jscheme-user — list for general users

Re: [Jscheme-user] ILC 2003 Programming Contest

[Timothy J. H. <tim...@ma...>]

On Monday, August 25, 2003, at 01:38 PM, Ken Anderson wrote:

> Here's a Lisp programming contest:
> http://www.ravenbrook.com/doc/2003/05/28/contest/
> Though they want a common lisp solution, it might be fun to come up 
> with one in JScheme too.  I'm interested in any suggestions on what a 
> good strategy would be.

   Here is a logic programming inspired solution strategy.... this is 
how N-queens and similar problems are solved
using Constraint Logic Programming. We could translate it into Jscheme 
without too much trouble....

The initial approach would be to use a constrained backtracking method.

For each of the 64 spaces on the board we would keep track of which 
particular puzzle pieces could go there.
This task is simplified if we consider each orientation of the 13 
pieces as a different piece. (So there are now
13*4 = 52 different pieces, and a constraint that exactly one piece 
from each group of 4 must be used).

Similarly, we can number the cells on each (oriented) puzzle piece, so 
that we would specify that
in a particular frame position (say row=5,col=3)
   cell 3 of piece "J" with orientation "N" could be in that position.
The idea here then is that one first initializes all 64 frame positions 
to see which puzzle pieces they could contain
(loop over all 13 pieces, 4 orientations and 3-6 cells of each piece).
There would be about 13*4*5 = 260 puzzle placements possible
at each of the 64 positions of the initial board. A board state could 
be an 8x8 array of bit strings of size 260 or so.

Then one starts at the top row and col and selects one of the possible 
pieces (e.g. the first).
All other pieces are removed from that position and each removed piece 
is then used to remove
possibilities from neighboring cells. Moreover, each removal is cached 
in a list so that it can be restored on backtracking.
(On backtracking, the next choice will be made, etc until all choices 
have been made (at which point deeper backtracking
is called).

One then takes the "next" board position that contain more than one 
elements and non-deterministically chooses one repeating the
same constraint process. The "next" board position can be chosen 
arbitrarily (e.g. the position containing the fewest positions
is a good choice....)
If a frame position becomes empty (i.e. no puzzle piece could cover 
that cell) then you backtrack to
the last choice and make a new choice....

The "tabs" constraint is handled by noticing that each cell of a puzzle 
piece can be labelled with "Black" or "White" and
each position in the Chessboard can be labelled with "Black" or "White" 
in a checkerboard pattern. If this is done, then the
tabs will match iff the even/odd properties match. This is just a 
further restriction on which puzzle positions can go in which
board positions.


Implementing this will require implementing a choice point stack. a 
trail stack (to store the backtracking info),
and deciding on how to represent the current set of possible puzzle 
placements at each position in the board....

we could use an array of hashsets??? or an 8x8 array of bit strings??

Cheers,
---Tim---



> k
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: VM Ware
> With VMware you can run multiple operating systems on a single machine.
> WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines
> at the same time. Free trial click 
> here:http://www.vmware.com/wl/offer/358/0
> _______________________________________________
> Jscheme-user mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-user

[Jscheme-user] ILC 2003 Programming Contest

[Ken A. <kan...@bb...>]

Here's a Lisp programming contest:
http://www.ravenbrook.com/doc/2003/05/28/contest/
Though they want a common lisp solution, it might be fun to come up with one in JScheme too.  I'm interested in any suggestions on what a good strategy would be.

k

[Jscheme-user] Fwd: LWN thread on gnucash, mentions guile

[Ken A. <kan...@bb...>]

Gnucash is a large Linux application that uses Guile scheme.
Perhaps there are some lessons we can learn from this discussion.
One is keep the cost of entry low and API's clearly documented.


>To: gui...@gn...
>From: Greg Troxel <gd...@ir...>
>Date: 12 Aug 2003 07:03:48 -0400
>Lines: 15
>User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3
>Subject: LWN thread on gnucash, mentions guile
>X-BeenThere: gui...@gn...
>X-Mailman-Version: 2.1.2
>List-Id: General Guile related discussions  <guile-user.gnu.org>
>List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/guile-user>,
>        <mailto:gui...@gn...?subject=unsubscribe>
>List-Archive: <http://mail.gnu.org/pipermail/guile-user>
>List-Post: <mailto:gui...@gn...>
>List-Help: <mailto:gui...@gn...?subject=help>
>List-Subscribe: <http://mail.gnu.org/mailman/listinfo/guile-user>,
>        <mailto:gui...@gn...?subject=subscribe>
>Sender: guile-user-bounces+kanderson=bb...@gn...
>Old-X-Spam-Status: NO
>Old-X-Scanned-By: MIMEDefang 2.28 (www . roaringpenguin . com / mimedefang)
>Old-X-Scanned-By: MIMEDefang 2.35
>X-Scanned-By: Spam Assassin
>X-Spam-Status: No, hits=-20.7 required=5.0
>        tests=AWL,BAYES_01,SIGNATURE_LONG_SPARSE,USER_AGENT_GNUS_UA
>        autolearn=ham version=2.52
>X-Spam-Level: 
>X-Spam-Checker-Version: SpamAssassin 2.52 (1.174.2.8-2003-03-24-exp)
>
>  http://lwn.net/Articles/43856/
>
>This has some discussion of whether using guile was a good idea or
>not.
>
>Interestingly, that discussion focuses on whether "Scheme is too
>hard"; I didn't see any "guile is broken" comments.  So perhaps half
>the battle has been won.
>
>(They are looking for help - the thread is in response to a call for
>help from one of the lead gnucash developers.) 
>
>-- 
>
>        Greg Troxel <gd...@ir...>
>
>
>_______________________________________________
>Guile-user mailing list
>Gui...@gn...
>http://mail.gnu.org/mailman/listinfo/guile-user

Re: [Jscheme-user] macro question

[Geoffrey K. <ge...@kn...>]

That's exactly what I was aiming for, thanks!  --Geoffrey

On Friday, Aug 8, 2003, at 09:29 US/Eastern, Ken Anderson wrote:

> You can take this one step further by writing a function producing 
> function:
>
>> (define (picker x)
>     (lambda () (vector-ref x (.nextInt random-gen (vector-length x)))))
> (lambda picker (x)...)
>> (define rocks #(1 2 3 4 5))
> #(1 2 3 4 5)
>> (define rock (picker rocks))
> (lambda picker~0 ()...)
>> (rock)
> 3
>> (rock)
> 2
>
> At 06:35 AM 8/8/2003 -0400, Geoffrey Knauth wrote:
>> Never mind, I guess that was a silly question.  For now I've done 
>> this:
>>
>> (define (pick-one v)
>>  (vector-ref v (.nextInt random-gen (vector-length v))))
>>
>> On Friday, Aug 8, 2003, at 00:00 US/Eastern, Geoffrey Knauth wrote:
>>
>>> Let's say I have:
>>>
>>> (define (stone-size)
>>>  (vector-ref stone-sizes
>>>              (.nextInt random-gen (vector-length stone-sizes))))
>>> (define (stone-thickness)
>>>  (vector-ref stone-thicknesses
>>>              (.nextInt random-gen (vector-length 
>>> stone-thicknesses))))
>>> (define (caulking-type)
>>>  (vector-ref caulking-types
>>>              (.nextInt random-gen (vector-length caulking-types))))
>>>
>>> Notice the repetition?  How can I use a JScheme macro to write this 
>>> once?  Ignore the fact that "es" pluralizes "thickness" while "s" 
>>> pluralizes "size" and "type."
>>>
>>> Thanks,
>>> Geoffrey

Re: [Jscheme-user] macro question

[Ken A. <kan...@bb...>]

You can take this one step further by writing a function producing function:

> (define (picker x)
    (lambda () (vector-ref x (.nextInt random-gen (vector-length x)))))
(lambda picker (x)...)
> (define rocks #(1 2 3 4 5))
#(1 2 3 4 5)
> (define rock (picker rocks))
(lambda picker~0 ()...)
> (rock)
3
> (rock)
2

At 06:35 AM 8/8/2003 -0400, Geoffrey Knauth wrote:
>Never mind, I guess that was a silly question.  For now I've done this:
>
>(define (pick-one v)
>  (vector-ref v (.nextInt random-gen (vector-length v))))
>
>On Friday, Aug 8, 2003, at 00:00 US/Eastern, Geoffrey Knauth wrote:
>
>>Let's say I have:
>>
>>(define (stone-size)
>>  (vector-ref stone-sizes
>>              (.nextInt random-gen (vector-length stone-sizes))))
>>(define (stone-thickness)
>>  (vector-ref stone-thicknesses
>>              (.nextInt random-gen (vector-length stone-thicknesses))))
>>(define (caulking-type)
>>  (vector-ref caulking-types
>>              (.nextInt random-gen (vector-length caulking-types))))
>>
>>Notice the repetition?  How can I use a JScheme macro to write this once?  Ignore the fact that "es" pluralizes "thickness" while "s" pluralizes "size" and "type."
>>
>>Thanks,
>>Geoffrey
>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>Data Reports, E-commerce, Portals, and Forums are available now.
>Download today and enter to win an XBOX or Visual Studio .NET.
>http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
>_______________________________________________
>Jscheme-user mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-user

Re: [Jscheme-user] macro question

[Geoffrey K. <ge...@kn...>]

Never mind, I guess that was a silly question.  For now I've done this:

(define (pick-one v)
   (vector-ref v (.nextInt random-gen (vector-length v))))

On Friday, Aug 8, 2003, at 00:00 US/Eastern, Geoffrey Knauth wrote:

> Let's say I have:
>
> (define (stone-size)
>   (vector-ref stone-sizes
>               (.nextInt random-gen (vector-length stone-sizes))))
> (define (stone-thickness)
>   (vector-ref stone-thicknesses
>               (.nextInt random-gen (vector-length stone-thicknesses))))
> (define (caulking-type)
>   (vector-ref caulking-types
>               (.nextInt random-gen (vector-length caulking-types))))
>
> Notice the repetition?  How can I use a JScheme macro to write this 
> once?  Ignore the fact that "es" pluralizes "thickness" while "s" 
> pluralizes "size" and "type."
>
> Thanks,
> Geoffrey

[Jscheme-user] macro question

[Geoffrey K. <ge...@kn...>]

Let's say I have:

(define (stone-size)
   (vector-ref stone-sizes
               (.nextInt random-gen (vector-length stone-sizes))))
(define (stone-thickness)
   (vector-ref stone-thicknesses
               (.nextInt random-gen (vector-length stone-thicknesses))))
(define (caulking-type)
   (vector-ref caulking-types
               (.nextInt random-gen (vector-length caulking-types))))

Notice the repetition?  How can I use a JScheme macro to write this 
once?  Ignore the fact that "es" pluralizes "thickness" while "s" 
pluralizes "size" and "type."

Thanks,
Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-user] Jscheme at Millennium Pharm.

[Ken A. <kan...@bb...>]

At 11:52 AM 8/5/2003 -0400, Jonathan A Rees wrote:
>  - I use Java collections heavily, and wrote some simple procedures
>    to make doing so easier from jscheme (I wish that lists were
>    Lists! - they could be, you know)

Rusty pointed out that you may not know about elf/iterate.scm which provides map* and for-each* for iterating over collections.

In fact, i'd like to make these map and for-each, though map* currently only works for 1 argument.

There's plenty of opportunity to make things more java compatible, with out loosing scheme compatibility, for example (load) should take a URL and any InputStream or Reader as an argument. 

Re: [Jscheme-user] Jscheme at Millennium Pharm.

[Ken A. <kan...@bb...>]

This is a great example, thanks!

At 11:52 AM 8/5/2003 -0400, Jonathan A Rees wrote:

>Thought it might be time for me to speak up on what I'm doing, just
>FYI... I work at Millennium Pharmaceuticals in the computational
>biology group, and am maintaining a sort of 'knowledge base' - a
>little database of interactions between biological objects.  This is
>used in the analysis of transcription profiling experiments.  We're
>dealing with a big curation problem right now in reconciling
>information from two different sources (internally curated
>vs. Ingenuity-derived).  I'm using Jscheme for all of the analysis at
>this point.  This means reading tab-delimited files, writing reports
>of many different kinds, doing lookups and filtering and conversions
>and set operations and so on.
>
>I also use jscheme for debugging a web services-based application
>(that uses GLUE to do SOAP), and for reading an S-expression-syntax
>configuration file that's then used by Java code.
>
>We got the seal of approval from the MPI legal department to use
>jscheme in applications developed here, FWIW.
>
>Random particulars:
>
>  - I use Java collections heavily, and wrote some simple procedures
>    to make doing so easier from jscheme (I wish that lists were
>    Lists! - they could be, you know)

Well, i've never thought of that!  We have not used Collections so JScheme can be used on the oldest browsers, but maybe its time to move on and take advantage of the newer versions of Java.

>  - I have classes for tab-delimited table input and output, and
>    use these from jscheme; I have a number of other classses and
>    methods exclusively for use from jscheme (e.g. ProcedureComparator
>    for calling Collections.sort with a scheme-supplied comparison
>    function)

ProcedureComparator must be like elf.SchemeComparator.  I wrote some tab delimited code recently too.  The first line is field names and i return each row as a HashTable, and an accessor is automatically written for each field.  You could do something more space efficient if you have a lot of data.

>  - I wish I could interrupt runaway evaluations without losing all 
>    my state (there's probably a way to do this but I don't know it)

see elf/interrupt.scm.  This puts up an interrupt button.

>  - I wish map and for-each would throw exceptions when a list
>    argument isn't a list - current behavior makes bugs harder to
>    track down

Good idea.

>  - I wish the jscheme and java backtraces were interleaved properly
>    (right now I often retry a failing evaluation with 
>    (tryCatch ... .printStackTrace) wrapped around the form)

Do you mean when scheme calls java which calls scheme ...?  can you give a simple example?

>  - I wish compatibility with R^nRS were a bit higher, esp. regarding
>    numbers.  (/ 1 2) => 0 etc. are very frustrating, makes reuse of 
>    existing Scheme code more difficult.

Yes, we chose to move closer to Java, and unfortunately left some Scheme behind.  I wish we could fit better with Scheme, for example we don't support any  SRFI's, probably because we get what we need from Java.

I've been wondering if we could provide extended math as a separate module.
Bigrats - big rational numbers can be useful for things like knowing how good your numerical approximation is.  

Tim, what about one for interval arithmetic?

>Overall I'm much happier using jscheme than Perl or Java, the other
>primary choices around here.  'wc' tells me my jscheme files total
>about 3000 lines.  And the best part is - I didn't have to write it.
>
>Cheers
>Jonathan
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>Data Reports, E-commerce, Portals, and Forums are available now.
>Download today and enter to win an XBOX or Visual Studio .NET.
>http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
>_______________________________________________
>Jscheme-user mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-user

Re: [Jscheme-user] Jscheme at Millennium Pharm.

[Geoffrey K. <ge...@kn...>]

This is music to my ears, not because I ever doubted the creativity and 
succintness JScheme enables, but because you convinced your management 
and legal folks to think outside the box and approve development.  
That's been the major stumbling block in my work.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

On Tuesday, Aug 5, 2003, at 11:52 US/Eastern, Jonathan A Rees wrote:

>
> Thought it might be time for me to speak up on what I'm doing, just
> FYI... I work at Millennium Pharmaceuticals in the computational
> biology group, and am maintaining a sort of 'knowledge base' - a
> little database of interactions between biological objects.  This is
> used in the analysis of transcription profiling experiments.  We're
> dealing with a big curation problem right now in reconciling
> information from two different sources (internally curated
> vs. Ingenuity-derived).  I'm using Jscheme for all of the analysis at
> this point.  This means reading tab-delimited files, writing reports
> of many different kinds, doing lookups and filtering and conversions
> and set operations and so on.
>
> I also use jscheme for debugging a web services-based application
> (that uses GLUE to do SOAP), and for reading an S-expression-syntax
> configuration file that's then used by Java code.
>
> We got the seal of approval from the MPI legal department to use
> jscheme in applications developed here, FWIW.
>
> Random particulars:
>
>   - I use Java collections heavily, and wrote some simple procedures
>     to make doing so easier from jscheme (I wish that lists were
>     Lists! - they could be, you know)
>   - I have classes for tab-delimited table input and output, and
>     use these from jscheme; I have a number of other classses and
>     methods exclusively for use from jscheme (e.g. ProcedureComparator
>     for calling Collections.sort with a scheme-supplied comparison
>     function)
>   - I wish I could interrupt runaway evaluations without losing all
>     my state (there's probably a way to do this but I don't know it)
>   - I wish map and for-each would throw exceptions when a list
>     argument isn't a list - current behavior makes bugs harder to
>     track down
>   - I wish the jscheme and java backtraces were interleaved properly
>     (right now I often retry a failing evaluation with
>     (tryCatch ... .printStackTrace) wrapped around the form)
>   - I wish compatibility with R^nRS were a bit higher, esp. regarding
>     numbers.  (/ 1 2) => 0 etc. are very frustrating, makes reuse of
>     existing Scheme code more difficult.
>
> Overall I'm much happier using jscheme than Perl or Java, the other
> primary choices around here.  'wc' tells me my jscheme files total
> about 3000 lines.  And the best part is - I didn't have to write it.
>
> Cheers
> Jonathan

[Jscheme-user] Jscheme at Millennium Pharm.

[Jonathan A R. <ja...@mu...>]

Thought it might be time for me to speak up on what I'm doing, just
FYI... I work at Millennium Pharmaceuticals in the computational
biology group, and am maintaining a sort of 'knowledge base' - a
little database of interactions between biological objects.  This is
used in the analysis of transcription profiling experiments.  We're
dealing with a big curation problem right now in reconciling
information from two different sources (internally curated
vs. Ingenuity-derived).  I'm using Jscheme for all of the analysis at
this point.  This means reading tab-delimited files, writing reports
of many different kinds, doing lookups and filtering and conversions
and set operations and so on.

I also use jscheme for debugging a web services-based application
(that uses GLUE to do SOAP), and for reading an S-expression-syntax
configuration file that's then used by Java code.

We got the seal of approval from the MPI legal department to use
jscheme in applications developed here, FWIW.

Random particulars:

  - I use Java collections heavily, and wrote some simple procedures
    to make doing so easier from jscheme (I wish that lists were
    Lists! - they could be, you know)
  - I have classes for tab-delimited table input and output, and
    use these from jscheme; I have a number of other classses and
    methods exclusively for use from jscheme (e.g. ProcedureComparator
    for calling Collections.sort with a scheme-supplied comparison
    function)
  - I wish I could interrupt runaway evaluations without losing all 
    my state (there's probably a way to do this but I don't know it)
  - I wish map and for-each would throw exceptions when a list
    argument isn't a list - current behavior makes bugs harder to
    track down
  - I wish the jscheme and java backtraces were interleaved properly
    (right now I often retry a failing evaluation with 
    (tryCatch ... .printStackTrace) wrapped around the form)
  - I wish compatibility with R^nRS were a bit higher, esp. regarding
    numbers.  (/ 1 2) => 0 etc. are very frustrating, makes reuse of 
    existing Scheme code more difficult.

Overall I'm much happier using jscheme than Perl or Java, the other
primary choices around here.  'wc' tells me my jscheme files total
about 3000 lines.  And the best part is - I didn't have to write it.

Cheers
Jonathan

Re: [Jscheme-user] scheme-package - another Scheme in Java

[Geoffrey K. <ge...@kn...>]

On Monday, Aug 4, 2003, at 09:50 US/Eastern, Ken Anderson wrote:
> Access from Scheme to Java uses a pre javadot style.

For me, javadot notation is the innovation that makes JScheme a winner 
for me.  Especially when working in an environment where management 
dictates that, say, 90% of code must be in Java, the javadots help with 
the argument that, hey,

1.  JScheme is written in Java.

2.  See all these dots (Java method calls)?  Those are calls to 
standard JARs.

3.  See how JScheme calls to Java are more succinct than plain Java?

4.  You can translate/expand JScheme to plain Java if you have to.  You 
could probably even write a JScheme program to do that for you.

Still, it's nice to hear about "the scheme package" from 
http://koala.ilog.fr/shillion/sp/index.html .  I'll have to give it a 
whirl.  The more Scheme out there being used and refined, the better.

We need a JScheme "killer app," the way Emacs impressed people with the 
power of Lisp, and GIMP helped legitimize Scheme in the view of users.  
Any Scheme killer app would be nice.  It's confining to listen to 
people who think all software should be written in the fad of the 
decade {COBOL, C, Ada, Java, ...}.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-user] scheme-package - another Scheme in Java

[Ken A. <kan...@bb...>]

Here's a scheme in Java i just found:
http://koala.ilog.fr/shillion/sp/index.html

I didn't see any more recent development than 1999, but it looks well 
written.  It uses a much more object oriented style than JScheme does, 
so the code looks cleaner, though you have to unrwap objects to get the 
the java values inside.  Also, each standard piece of syntax is 
represented by a class.

Access from Scheme to Java uses a pre javadot style.

k

Re: [Jscheme-user] Re: Jscheme Webapp

[Timothy J. H. <ti...@cs...>]

On Wednesday, July 30, 2003, at 09:59 AM, Ken Anderson wrote:

> I think Tim can answer the question about webapp.
The webapp should now be available from those links, but you can also  
just download the source
using CVS
   % cvs  
-d:pserver:ano...@cv...:/cvsroot/jscheme login
    (when prompted for a password for anonymous, simply press the Enter  
key)
   % cvs -z3  
-d:pserver:ano...@cv...:/cvsroot/jscheme co  
jscheme
and then build jscheme
   % src/build/bootstrap
then copy the jscheme jar into the webapp lib
    % cp lib/jscheme.jar  src/webapp/jscheme/WEB-INF/lib
If you then copy src/webapp/jscheme into a servlet container (e.g.  
Tomcat or Jetty or ....)
and restart the servlet container, the webapp should work...
Please tell me if you have any problems....
----Tim---

P.S. to get the mail demo to work you have to add the activation.jar  
file from the JavaMail site

> You can certain get the stuff from cvs.
>
> The editor i use is xemacs. The file src\emacs\scheme-init.el
> is an emacs extension that will let you run jscheme by typing the  
> command
> M-x jscheme
>
> There is also a plugin for Eclipse, but i haven't tried it yet:
> http://zclipse.org/projects/jscheme/
> k
> At 02:59 PM 7/27/2003 -0700, you wrote:
>> Dear Ken,
>>
>> First of all, thank you for your work on the Jscheme project.  I have  
>> a couple of questions:
>>
>> 1.  I am unable to access the webapp modules.  When I click on either  
>> of the links in the
>> "Download the jscheme webapp  
>> (<http://jscheme.sourceforge.net/jscheme/download/ 
>> jschemewebapp.zip>zip or  
>> <http://jscheme.sourceforge.net/jscheme/download/ 
>> jschemewebapp.jar>jar)" I get a "The Page cannot be found" response.   
>> Is the webapp sitll available?
>>
>> 2.  Can you recommend a programmer's editor for Scheme?
>>
>> Many thanks,
>>
>> Michael
>>
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> Data Reports, E-commerce, Portals, and Forums are available now.
> Download today and enter to win an XBOX or Visual Studio .NET.
> http://aspnet.click-url.com/go/psa00100003ave/ 
> direct;at.aspnet_072303_01/01
> _______________________________________________
> Jscheme-user mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-user

Re: [Jscheme-user] Re: Jscheme Webapp

[Michael R H. <bu...@su...>]

On Wed, 2003-07-30 at 09:59, Ken Anderson wrote:

> There is also a plugin for Eclipse, but i haven't tried it yet:
> http://zclipse.org/projects/jscheme/

Thanks for mentioning it ;-)

Some details on the current state...

Right now it does some basic (and imperfect) syntax highlighting and
installs jscheme icons into the Eclipse workbench. One thing that I find
really useful is that it loads up all the symbols in elf/basic.scm so
that it can embolden at least the basic jscheme function names.

Currently, the user has to load the jscheme.jar into the project to run
scheme files (and use the standard java running tools to call
jscheme.REPL), but I'm working (slowly) on an eval button and better
launching tools.

Maybe I should get some screenshots and intro docs up on the web site...

I do have a changelog and todo list for those interested:

http://cvs.zclipse.org/cgi-bin/viewcvs.cgi/zclipse/org.zclipse.jscheme/CHANGELOG?rev=HEAD&content-type=text/vnd.viewcvs-markup
http://cvs.zclipse.org/cgi-bin/viewcvs.cgi/zclipse/org.zclipse.jscheme/TODO?rev=HEAD&content-type=text/vnd.viewcvs-markup

mike

-- 
Michael R Head <bu...@su...>
http://suppressingfire.org/~burner/gpg.key.txt

[Jscheme-user] Re: Jscheme Webapp

[Ken A. <kan...@bb...>]

I think Tim can answer the question about webapp.
You can certain get the stuff from cvs.

The editor i use is xemacs. The file src\emacs\scheme-init.el
is an emacs extension that will let you run jscheme by typing the command
M-x jscheme

There is also a plugin for Eclipse, but i haven't tried it yet:
http://zclipse.org/projects/jscheme/
k
At 02:59 PM 7/27/2003 -0700, you wrote:
>Dear Ken,
> 
>First of all, thank you for your work on the Jscheme project.  I have a couple of questions:
> 
>1.  I am unable to access the webapp modules.  When I click on either of the links in the 
>"Download the jscheme webapp (<http://jscheme.sourceforge.net/jscheme/download/jschemewebapp.zip>zip or <http://jscheme.sourceforge.net/jscheme/download/jschemewebapp.jar>jar)" I get a "The Page cannot be found" response.  Is the webapp sitll available?
> 
>2.  Can you recommend a programmer's editor for Scheme?
> 
>Many thanks,
> 
>Michael
> 

[Jscheme-user] [Fwd: [ANN] Viva - Open Source Java Site Goes Live]

[Michael R H. <bu...@cs...>]

This just came across my inbox from the debian-java list. 

Perhaps the JScheme project would be interested in getting listed under
the "Open-Source Scripting Languages For Java" section?

mike
-- 
Michael R Head <bu...@cs...>
Brandeis University CS graduate student

[Jscheme-user] elf/sort.scm

[Ken A. <kan...@bb...>]

elf/sort.scm is a group of procedures i've been using that let you analyze Scheme data in a relatively relational way, so its easy to combine expressions to get the query you want.

So for example 

(map* (project (unique x) (count y))
      (group-by (project x y) data)

is the same as the SQL query :
select unique(x), count(y) from data group by x, y

Rusty wanted me to add the histogram code we've been using also, which is a great way to summarize aspects of your data.

We've been successfully analyzing  60,000+ element datasets with JScheme.

[Jscheme-user] Re: friedman: object-oriented style (in Scheme)

[Ken A. <kan...@bb...>]

This is a very interesting paper!  I think it will be as important as Gregor's Tiny CLOS, and PCL.


At 09:50 AM 7/14/2003 -0400, Geoffrey Knauth wrote:
>This just showed up on LtU:  
>http://www.cs.indiana.edu/hyplan/dfried/ooo9.pdf
>
>(invited talk for Oct 2003 Int'l Lisp conference)
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-user] less perl, more jscheme?

[Geoffrey K. <ge...@kn...>]

On Thursday, Jun 26, 2003, at 16:05 US/Eastern, Ken Anderson wrote:
> If you didn't have the "03" conversion problem, you could do something 
> like this
> (map* identity (open-input-file "data.file"))

Wonderful, thanks!  I'll use that model for general-purpose files.
Only one file, with yymmdd data, needed the yyyymmdd transformation.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-user] less perl, more jscheme?

[Ken A. <kan...@bb...>]

At 02:12 PM 6/26/2003 -0400, Ken Anderson wrote:
>At 08:17 PM 6/25/2003 -0400, Geoffrey Knauth wrote:
>>This worked perfectly on the first try, many thanks!  --Geoffrey
>
>It's also nice that its less code then the Perl. 

The version i sent iterates over the lines of a BufferedReader.
If you didn't have the "03" conversion problem, you could do something like this

(map* identity (open-input-file "data.file"))

which maps over each expression in the input file. 

Re: [Jscheme-user] less perl, more jscheme?

[Ken A. <kan...@bb...>]

At 08:17 PM 6/25/2003 -0400, Geoffrey Knauth wrote:
>This worked perfectly on the first try, many thanks!  --Geoffrey

It's also nice that its less code then the Perl. 

Re: [Jscheme-user] less perl, more jscheme?

[Geoffrey K. <ge...@kn...>]

This worked perfectly on the first try, many thanks!  --Geoffrey

On Wednesday, Jun 25, 2003, at 16:33 US/Eastern, Ken Anderson wrote:

> Try something like this:
>
> (load "elf/basic.scm")
>
> (define data
>   (map* (lambda (line)
>           (string->expr (string-append "(20" (.substring line 1))))
>         (java.io.BufferedReader. (java.io.FileReader. 
> "d:/temp/geof.data"))))
>
>
> At 03:51 PM 6/25/2003 -0400, Geoffrey Knauth wrote:
>> I have some code which reads a bunch of data in sexps:
>>
>> (yymmdd blah blah blah ...)
>> (yymmdd blah blah blah ...)
>> (yymmdd blah blah blah ...)
>>
>> I can read all of them with this JScheme code:
>>
>> (load "elf/load.scm")
>> (define in (openInputPort "foo"))
>> (define entries (eval (read in)))
>> (.close in)
>>
>> if I first process the file with this Perl script:
>>
>> print "'(\n";
>> while (<>) {
>>    chomp $_;
>>    $_ =~ s/^\(03/\(2003/;   # so that yymmdd won't be interpreted by 
>> JScheme as octal
>>    print "$_\n";
>> }
>> print ")\n";
>>
>> which puts '( and ) around the entire contents of the original file.
>>
>> Is there a way to do the whole thing in JScheme and get rid of the 
>> Perl pre-processing script?
>>
>> I can handle the regexp substitution.  What I was having trouble 
>> doing was using eval and read to ingest the data that lacked the 
>> surrounding '( and ).
>>
>> Geoffrey
>> --
>> Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-user] less perl, more jscheme?

[Ken A. <kan...@bb...>]

Try something like this:

(load "elf/basic.scm")

(define data
  (map* (lambda (line)
          (string->expr (string-append "(20" (.substring line 1))))
        (java.io.BufferedReader. (java.io.FileReader. "d:/temp/geof.data"))))


At 03:51 PM 6/25/2003 -0400, Geoffrey Knauth wrote:
>I have some code which reads a bunch of data in sexps:
>
>(yymmdd blah blah blah ...)
>(yymmdd blah blah blah ...)
>(yymmdd blah blah blah ...)
>
>I can read all of them with this JScheme code:
>
>(load "elf/load.scm")
>(define in (openInputPort "foo"))
>(define entries (eval (read in)))
>(.close in)
>
>if I first process the file with this Perl script:
>
>print "'(\n";
>while (<>) {
>    chomp $_;
>    $_ =~ s/^\(03/\(2003/;   # so that yymmdd won't be interpreted by JScheme as octal
>    print "$_\n";
>}
>print ")\n";
>
>which puts '( and ) around the entire contents of the original file.
>
>Is there a way to do the whole thing in JScheme and get rid of the Perl pre-processing script?
>
>I can handle the regexp substitution.  What I was having trouble doing was using eval and read to ingest the data that lacked the surrounding '( and ).
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: INetU
>Attention Web Developers & Consultants: Become An INetU Hosting Partner.
>Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
>INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php
>_______________________________________________
>Jscheme-user mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-user

[Jscheme-user] less perl, more jscheme?

[Geoffrey K. <ge...@kn...>]

I have some code which reads a bunch of data in sexps:

(yymmdd blah blah blah ...)
(yymmdd blah blah blah ...)
(yymmdd blah blah blah ...)

I can read all of them with this JScheme code:

(load "elf/load.scm")
(define in (openInputPort "foo"))
(define entries (eval (read in)))
(.close in)

if I first process the file with this Perl script:

print "'(\n";
while (<>) {
     chomp $_;
     $_ =~ s/^\(03/\(2003/;   # so that yymmdd won't be interpreted by 
JScheme as octal
     print "$_\n";
}
print ")\n";

which puts '( and ) around the entire contents of the original file.

Is there a way to do the whole thing in JScheme and get rid of the Perl 
pre-processing script?

I can handle the regexp substitution.  What I was having trouble doing 
was using eval and read to ingest the data that lacked the surrounding 
'( and ).

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk