI toolkit), Core (which is an esoteric MP arithmetic
> library) and then the scripting language Scheme.
> Then the Scheme functions 'line' and 'section' and 'draw'
> would call C++ functions that would do the job.
DrScheme actually uses wxWindow to make its GUI portable :-)
To me it would seem simpler to use Scheme as the main language
(that is for both GUI and scripting) then you only need to
make an extension for Core.
One could either use the FFI:
http://download.plt-scheme.org/doc/202/html/mzc/mzc-Z-H-2.html#%_chap_2
or use SWIG http://www.swig.org
> MzScheme look interesting and I shall take a closer look at
> it. Didi I understand the user manual correctly: I should
> need embedding and extension at the same time?
As far as can tell, extension simply means you provide extra
primitives to mzscheme. You therefore need extensions if you
want to call your own primitives from within Scheme. Regardless
of whether you choose to embed or not.
Out of curiosity: What is Core and where can I read about it?
For once I had no success using Google.
--
Jens Axel Søgaard
JB
> DrScheme actually uses wxWindow to make its GUI portable
> :-) To me it would seem simpler to use Scheme as the main
> language (that is for both GUI and scripting) then you
> only need to make an extension for Core.
The program would have to run on a computer without any
sensible infrastructure. I am afraid I cannot make
executables with Scheme. (This is for example a serious
problem with CL.)
> One could either use the FFI:
>
>
http://download.plt-scheme.org/doc/202/html/mzc/mzc-Z-H-2.html#%_chap_2
>
> or use SWIG http://www.swig.org
That would be horrible.
>> MzScheme look interesting and I shall take a closer look
>> at it. Didi I understand the user manual correctly: I
>> should need embedding and extension at the same time?
>
> As far as can tell, extension simply means you provide
> extra primitives to mzscheme. You therefore need
> extensions if you want to call your own primitives from
> within Scheme. Regardless of whether you choose to embed
> or not.
So I should really need both.
> Out of curiosity: What is Core and where can I read about
> it? For once I had no success using Google.
Core is a MP arithmetic package. When you deal with
geometry, you need precise calculations, that is, the first
three or four digits of your results should be exact. Core
offers such precision. (The other package I know about,
that offers the same precision, is LEDA, but it is not
free.)
Look at http://www.cs.nyu.edu/exact/core/
The strange thing is, that from the three components I
should need, that is Guile/Scheme, wxWindows and Core, only
the latter is working at the moment. The Core distribution
I downloaded was slightly buggy, but the Core people helped
me...
--
JB
Jens Axel Søgaard
> That would be horrible.
It would demand some work - yes. But technically it would
be a nice solution.
> Core is a MP arithmetic package. When you deal with
> geometry, you need precise calculations, that is, the first
> three or four digits of your results should be exact. Core
> offers such precision. (The other package I know about,
> that offers the same precision, is LEDA, but it is not
> free.)
> Look at http://www.cs.nyu.edu/exact/core/
This looks interesting indeed. The ability to calculate
as many decimals as needed is very appealing and elegant.
Hm. Reminds me. Have you checked out SISC? Is a
Scheme in Java:
http://sisc.sourceforge.net/manual/html/ch03.html#Numbers
SISC (1.6.0-beta) - main
> (max-precision 1000)
> (min-precision 1000)
> (sqrt 2)
1.4142135623730950488016887242096980785696718753
769480731766797379907324784621070388503875343276
415727350138462309122970249248360558507372126441
214970999358314132226659275055927557999505011527
820605714701095599716059702745345968620147285174
186408891986095523292304843087143214508397626036
279952514079896872533965463318088296406206152583
523950547457502877599617298355752203375318570113
543746034084988471603868999706990048150305440277
903164542478230684929369186215805784631115966687
1301301561856898723723
The on-line demo is at
http://sisc.sourceforge.net/sisc-online.html
I know this is not the same thing as Core, but
it is an easy start. Note that you can call Java
functions from within SISC - and thus make the GUI
using the standard Java API.
--
Jens Axel Søgaard
Friedrich Dominicus
> Jens Axel Søgaard wrote:
>
> > DrScheme actually uses wxWindow to make its GUI portable
> > :-) To me it would seem simpler to use Scheme as the main
> > language (that is for both GUI and scripting) then you
> > only need to make an extension for Core.
>
> The program would have to run on a computer without any
> sensible infrastructure. I am afraid I cannot make
> executables with Scheme. (This is for example a serious
> problem with CL.)
You don't mean this seriously do you? You can make executable at least
with LispWorks and Franz Allegro Lisp.
Friedrich
Siegfried Gonzi
> You don't mean this seriously do you? You can make executable at least
> with LispWorks and Franz Allegro Lisp.
Let's be serious: he needs his stuff for a school. They are not
braindamaged that they will spend many thousands of dollars for a
LispWorks or Franz Allegro distribution; an excecutable requires the
full licence.
Not going to say I am against paying for software (I paid the freeware
license for my Macintosh LateX distribution, though, they would have
been free if I had used it that).
S. Gonzi
Friedrich Dominicus
> Friedrich Dominicus wrote:
>
>
> > You don't mean this seriously do you? You can make executable at least
> > with LispWorks and Franz Allegro Lisp.
>
> Let's be serious: he needs his stuff for a school. They are not
> braindamaged that they will spend many thousands of dollars for a
> LispWorks or Franz Allegro distribution; an excecutable requires the
> full licence.
For what should thy need executable than? And I bet they could get a
decent offer from both Xanalys and Franz if they would ask for it.
But I won't argue about it any longer. this is c.l.scheme isn't it?
Regards
Friedrich
JB
> You don't mean this seriously do you? You can make
> executable at least with LispWorks and Franz Allegro Lisp.
First, those great products are not free. Though we are a
rich country, my school is so poor that we cannot even
afford to buy blackboards with grids on them.
And the second point is that I am not sure if they could be
embedded. I thought more of ECL Lisp when I wrote about CL
but there are several problems with that option too.
I also investigated OpenLisp and TISL.
There are always some problems with these products.
--
JB
JB
> Hm. Reminds me. Have you checked out SISC? Is a
> Scheme in Java:
>
> http://sisc.sourceforge.net/manual/html/ch03.html#Numbers
>
> SISC (1.6.0-beta) - main
> > (max-precision 1000)
> > (min-precision 1000)
> > (sqrt 2)
>
> 1.4142135623730950488016887242096980785696718753
> 769480731766797379907324784621070388503875343276
> 415727350138462309122970249248360558507372126441
> 214970999358314132226659275055927557999505011527
> 820605714701095599716059702745345968620147285174
> 186408891986095523292304843087143214508397626036
> 279952514079896872533965463318088296406206152583
> 523950547457502877599617298355752203375318570113
> 543746034084988471603868999706990048150305440277
> 903164542478230684929369186215805784631115966687
> 1301301561856898723723
I had seen this. Clisp offers the same precision. (I think I
had checked virtually everything.)
--
JB
Rainer Joswig
> > You don't mean this seriously do you? You can make executable at least
> > with LispWorks and Franz Allegro Lisp.
>
> Let's be serious: he needs his stuff for a school. They are not
> braindamaged that they will spend many thousands of dollars for a
> LispWorks or Franz Allegro distribution; an excecutable requires the
> full licence.
The academic license for LispWorks for Windows Professional is
$540 dollars and includes royalty-free application distribution.
The academic license for LispWorks for Windows Enterprise is $1680.
http://www.lispworks.com/buy/index.html
From http://www.lispworks.com/products/lww.html :
Professional Edition
The Professional Edition includes everything you need for
commercial Common Lisp software development and
application delivery. CLIM 2.0 is included to further increase
program portability. Applications developed with the
Professional edition can be distributed free of charge.
Enterprise Edition
The Enterprise Edition includes all the features of the
Professional Edition, and provides further support for software
needs of the modern enterprise, including:
+ Database access through object-oriented SQL/ODBC
libraries.
+ Industry standard distributed computing through LispWorks
ORB.
+ Expert system programming through KnowledgeWorksTM
which has an embedded Prolog compiler.
Distribution of applications developed with the Enterprise
Edition in version 4.2 and above no longer require the payment
of a runtime fee.
Friedrich Dominicus
>
> First, those great products are not free. Though we are a
> rich country, my school is so poor that we cannot even
> afford to buy blackboards with grids on them.
Well without having the need for "excetutables" you can use the
personal editions. With LispWorks even CAPI comes along bundeled with
it. So you can write either Windows programs and/or Linux Programs.
>
> And the second point is that I am not sure if they could be
> embedded. I thought more of ECL Lisp when I wrote about CL
> but there are several problems with that option too.
Strange you post you have problems with Guile too. What is your
problem with ECL than?
> I also investigated OpenLisp and TISL.
> There are always some problems with these products.
Well as you notices those are "free". You're "free" to improve on
them.
Regards
Friedrich
JB
> Siegfried Gonzi <siegfri...@kfunigraz.ac.at> wrote in
> message news:<3DC62589...@kfunigraz.ac.at>...
>
>> > You don't mean this seriously do you? You can make
>> > executable at least with LispWorks and Franz Allegro
>> > Lisp.
>>
>> Let's be serious: he needs his stuff for a school. They
>> are not braindamaged that they will spend many thousands
>> of dollars for a LispWorks or Franz Allegro distribution;
>> an excecutable requires the full licence.
>
>
> The academic license for LispWorks for Windows
> Professional is $540 dollars and includes royalty-free
> application distribution.
Rainer, I do not know the situation of schools in Hamburg,
but in Bavaria, the communities pay for the equipment of
the schools and the state pays the slaries of the teachers.
Now the small town where I live is virtually bankrupt. They
almost had to close down their theatre last year.
So paying even $54 is completely out of question. As far as
I am concerned, I think that volunteering my time is enough
and do not want to support the community financially.
Let me tell you an example: When I want to use a coloured
chalk, I have to use thick and cheap ones as the thin
pieces of chalk are a bit more expensive and we cannot
afford it. I also have to pay for the tranparecies I use in
my classes.
I am sure that the price for an academic license is
justified, but I do not think that they have sold many
licenses in Bavaria.
--
JB
JB
> Well without having the need for "excetutables" you can
> use the personal editions. With LispWorks even CAPI comes
> along bundeled with it. So you can write either Windows
> programs and/or Linux Programs.
Is the Personal Edition of LispWorks embeddable?
>> And the second point is that I am not sure if they could
>> be embedded. I thought more of ECL Lisp when I wrote
>> about CL but there are several problems with that option
>> too.
> Strange you post you have problems with Guile too. What is
> your problem with ECL than?
You are writing to me as though I were not a sincere person.
You are not putting it explicitly but I have a fine antenna
for this as everybody else has, I suppose. I do not think
that I was inpolite to you even on one occasion. (Though I
must admit that I may have forgotten it.)
Now to answer you question:
A studied the manuals for ECL. At the beginning they say
that ECL /was/ designed to be embeddable but it is not
clear to me if it still is. Later in the manuals,
/extending/ ECl is discussed at some length but /embedding/
it is not mentioned any more. So I lost my courage.
For Guile, on the other hand, there is a very good tutorial.
I tried to run the programs from the tutorial and my system
crashed immediately. I you were member of the right mailing
lists, you could see that I followed up on this and finally
found out why and what happened and how to circumvent the
problem.
But I need C++ and at the time Guile only works with C. So I
shall try using extern "C" directives but this will take
time and I do not think that asking for help is in any way
negative.
>> I also investigated OpenLisp and TISL.
>> There are always some problems with these products.
> Well as you notices those are "free". You're "free" to
> improve on them.
Again the same attitude. Why are you doing this? (I remember
that I once had a problem with a German guy on c.l.l and
finally I wrote him a private letter in which I politely
asked him to shut the **** up. I do not remember whether
that were you.)
--
JB
Siegfried Gonzi
> So paying even $54 is completely out of question. As far as
> I am concerned, I think that volunteering my time is enough
> and do not want to support the community financially.
> Let me tell you an example: When I want to use a coloured
> chalk, I have to use thick and cheap ones as the thin
> pieces of chalk are a bit more expensive and we cannot
> afford it. I also have to pay for the tranparecies I use in
> my classes.
You shouldn't worry too much. I do not know of any school in a German
speaking country which would be that lunatic and pays many thousand of
dollars for a software license which will become scarcely applied.
However, universities have a budget for such license issues and they pay
if needed.
It is questionable whether school childern should be confronted with
programming languages at all. Maybe they should become exposed to some
LOGO or Scheme programming but nothing more. I know of insane beeings
who want to confront school children with C++.
You should rather teach them that programming for the sake of
programming is irrelevant.
Regards,
S. Gonzi
JB
> It is questionable whether school childern should be
> confronted with programming languages at all. Maybe they
> should become exposed to some LOGO or Scheme programming
> but nothing more.
This is of course a speculative statement as there is no way
to falsify it by experiment.
> I know of insane beeings who want to
> confront school children with C++.
I did that myself before I new that Scheme or Python existed
but now I should not do it again. At the moment I use
Python but by using Scheme you make the structure of the
problem more explicit and this is why I am going to script
my application with Scheme or another Lisp dialect and not
with Python.
--
JB
Siegfried Gonzi
> This is of course a speculative statement as there is no way
> to falsify it by experiment.
One has to be fair and concede that it all depends on the type of
education institution. As you maybe know, we in Austria have a school
system where you can decide whether you want become educated in
classical things like poetry, writing, math,...or where you become
educated in an engineering discipline.
The latter is mostly used by people who want not go to the university
thereafter, but want a harder education than the classical one. Before I
went to university I decided to go through the taff way and enrolled to
a mechanical engineering school. But I was never interested in
engineering and decided in the following, after obtaining the degree, to
jump onto a university career.
Nowadays the programming language education in schools is way more
sophisticated than it was back to 1989. Even during my university time
in physics (now I am in a PhD position) there were nearly no courses
held in programming; they expect that you learn programming in your leisure.
But I highly appreciate when teachers present Scheme to childern. Not
that I believe Scheme is the best thing since sliced bread, but it is
good to move beyond ones Java/C++ horizon.
Regards,
S. Gonzi
Jens Axel Søgaard
> Jens Axel Søgaard wrote:
>
>
>>Hm. Reminds me. Have you checked out SISC? Is a
>>Scheme in Java:
> I had seen this. Clisp offers the same precision. (I think I
> had checked virtually everything.)
Do you know the exact reals library by Boehm et al:
http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/CRCalc-force.html
It is written in Java and I imagine it would be quite easy to
call it from SISC (or Bigloo!).
--
Jens Axel Søgaard
By know you probably have guessed I don't like C++ :-)
JB
> JB wrote:
>> Jens Axel Søgaard wrote:
>>
>>
>>>Hm. Reminds me. Have you checked out SISC? Is a
>>>Scheme in Java:
>
>> I had seen this. Clisp offers the same precision. (I
>> think I had checked virtually everything.)
>
> Do you know the exact reals library by Boehm et al:
>
>
http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/CRCalc-force.html
>
> It is written in Java and I imagine it would be quite easy
> to call it from SISC (or Bigloo!).
>
I vaguely remember that have seen this before. You cannot
compare to CRs, can you? I think this is not excatly what I
am after.
At the moment I have a few possibilities I can explore. The
problem with Guile may be harder than I thought. They use
GTK+ which is in C and therefor it may not be possible to
use Guile from within C++ at all.
--
JB
Neil W. Van Dyke
MzScheme is embeddable, is extensible via shared libraries, runs on the
popular OS platforms, has an active user community, has a number of
standard and add-on libraries, is well-documented, and is freely
available.
There are several other free Scheme implementations that would also
probably do what you need.
--
Neil W. Van Dyke
http://www.neilvandyke.org/
JB
> Have you considered MzScheme
> ("http://www.plt-scheme.org/")?
>
> MzScheme is embeddable, is extensible via shared
> libraries, runs on the popular OS platforms, has an active
> user community, has a number of standard and add-on
> libraries, is well-documented, and is freely available.
>
> There are several other free Scheme implementations that
> would also probably do what you need.
>
I have just responded to Jens Axel:
"At the moment I have a few possibilities I can explore..."
and MzScheme was one of those very interesting
possibilities. It looks very good. If you read one of my
earlier postings in this thread, you can see that I am
dealing with the technical problems already:
I wrote:
"Here is what I should like to do:
I should write a c++ program, that uses wxWindows (which is
a GUI toolkit), Core (which is an esoteric MP arithmetic
library) and then the scripting language Scheme.
I am having tecnical problems with Guile and I should prefer
a Scheme that offers macros anyway. (I do not know if Guile
has macro support.) What I should be be able to
do, are calls like
(define g (line '(3 4) '(4 5)))
(define h (line '(-1 2) '(3 0)))
(define p (section g h))
(draw g h p)
Then the Scheme functions 'line' and 'section' and 'draw'
would call C++ functions that would do the job.
MzScheme look interesting and I shall take a closer look at
it. Didi I understand the user manual correctly: I should
need embedding and extension at the same time?"
--
JB
Friedrich Dominicus
>
> You are writing to me as though I were not a sincere person.
?
> You are not putting it explicitly but I have a fine antenna
> for this as everybody else has, I suppose. I do not think
> that I was inpolite to you even on one occasion. (Though I
> must admit that I may have forgotten it.)
Well to put it as clear as possible.
- you do not want to spend any money
- you do have problems with any of the free alternatives
So I pointed out more than one alternative, but all things do not
satisfy you. This does not work that is too expensive, that must
there. I call this "whining". So if you do not have the money than
take what's available for free, and invest your time.
>
> Now to answer you question:
>
> A studied the manuals for ECL. At the beginning they say
> that ECL /was/ designed to be embeddable but it is not
> clear to me if it still is. Later in the manuals,
> /extending/ ECl is discussed at some length but /embedding/
> it is not mentioned any more. So I lost my courage.
Just a few links from the documentation:
http://ecls.sourceforge.net/ecldev.html#Building%20programs
embedding C code:
http://ecls.sourceforge.net/ecldev.html#Embedding%20C%20code
here's something to read about the C-side calling Lisp:
http://ecls.sourceforge.net/ecldev.html#Objects%20representation
Than there is a mailing list in which you could ask for an example of
how to embedd it.
>
> Again the same attitude. Why are you doing this? (I remember
> that I once had a problem with a German guy on c.l.l and
> finally I wrote him a private letter in which I politely
> asked him to shut the **** up. I do not remember whether
> that were you.)
Well even if I had been that, what does it have to do with your
problem? You got more than one good suggestion. Everything was not
right, and if one had a suggestion you come with a new requirement.
This is what gets on my nerves.
Friedrich
William D Clinger
> Let's be serious: he needs his stuff for a school. They are not
> braindamaged that they will spend many thousands of dollars for a
> LispWorks or Franz Allegro distribution; an excecutable requires the
> full licence.
In my experience, it is not unusual for a school, corporation, or
other institution to balk at paying a few thousand dollars to buy
the right tool for the job, while paying its employees tens of
thousands, or hundreds of thousands, or sometimes even millions
of dollars to work around the problems that result from using the
wrong tool. Why this "brain-damage"? The tool budget may be too
small, or there may be some penny-wise/pound-foolish bean counter
whose approval is needed for each purchase, or some suit may be
prejudiced against the right tool, or whatever.
I am not suggesting that the specific case we are discussing is
an example of this, but I do not think we have enough information
to dismiss the possibility out of hand.
Will
Neil W. Van Dyke
> MzScheme look interesting and I shall take a closer look at it. Didi I
> understand the user manual correctly: I should need embedding and extension
> at the same time?"
You can embed and extend MzScheme at the same time, yes.
This is all documented in "Inside PLT MzScheme":
http://download.plt-scheme.org/doc/202/html/insidemz/index.htm
There is also some helpful example code when you instead MzScheme.
Other PLT documentation (including on the DrScheme programming
environment and the MrEd graphical toolkit) is available at:
http://download.plt-scheme.org/doc/
You may even find that PLT's existing GUI toolkit and pedagocical tools
already do what you want, or are a good starting point.
Good luck.
JB
> JB <jbl...@hotmail.com> writes:
>
>>
>> You are writing to me as though I were not a sincere
>> person.
> ?
>
>> You are not putting it explicitly but I have a fine
>> antenna for this as everybody else has, I suppose. I do
>> not think that I was inpolite to you even on one
>> occasion. (Though I must admit that I may have forgotten
>> it.)
> Well to put it as clear as possible.
> - you do not want to spend any money
> - you do have problems with any of the free alternatives
>
> So I pointed out more than one alternative, but all things
> do not satisfy you. This does not work that is too
> expensive, that must there. I call this "whining". So if
> you do not have the money than take what's available for
> free, and invest your time.
I did not ask you for your "help". I know whom you are
imitating but without the master's strange nordic charm,
technical knowledge and genuine devotion this looks less
than half as good.
Now let us proceed. I started with Guile as Guile has the
reputation of being *the* scripting language. I experienced
problems and I asked for help. You did not offer any
technical help on this.
Then other people suggested MsScheme that looked very good.
Now I admit to having overseen one of the pages in the ECL
manual and you gave a link to this page. Thank you very
much. I am also taking a look at OpenLisp. At the same
time I am having other problems too and as I wrote twice in
this thread, I shall need time to evaluate all this. So
while nobody could tell me how to solve my Guile problem, I
feel that I have made some headway.
By the way: I knew about LispWorks and my scripting needs
are so minimal that the memory restrictions of the personal
edition would not pose any problems. But I did not know
that LispWorks was embeddable. Is it? You have not aswered
this question yet. I am not looking for /some/ programming
language, I am looking for scripting a C++ application from
Scheme or Lisp. (I have a working prototype in Python
already where scripting is no problem.)
> Well even if I had been that, what does it have to do with
> your problem?
If you were that person, that would mean a lot. That would
mean that I had told to you that I did not wish any
communication with you until you send me your apology and
this would mean that you are intruding. (Then you should
watch your idol more closely as he may be rude and callous
at times, but he is never indecent.)
But, sigh, this belongs to usenet. So thank you for your
valuable ECL hint.
--
J.... B....
Shriram Krishnamurthi
Siegfried Gonzi <siegfri...@kfunigraz.ac.at> writes:
> Let's be serious: he needs his stuff for a school. They are not
> braindamaged that they will spend many thousands of dollars for a
> LispWorks or Franz Allegro distribution; an excecutable requires the
> full licence.
Maybe you don't appreciate how harsh the term "braindamaged" is -- or
maybe you do. But if a bunch of professionals want to put in the time
to actually produce an implementation, why should they not be
compensated for it? I don't know many university professors (which is
what you seem to be) who work for free, or who do research without any
kind of grant support. Calling the fruits of those labors "free" is
utterly false anyway. The companies just make the costs more
explicit.
(There may be one difference, which is that the professor will
probably continue to *think* about the same set of things with or
without funding -- at least until it becomes too cold because there's
no money to buy oil for the wick lamps. At the same time, it's
interesting to see how "Lisp thinking" continues to influence the many
people who work at companies like Sun, some even openly admitting to
it. So even that difference may not count.)
Okay, sorry, off-topic. I couldn't resist.
Shriram
JB
> Maybe you don't appreciate how harsh the term
> "braindamaged" is -- or
> maybe you do. But if a bunch of professionals want to put
> in the time to actually produce an implementation, why
> should they not be compensated for it?
Nobody denies that. But we simply cannot afford it. I think
the prices are made for US universities and then of course
the company cannot say that for German highschools they
make a special price and this is less than one tenth of the
price for US universities.
The funny thing is that as far my problem is concerned, the
product we are talking about, is /free/, but probably of no
use.
--
JB
Rainer Joswig
JB <jbl...@hotmail.com> wrote:
> So paying even $54 is completely out of question. As far as
> I am concerned, I think that volunteering my time is enough
> and do not want to support the community financially.
> Let me tell you an example: When I want to use a coloured
> chalk, I have to use thick and cheap ones as the thin
> pieces of chalk are a bit more expensive and we cannot
> afford it. I also have to pay for the tranparecies I use in
> my classes.
>
> I am sure that the price for an academic license is
> justified, but I do not think that they have sold many
> licenses in Bavaria.
Janos, just a guess. If you develop a Lisp application for
use in schools, Xanalys might be interested in supporting you.
Just a guess, since I can't speak for Xanalys.
This also depends whether LispWorks is a useful
tool for your application. *IFFFFFF* you were interested why not
contact Xanalys in Germany? They have a very helpful sales
person here - Rolf Mach ( rmach @ xanalys.com ).
JB
I reread the ECL manual. I had not overseen it, I did not
understand it. From the explanations there my way to
something like
--- c-code ---
void start_interpreter(char *str)
{
lisp_eval_buffer(str);
// catch errors from the interpreter
}
where lisp_eval_buffer will call C++-functions (this seems
to be well documented: the tak example),
is very, very long.
By reading the manual, I had the impression that though ECL
was designed to be embeddable (when it was called EcoLisp)
this idea was given up and it is no more embeddable or at
least the author does not care.
JB
> In article <aq5pi8$6qi8u$1...@ID-167393.news.dfncis.de>,
> JB <jbl...@hotmail.com> wrote:
>
>> So paying even $54 is completely out of question. As far
>> as I am concerned, I think that volunteering my time is
>> enough and do not want to support the community
>> financially. Let me tell you an example: When I want to
>> use a coloured chalk, I have to use thick and cheap ones
>> as the thin pieces of chalk are a bit more expensive and
>> we cannot afford it. I also have to pay for the
>> tranparecies I use in my classes.
>>
>> I am sure that the price for an academic license is
>> justified, but I do not think that they have sold many
>> licenses in Bavaria.
>
> Janos, just a guess. If you develop a Lisp application for
> use in schools, Xanalys might be interested in supporting
> you.
BUT THE PRODUCT IS FREE!!! (As far as I am concerned.)
> Just a guess, since I can't speak for Xanalys.
> This also depends whether LispWorks is a useful
> tool for your application. *IFFFFFF* you were interested
> why not contact Xanalys in Germany? They have a very
> helpful sales person here - Rolf Mach ( rmach @
> xanalys.com ).
I quote from another posting by me in this thread:
"Nobody denies that. (I mean that the price is ok.) But we
simply cannot afford it. I think
the prices are made for US universities and then of course
the company cannot say that for German highschools they
make a special price and this is less than one tenth of the
price for US universities.
The funny thing is that as far my problem is concerned, the
product we are talking about, is free, but probably of no
use."
But let me ask you directly.
I'd like to build an application consisting of an esoteric
C++ library, wxWindows (which does not work on Cygwin yet,
though it seems that the wxWindows guys have found the
problem at least) and a scripting language that should be
Lisp or Scheme. So I shall have to make those four
ingredients work together.
I know that there are another type of solution: using SWIG.
What would you do?
--
JB
David Rush
> In my experience, it is not unusual for a school, corporation, or
> other institution to balk at paying a few thousand dollars to buy
> the right tool for the job, while paying its employees tens of
> thousands, or hundreds of thousands, or sometimes even millions
> of dollars to work around the problems that result from using the
> wrong tool. Why this "brain-damage"? The tool budget may be too
> small, or there may be some penny-wise/pound-foolish bean counter
> whose approval is needed for each purchase, or some suit may be
> prejudiced against the right tool, or whatever.
Ah Will, this requires a whole book to answer properly...
I had a revelatory experience a few years back while I was developing
my cabinetry skills. I was developing a bit of skill at cutting
mortise-and-tenon joints, and having a lot of fun at designing using
them, but I was still frustated by what I thought was my beginner's
inability to really cut a through mortise with clean edges (important
for strong glue joints) that matched the initial mark-up on the wood.
Then I bought myself some mortising chisels (after convincing my wife
with an apropriate project). Literally overnight, the quality of my
mortise and tenon joints improved at least threefold. I could chop
them quicker (easily twice as fast) and more accurately (which is much
harder to quantify).
Prior to this experience, I had believed the old adage "It's a poor
workman who blames his tools". Now I believe something similiar, but
also different: "It's a poor workman who doesn't know when to blame
his tools". The relevance to your observation, for the chronically
dense (of which there are many on Usenet), is that in the software
business the ones that pay for the tools *don't* use them, so they
have no actual knowledge of the possibility of one tool being better
than another. And in much the same way that my wife had to be
convinced that spending the U$70 or so on chisels was a good idea, so
do the bean counters need to be convinced.
Except, of course, that they're not really listening (and the thousand
other reasons why corporate software development largely sucks). And
now you know why I program recreationally and why I use Scheme and
other advanced tools for doing it. It's a *lot* more fun working with
good tools.
david rush
--
Coffee should be black as hell, strong as death and sweet as love.
-- Turkish proverb
Robert Uhl <ruhl@4dv.net>
>
> I'd like to build an application consisting of an esoteric C++
> library, wxWindows (which does not work on Cygwin yet, though it
> seems that the wxWindows guys have found the problem at least) and a
> scripting language that should be Lisp or Scheme. So I shall have
> to make those four ingredients work together.
>
> I know that there are another type of solution: using SWIG.
>
> What would you do?
I had a similar situation with travlib and travtrack. I wanted to use
SWIG, but at the time it didn't seem to do quite what I wished it to.
Now that I've had to implement Guile bindings by hand, I wish that I'd
instead modified it to do what I wanted it to do, and then used it.
OTOH, it has been nice to have full control over how C functions are
wrapped, particularly given the data structures I am using in C and
Scheme.
If what you're doing is fairly straightforward, I'd go with SWIG, but
if it's complex I'd prob. still do it by hand. But then, it's quite
possible that SWIG already has hooks for that which I missed (or which
weren't there then).
I'd definitely use Guile as the implementation. I've been _very_
happy with it. I can load my libraries in from a Guile session (or
from any Guile code), and I can use Guile from within my travtrack
application. It's very flexible, and very nice.
--
Robert Uhl <ru...@4dv.net>
I went to my first computer conference at the New York Hilton about 20
years ago. When somebody there predicted the market for
microprocessors would eventually be in the millions, someone else
said, `Where are they all going to go? It's not like you need a
computer in every doorknob!'
Years later, I went back to the same hotel. I noticed the room
keys had been replaced by electronic cards you slide into slots in the
doors.
There was a computer in every doorknob.
--Danny Hillis
felix
>
> Here is what I should like to do:
> I should write a c++ program, that uses wxWindows (which is
> a GUI toolkit), Core (which is an esoteric MP arithmetic
> library) and then the scripting language Scheme.
> I am having tecnical problems with Guile and I should prefer
> a Scheme that offers macros anyway. (I do not know if Guile
> has macro support.) What I should be be able to
> do, are calls like
>
> (define g (line '(3 4) '(4 5)))
> (define h (line '(-1 2) '(3 0)))
>
> (define p (section g h))
> (draw g h p)
>
> Then the Scheme functions 'line' and 'section' and 'draw'
> would call C++ functions that would do the job.
>
You might want to try the Chicken Scheme system
(http://www.call-with-current-continuation.org), which
offers a couple of features you seem to be looking for:
- Works on Windows (MSVC or Cygwin)
- Generates *compiled* standalone executables
- Allows *compiled* code to be embedded into C or C++
code
- The generated code can be compiled as C++, but FFI
calls can be a bit tricky
- Can be easily extended
- Macros (`define-macro' or R5RS)
cheers,
felix
Shriram Krishnamurthi
> Now let us proceed. I started with Guile as Guile has the
> reputation of being *the* scripting language.
That's b.s. There are lots and lots of fine Scheme implementations
out there ... and there's Guile, too. If you are going to invest time
in a significant project (which it looks like you are), please spend a
little time surveying the field, rather than falling for hype and
vacuous advertisement like this.
Good luck with your effort!
Shriram
Shriram Krishnamurthi
> Shriram Krishnamurthi wrote:
> > But if a bunch of professionals want to put
> > in the time to actually produce an implementation, why
> > should they not be compensated for it?
>
> Nobody denies that. But we simply cannot afford it.
Please note that I was criticizing Herr Gonzi, not you. I'm very
sympathetic to your problems. All software I've produced I've
distributed free of charge. We even distribute our book on-line free
of charge, and it's used by people at poorer schools not only in
"third world" countries but even in the US.
Indeed, don't assume these prices are "made for US universities" in a
single broad brush -- many US universities also balk at such prices.
Of course, these same universities that give academic departments a
few hundreds of dollars a year for software and books spend millions
on critical infrastructure like seating at the football stadium
... which takes us back, sort of, to Will Clinger's point.
Shriram
Shriram Krishnamurthi
> I'd like to build an application consisting of an esoteric
> C++ library, wxWindows (which does not work on Cygwin yet,
> though it seems that the wxWindows guys have found the
> problem at least) and a scripting language that should be
> Lisp or Scheme.
Given that I have PLT Scheme (MzScheme), which integrates wxWindows,
running on my Windows box, I really don't understand what the problem
is here. You would already have the wxWindows and Lisp/Scheme part
locked down tight. And since PLT Scheme supports linking with C++
applications, including embedding, you have that part down also.
I don't think I'm saying anything different from what a few other
posters have said re. using PLT Scheme. I'm sorry if I've missed a
post in which you explained the problem, but perhaps you could repeat
your explanation?
Shriram
JB
> I'd definitely use Guile as the implementation. I've been
> _very_ happy with it. I can load my libraries in from a
> Guile session (or from any Guile code), and I can use
> Guile from within my travtrack application. It's very
> flexible, and very nice.
The problem is that I cannot compile the tutorial file as an
.cpp file. I can only compile it as a .c file.
--
JB
Siegfried Gonzi
> Maybe you don't appreciate how harsh the term "braindamaged" is -- or
> maybe you do. But if a bunch of professionals want to put in the time
> to actually produce an implementation, why should they not be
> compensated for it? I don't know many university professors (which is
> what you seem to be) who work for free, or who do research without any
> kind of grant support.
[Some small ceaveat: I am not a professor. I was just accepted for a PhD
and actually I am pursuing my PhD degree in physics.]
I think it is not valuable to step deeper into the discussion, because I
am the last person on earth who denies that people in private companies
have to feed their families.
[I bought my new telescope here in Austria, though, the prices in
Germany would have been much lower due to smaller taxes and for reason
of its own. But I thought it is better to support my local dealer. And
you know my PhD salary?]
As I wrote it: all depends on the type of school. We have engineering
shools in Austria which are willing to buy software licenses, because
the schools are specialized in software engineering (as there are shools
which are specialized in chemistry, mechanical engineering, business,...).
But the German schools are high schools in a classical sense.
When I talk of schools I mean schools and not universities. Universities
have a big budget for software licenses.
One shouldn't dismiss the fact the especially in German speaking
countries the money for the schools and universities stems excusively
from the govermnent. There is no such a tradition as with the Americas
where private organizations sponsor schools and universities. Though, I
have to concur that such a system does not impede a school teacher to
contact the next software dealer and negotiate a special license
contract for his school.
S. Gonzi
Marius Vollmer
> The problem is that I cannot compile the tutorial file as an
> .cpp file. I can only compile it as a .c file.
Please provide more details, on <bug-...@gnu.org>. I'm sure people
will help you there when they can reprodice your problem. It is
difficult to help you without knowing more.
Your test program might not compile with a C++ compiler because it
might, well, not be written in C++. Or there might be a deep problem
with Guile's stack tricks that violate some assuptions that C++
exceptions make. We can't say without seeing the program and the
output of the compiler.
To encourage you a bit: I'm using Guile in a C++ program without any
tricks and it is working fine.
--
GPG: D5D4E405 - 2F9B BCCC 8527 692A 04E3 331E FAF8 226A D5D4 E405
MJ Ray
> Thx. I use Linux. But I teach and at school we use Windows. There is
> absolutely no chance of changing this. [...]
If you never ask them to offer GNU/Linux, it will never change.
To a related post: please do not use SuSE. It contains proprietary software
and we do not wish to replace one vendor lock-in with another.
> Besides, using Linux would not help, the problems were the same.
Huh?
MJ Ray
> I could not make Guile work from a .cpp source file, I had to use a .c file.
> Now I am going to use wxWindows in my application and so I should need c++.
> Is there a way of solving this problem?
http://www.swig.org/ would be my first try.
> Is there a manual for Guile in some easily readable form (html,pdf,...)?
http://www.gnu.org/software/guile/docs/docs.html or "info guile-ref" may
work for you.
MJR
JB
> Janos Blazi <jbl...@hotmail.com> wrote:
>> Thx. I use Linux. But I teach and at school we use
>> Windows. There is absolutely no chance of changing this.
>> [...]
>
> If you never ask them to offer GNU/Linux, it will never
> change.
>
> To a related post: please do not use SuSE. It contains
> proprietary software and we do not wish to replace one
> vendor lock-in with another.
I do not understand this. I use SuSE as I was told to by a
genuine Linux expert. This guy once was member of the
Debian team and still uses Debian but told me that for a
newbie SuSE was the best solution.
Which distribution should I use instead?
--
JB
JB
This is interesting. I posted my question to the main guile
maling list but received no response.
--
JB
Jens Axel Søgaard
> Jens Axel Søgaard wrote:
>>Do you know the exact reals library by Boehm et al:
>
> http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/CRCalc-force.html
>
>>It is written in Java and I imagine it would be quite easy
>>to call it from SISC (or Bigloo!).
>
> I vaguely remember that have seen this before. You cannot
> compare to CRs, can you?
Huh? Sure use one of the three compareTo-functions.
http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/com/sgi/math/CR.html
In case anyone is interested, the code Boehm has written is
very readable. Impressively so.
JB
> JB wrote:
>> Jens Axel Søgaard wrote:
>
>>>Do you know the exact reals library by Boehm et al:
>>
>>
http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/CRCalc-force.html
>>
>>>It is written in Java and I imagine it would be quite
>>>easy to call it from SISC (or Bigloo!).
>>
>> I vaguely remember that have seen this before. You cannot
>> compare to CRs, can you?
>
> Huh? Sure use one of the three compareTo-functions.
Sorry, but this does not work. Take a closer look. It is the
fundamental problem of the computability theory of reals
that they cannot be tested for equality. This is
theoretically not possible and it is not possible in our
software either. Actually in our case comparing two equal
reals is a process that does not terminate. (Take a closer
look at those CompareTo functions!)
The solution is that you implement only the constructible
algebraic numbers, where testing for equality is possible.
--
JB
Jens Axel Søgaard
> Jens Axel Søgaard wrote:
>>JB wrote:
>>>Jens Axel Søgaard wrote:
>>
>>>>Do you know the exact reals library by Boehm et al:
>
> http://www.hpl.hp.com/personal/Hans_Boehm/crcalc/CRCalc-force.html
>
>>>>It is written in Java and I imagine it would be quite
>>>>easy to call it from SISC (or Bigloo!).
>>>
>>>I vaguely remember that have seen this before. You cannot
>>>compare to CRs, can you?
>>Huh? Sure use one of the three compareTo-functions.
> Sorry, but this does not work. Take a closer look. It is the
> fundamental problem of the computability theory of reals
> that they cannot be tested for equality.
Sure. But you can stil compare them and determine whether their
difference is less than some given precision. "Compare" is note
same as "determine equality".
But frankly I don't see the relevance of this point in conncetion with
CR - you have the same problem in CORE.
JB
I hope O do not have this problem with Core.
--
JB
Robert Uhl <ruhl@4dv.net>
>
> > Now let us proceed. I started with Guile as Guile has the
> > reputation of being *the* scripting language.
>
> That's b.s. There are lots and lots of fine Scheme implementations
> out there ... and there's Guile, too.
Guile has the advantage of being part of the GNU project, which is
important to many of us, and completely unimportant to others. My
biggest problem with Guile is that it is still very much a
work-in-progress, and thus later versions are typically _much_
improved on earlier ones, and sometimes fail to be
backwards-compatible. Combined with the fact that my distro comes
with an older version, this means that I must maintain parallel
installs, which is a chore.
The only other real issue is that Guile hasn't taken off as much as
one might hope. If it were part of every (reasonably extensible) GNU
app, imagine how cool that'd be. We'll be there someday, though, and
that's why I use Guile.
It is lacking in some areas as an implementation (the lack of a really
nice numerical tower is a bother at times--OTOH I like not having to
include a decimal when using it as a calculator), but is IMHO fairly
decent and more than usable.
--
Robert Uhl <ru...@4dv.net>
Every man, woman, and responsible child has a natural, fundamental,
and inalienable human, individual, civil, and Constitutional right
(within the limits of the Non-Aggression Principle) to obtain, own,
and carry, openly or concealed, any weapon--handgun, shotgun, rifle,
machinegun, anything--any time, anywhere, without asking anyone's
permission. --L. Neil Smith
Robert Uhl <ruhl@4dv.net>
>
> > Is there a manual for Guile in some easily readable form
> > (html,pdf,...)?
>
> http://www.gnu.org/software/guile/docs/docs.html or "info guile-ref"
> may work for you.
It'd be nice if the docs section had a link to the reference manual...
--
Robert Uhl <ru...@4dv.net>
It's not like I burst forth from my father's brow holding a fistful
of cash and 10 years of UNIX sysadmin experience. --Mike Sphar
Anton van Straaten
> important to many of us, and completely unimportant to others.
The importance of that is unfortunate, though. Surely it would be better
for the GNU project to develop standards to allow different Schemes to be
used? Why bless one particular Scheme over all others? Of course, the onus
would be on implementors to meet the requirements for interfacing as an
extension language - Microsoft's Active Scripting is/was one example of this
kind of thing, and the Java VM provides another example of standards being
used to avoid forcing use of one particular implementation.
In this area, the GNU project seems perhaps unintentionally to be following
a classic "lock-in" strategy. It's not clear to me that being locked in to
a particular free implementation of something is that much better than being
locked into a commercial product. Lock-in is a removal of freedom, quite
closely related to the kind of freedom which the GPL is designed to protect.
> The only other real issue is that Guile hasn't taken off as much as
> one might hope.
...
> It is lacking in some areas as an implementation (the lack of a really
> nice numerical tower is a bother at times
I rest my case. There are other GPL'd and LGPL'd Schemes out there, that
are more full-featured, more widely used, better performing, etc...
Anton
David Rush
ru...@4dv.net (Robert Uhl <ru...@4dv.net>) writes:
> Shriram Krishnamurthi <s...@cs.brown.edu> writes:
> >
> > > Now let us proceed. I started with Guile as Guile has the
> > > reputation of being *the* scripting language.
> >
> > That's b.s. There are lots and lots of fine Scheme implementations
> > out there ... and there's Guile, too.
>
> Guile has the advantage of being part of the GNU project,
O yay. Welcome to the 21st century. I appreciate GNU for helping (in a
big way) kick-start the free software movement, but I haven't seen too
much good coming out of there lately...
> biggest problem with Guile is that it is still very much a
> work-in-progress, and thus later versions are typically _much_
> improved on earlier ones,
Bog knows they couldn't be worse. Slow. Bugs in hygienic macros and
call/cc. The world's most *bizarre* module system...I got very
thouroughly turned off guile between my discovery of it (and Scheme)
via SCWM in '97-98ish and my abandonment of it for development
*except* as part of SCWM in 2000. And now SCWM appears to be
dead...hmmm.
> The only other real issue is that Guile hasn't taken off as much as
> one might hope.
There are reasons for that.
> If it were part of every (reasonably extensible) GNU
> app, imagine how cool that'd be.
I shudder at the thought. Why not just embed ELisp in everything?
IOB:
> Every man, woman, and responsible child has a natural, fundamental,
> and inalienable human, individual, civil, and Constitutional right
> (to keep and bear arms)
Actually I have heard some very cogent arguments that it is not merely
a right, but a responsibility ;)
david rush
--
From the start...the flute has been associated with pure (some might
say impure) energy. Its sound releases something naturally untamed, as
if a squirrel were let loose in a church." --Seamus Heaney
Robert Uhl <ruhl@4dv.net>
> > Guile has the advantage of being part of the GNU project, which is
> > important to many of us, and completely unimportant to others.
>
> The importance of that is unfortunate, though. Surely it would be
> better for the GNU project to develop standards to allow different
> Schemes to be used? Why bless one particular Scheme over all
> others?
Well, at one time there was the gh interface, which I think was meant
to serve as a way to interface with multiple Scheme implementations.
It may not have started with Guile, though.
Guile isn't much more than a C API and a Scheme library. If one
wanted, one could re-implement that API and library, but why bother?
--
Robert Uhl <ru...@4dv.net>
`A person commits a class 4 felony when he knowingly uses or intends to
use armed force to prevent assault, rape, murder, or other attack upon
himself or another.'
--Illinois Revised Statutes, Chapter 38, Paragraph 24
Anton van Straaten
> wanted, one could re-implement that API and library, but why bother?
Well, I can think of many reasons to bother, both from the point of view of
the GNU project itself, and of the GNU user community. It's not so much
about expecting other people to reimplement those things as a kind of
unsanctioned plugin, but rather about creating an environment (technical and
otherwise) which encourages them to implement what's necessary to integrate
cleanly.
To remain focused on the original point, you said that "Guile has the
advantage of being part of the GNU project, which is important to many of
us". The question is, why is it important? As opposed to using some other
GPL'd Scheme? The answer, I assume, relates to its ability to integrate
with other GNU programs - the "design context" which Tom Lord mentioned the
other day. I'm suggesting that given the broader goals of the GNU project,
that the approach taken with Guile seems somewhat counter to those goals,
and at best will just result in Guile becoming another Emacs Lisp -
something you have to use because it's tied into other programs, not
something you choose to use because it's most suitable for your purposes.
In fact, Guile seems to have already become that, based on your comment in
quotes above.
Anton
Neil W. Van Dyke
> O yay. Welcome to the 21st century. I appreciate GNU for helping (in a
> big way) kick-start the free software movement, but I haven't seen too
> much good coming out of there lately...
I think there are a variety of reasons for that, some of them the FSF's
fault, and some not. One of the problems was that dotcom mania and then
"Linux" commercial viability did both: (1) enticed a lot of potential
GNU hackers out of their basements, dorm rooms, and sys-admin offices
with promises of instant millions, if only they'd go dotcom instead of
FSF; and (2) introduced so many "Linux" opportunists into the space that
people like RMS couldn't tell who was a genuine crunchy-granola Free
Software hacker, and who mainly saw dollar signs and empire-building.
But as for Scheme... Guile 1.6 has *finally* been released, and is
quite an improvement over 1.4. It'd be nice if this release gave Guile
a second wind and maybe interested a couple new key contributors who can
work with the existing ones. There's an obvious challenge for Guile to
show what it can do better than other embedables (than, say, MzScheme).
--
Neil W. Van Dyke
http://www.neilvandyke.org/
Shriram Krishnamurthi
> But as for Scheme... Guile 1.6 has *finally* been released, and is
> quite an improvement over 1.4. It'd be nice if this release gave Guile
> a second wind and maybe interested a couple new key contributors who can
> work with the existing ones. There's an obvious challenge for Guile to
> show what it can do better than other embedables (than, say, MzScheme).
Good for the 1.6 release. Since you seem to care about Guile, maybe
you can go ahead and answer the question you've raised: how *does* it
compare, and what does it do better?
By the way, a proper comparison would be not merely to embeddables
(after all, you have to *develop* the embedded program, and sometimes
it grows large enough that you can't write it all on one line).
Where's the programming environment? Where's the algebraic stepper?
Where's the equivalent of MrFlow, Check Syntax, Help Desk, Countours,
images as objects, language levels, language embedding (Algol 60)?
How about the equivalent of units, mixins, parameterizations, modules,
language definitions? send/suspend and send/finish?
(Or, you could pick something like Larceny or Bigloo or Gambit or Chez
Scheme or Stalin, and come up with a bunch of other questions.)
*That*'s the obvious challenge.
Shriram
Friedrich Dominicus
>
> By the way, a proper comparison would be not merely to embeddables
> (after all, you have to *develop* the embedded program, and sometimes
> it grows large enough that you can't write it all on one line).
> Where's the programming environment? Where's the algebraic stepper?
> Where's the equivalent of MrFlow, Check Syntax, Help Desk, Countours,
> images as objects, language levels, language embedding (Algol 60)?
Well I could not put it better. I have to admit I can see nothing in
Guile which would let me choose it over DrScheme or even MIT
Scheme. The only thing I can imagine would let me choose guile is if
there would be the "Guile-Emacs" and even there I would prefer a
DrEmacs ;-)
Regards
Friedrich
Rob Warnock
+---------------
| Shriram Krishnamurthi <s...@cs.brown.edu> writes:
| > Where's the programming environment? Where's the algebraic stepper?
| > Where's the equivalent of MrFlow, Check Syntax, Help Desk, Countours,
| > images as objects, language levels, language embedding (Algol 60)?
|
| Well I could not put it better. I have to admit I can see nothing in
| Guile which would let me choose it over DrScheme or even MIT Scheme.
| The only thing I can imagine would let me choose guile is if there
| would be the "Guile-Emacs" and even there I would prefer a DrEmacs ;-)
Already there, sort of: It's called MrEd, and is part of PLT Scheme.
It provides the editor for DrScheme, but can also be used as a standalone
multi-media file editor -- including a simple embedded web browser! --
and as a framework for building standalone GUI-based apps. (DrScheme
itself is built on top of MrEd.)
-Rob
-----
Rob Warnock, PP-ASEL-IA <rp...@rpw3.org>
627 26th Avenue <URL:http://www.rpw3.org/>
San Mateo, CA 94403 (650)572-2607
Neil W. Van Dyke
> Good for the 1.6 release. Since you seem to care about Guile, maybe
> you can go ahead and answer the question you've raised: how *does* it
> compare, and what does it do better?
I'm the wrong person to ask, since I switched to MzScheme. :) Anyway,
that's the challenge Guile needs to address: "why aren't we just using
MzScheme instead?"
Maybe they decide they want to squeeze a full-featured Scheme with a
CLOS-alike into 1MB, or push more of the footprint into shared text so
they can have *numerous* interpreters running, or have funky transparent
IPC and persistence, or a dynamic incremental optimizer, or something
else that might require serious internals mucking.
I'd welcome this kind of mutation in the ecology of Scheme implementations.
Friedrich Dominicus
> Friedrich Dominicus <fr...@q-software-solutions.com> wrote:
> +---------------
> | Shriram Krishnamurthi <s...@cs.brown.edu> writes:
> | > Where's the programming environment? Where's the algebraic stepper?
> | > Where's the equivalent of MrFlow, Check Syntax, Help Desk, Countours,
> | > images as objects, language levels, language embedding (Algol 60)?
> |
> | Well I could not put it better. I have to admit I can see nothing in
> | Guile which would let me choose it over DrScheme or even MIT Scheme.
> | The only thing I can imagine would let me choose guile is if there
> | would be the "Guile-Emacs" and even there I would prefer a DrEmacs ;-)
> +---------------
>
> Already there, sort of:
ho ho, slow down a bit I'm talking about Emacs not any
"would-like-to-be-a-grown-up-emacs" ;-)
I use Emacs for everything among others C-Development, Mail and News
and and and. That all is not available in anything else. It's the
sheer mass of useful and useless tools (till you need them) in Emacs
which makes it so outstanding.
So please do not get me wrong I do not want anyone starting on that
stuff just for my happiness. And as said the only reason that I would
look at Guile is if Guile-Emacs would be a bit more than varpoware.
Regards
Friedrich
David Rush
> rp...@rpw3.org (Rob Warnock) writes:
> > Friedrich Dominicus <fr...@q-software-solutions.com> wrote:
> > | Guile which would let me choose it over DrScheme or even MIT Scheme.
> > | The only thing I can imagine would let me choose guile is if there
> > | would be the "Guile-Emacs" and even there I would prefer a DrEmacs ;-)
> >
> ho ho, slow down a bit I'm talking about Emacs not any
> "would-like-to-be-a-grown-up-emacs" ;-)
>
> I use Emacs for everything among others C-Development, Mail and News
> and and and. That all is not available in anything else. It's the
> sheer mass of useful and useless tools (till you need them) in Emacs
> which makes it so outstanding.
M-x tetris
Because you can.
david rush
--
Don't you think it would be a useful item to add to your intellectual
toolkit to be capable of saying, when a ton of wet steaming bullshit
lands on your head, "My goodness, this appears to be bullshit?"
-- Douglas MacArthur Shaftoe, in _Cryptonomicon_
Shriram Krishnamurthi
> Maybe they decide they want to squeeze a full-featured Scheme with a
> CLOS-alike into 1MB, or push more of the footprint into shared text so
> they can have *numerous* interpreters running, or have funky transparent
> IPC and persistence, or a dynamic incremental optimizer, or something
> else that might require serious internals mucking.
>
> I'd welcome this kind of mutation in the ecology of Scheme implementations.
I do, too. I think it's terrific that Scheme has not only PLT Scheme
but also Bigloo, Gambit, Larceny, Chez, Stalin and Scheme 48. Each of
these meets a different, valuable "price point" and -- this is the
kicker -- ***does it really well***.
That still doesn't help me from feeling Guile is at the shallow end of
this gene pool.
Shriram
Hans-J. Boehm
can support a decision procedure for equality. The Java package I
wrote is no doubt much simpler, and can handle arbitrary constructive
reals, at least at the library level. (It explicitly includes
exponential and trig functions.) Hence equality is undecidable, and
inherently so.
Hans
JB
Exactly. Prof Weihrauch, who is one of the leading
authorities in the field of Computability Theory of the
real numbers, was my teacher and I think I understood these
things well.
But this means in no way that I should not appreciate your
effort, but it is not exactly what I was looking for. (By
the way, I think that Core only implements algebraic
numbers that are constructible by ruler and compass but
this is exactly what I am looking for.
--
JB
Bengt Kleberg
...deleted
>That still doesn't help me from feeling Guile is at the shallow end of
>this gene pool.
perhaps guile wants to be a full r[45]rs with a gnu public license for
everything? ie, there might be no technical reasons behind guiles
existence, but there might be ideological.
bengt
MJ Ray
> I do not understand this. I use SuSE as I was told to by a genuine Linux
> expert. This guy once was member of the Debian team and still uses Debian
> but told me that for a newbie SuSE was the best solution. Which
> distribution should I use instead?
Currently, I recommend Mandrake or special-purpose Debian-derived desktop
distribtions. Mandrake is far bigger and has a far slicker installation and
setup, but doesn't have dpkg and apt for smooth upgrades, so you'll need to
reinstall at some point, most likely.
The Debian project is a broad church. All Debian developers should respect
the project's free software guidelines and the Social Contract when working
for the project, but it does not necessarily mean that all their advice
will. Buyer beware, same as ever.
MJR
MJ Ray
> It'd be nice if the docs section had a link to the reference manual...
Indeed. Now that 1.6 is out, I'm not sure why it's not there yet.
MJ Ray
JB <jbl...@hotmail.com> wrote:
> I do not understand this. I use SuSE as I was told to by a genuine Linux
> expert. This guy once was member of the Debian team and still uses Debian
> but told me that for a newbie SuSE was the best solution. Which
> distribution should I use instead?
Currently, I recommend Mandrake or special-purpose Debian-derived desktop
Jens Axel Søgaard
But DrScheme, Bigloo, SISC and numerous others are also GPL'ed.
Actually some where turned down, when the decision was made
which implementation to choose to be the official FSF (GNU?)Scheme.
Google for the old rrrs maling lists.
Robert Uhl <ruhl@4dv.net>
>
> > Guile has the advantage of being part of the GNU project,
>
> O yay. Welcome to the 21st century. I appreciate GNU for helping
> (in a big way) kick-start the free software movement, but I haven't
> seen too much good coming out of there lately...
But the GNU Project is still the repository of what are IMHO the
official free software tools--or should be. But I can see how
opinions differ. It's important to me, anyway.
> > biggest problem with Guile is that it is still very much a
> > work-in-progress, and thus later versions are typically _much_
> > improved on earlier ones,
>
> Bog knows they couldn't be worse. Slow. Bugs in hygienic macros
> and call/cc. The world's most *bizarre* module system...
I've not noticed it being particularly slow. I was actually
pleasantly surprised by its speed at executing some _extraordinarily_
bad and unoptimised code I'd written one even.
I've not run across any of the mentioned bugs. The worst one I've
found so far is the stack depth limitation, which is IMHO
unconscionable.
I rather like the module system, but then I've not dealt with much
else.
> > If it were part of every (reasonably extensible) GNU app, imagine
> > how cool that'd be.
>
> I shudder at the thought. Why not just embed ELisp in everything?
Because Scheme is better than elisp...
What's wrong with having everything extensible? If the overhead is
kept to a minium (unloaded until needed, perhaps) it can only help.
After all, emacs is not the greatest text editor in the world for
nothing!
> > Every man, woman, and responsible child has a natural,
> > fundamental, and inalienable human, individual, civil, and
> > Constitutional right (to keep and bear arms)
>
> Actually I have heard some very cogent arguments that it is not
> merely a right, but a responsibility ;)
I'd almost agree with that. Those who wish to render themselves
helpless should be allowed to, though. Perhaps we could institute a
permit system, whereupon they would be required to pass a course
informing them of the risks associated therewith...
--
Robert Uhl <ru...@4dv.net>
Hell No! We won't relent to your artificial socioeconomic subsidisation
of political contributors and special interests!
Shriram Krishnamurthi
> > O yay. Welcome to the 21st century. I appreciate GNU for helping
> > (in a big way) kick-start the free software movement, but I haven't
> > seen too much good coming out of there lately...
>
> But the GNU Project is still the repository of what are IMHO the
> official free software tools--or should be.
No, you are confusing "free software" with "Free Software". If GNU
wants to set itself up as the "official repository" of "Free
Software", fine -- Stallman coined the term, he can protect it with
diving tackles (though, boy, that sure starts to look like the
behavior of the people he despises ...).
But "free software" is not bound by the diktats of Stallman, Lord and
others. Are you suggesting nobody else can claim to produce free
software, or did you merely fail to press the Shift key?
Shriram
MJ Ray
> But "free software" is not bound by the diktats of Stallman, Lord and
> others. Are you suggesting nobody else can claim to produce free
> software, or did you merely fail to press the Shift key?
For all his failings, rms defined the term as we know it, so he's got a bit
of a platform from which to say how he'd like to see people use it. He
makes points and tries to explain why he thinks that, then hopes that others
will agree with his reasoning. Most of the time, he's on the ball, but he's
human and not infallible.
Using the loaded term "diktat" makes it sound like Stallman, Lord, etc have
some other method than debate, assistance and persuasive argument to impose
their will. It is not so, is it? If anything, PLT have more to worry about
from any diktats of their host universities...
MJR
David Rush
MJ Ray <markj...@cloaked.freeserve.co.uk> writes:
> Shriram Krishnamurthi <s...@cs.brown.edu> wrote:
> > But "free software" is not bound by the diktats of Stallman, Lord and
> > others. Are you suggesting nobody else can claim to produce free
> > software, or did you merely fail to press the Shift key?
>
> For all his failings, rms defined the term as we know it
Err...mostly. I would say that RMS created an opportunity for
meaningful discussion about the economics, politics, and morality of
software devlopment by his definition of "free software". The notion
of free software has at least two different senses *now* (Free Beer vs
Free Speech).
> so he's got a bit of a platform from which to say how he'd like to
> see people use it.
And so did Humpty-Dumpty. The question is: Who is to be master?
> Using the loaded term "diktat" makes it sound like Stallman, Lord,
> etc have some other method than debate, assistance and persuasive
> argument to impose their will. It is not so, is it?
It depends on whether you think that the moralistic tone is
justifiable in the socio-eco-political context of the
debate. Arguments from authority ("I invented X, so I must be right")
have never gone down well with me, and I think many others feel the
same way.
david rush
--
Oh. Politeness. Good-bye.
-- Elson Roa (in _Against a Dark Background_)
Shriram Krishnamurthi
> For all his failings, rms defined the term as we know it, so he's got a bit
> of a platform from which to say how he'd like to see people use it.
(a) And then you go on to ask me about the term "diktat"?
(b) I made a distinction between "Free Software" and "free software"
which you quoted but then ignored.
> Using the loaded term "diktat" makes it sound like Stallman, Lord, etc have
> some other method than debate, assistance and persuasive argument to impose
> their will.
If you call treating Usenet as a write-only medium, or the initial
development of Guile, "debate, assistance and persuasive argument" ...
Were you here in the glorious mid-90s?
> If anything, PLT have more to worry about
> from any diktats of their host universities...
(a) There is not one but many host universities. No one school can
lay claim to the product.
(b) We have contributors from outside universities.
(c) The problem doesn't appear different from the development done for
Guile by people employed at universities and companies. (Are all
Guile developers unemployed?)
(d) We are moving to clarify our independence. We should have an
announcement about it in the coming months.
Shriram
MJ Ray
> This discussion is going political...
So it is. Does that worry you?
[...]
> And so did Humpty-Dumpty. The question is: Who is to be master?
No-one should be master. We should all have open minds and consider things
as they present themselves, seeing if they challenge or reinforce our ideas.
> [....] Arguments from authority ("I invented X, so I must be right") have
> never gone down well with me, and I think many others feel the same way.
The argument is sometimes presented via "I invented X, so please listen
awhile," rather than using it to claim supremacy.
MJR