Is there any code out there that, using Allegroserve
(or otherwise) implements some form of Lisp Listener
so that one can just point a browser at a web page
and get a Read-Eval-Print loop, or fancier?
(I have a few lines of code that implement a VERY simple
such mechanism, and it could be extended; but I'm guessing
there may be a number of such codes already existing, some
available freely.)
Thanks for info, pointers, etc.
Not really very nice but its a playground for something in future, the HTML
page to start with (optional
if you want to tweak the code). Feel free to tear it apart with criticism
(back it up with suggestions).
------------
<html>
<head>
<title>Lisp Engine</title>
</head>
<body>
<table border="0" cellspacing="0" cellpadding="0" width="100%">
<tbody><tr>
<td width="20%" valign="top"> </td>
<td width="60%" valign="top">
<center>
<h1>Lisp web interface</h1>
</center>
<p>
<font face="Verdana, Arial, Helvetica, sans-serif" size="2">
Type a Lisp expression in the window ... Then click and pray. If its a
definition like DEFUN, DEFCLASS or DEFMETHOD ... it will save to the
running Lisp image and get reloaded as soon as you click the back button
or the link, which will let you call it the next time around.
<p>
To indent your Lisp code and match parentheses use Emacs or vi and just
paste the code in the space below.
<p>
</font></p>
<pre>
Examples : <p>
((lambda (x)(* x x)) 5)
(defun square (x)(* x x))
</pre>
<p>
<font face="Verdana, Arial, Helvetica, sans-serif" size="2">
<form name="input" action="cgi-bin/stdio3.lisp" method="post">
<b>S-Expr :</b><br>
<textarea name="Expr" rows="15" cols="70"></textarea>
<p>
<input type="submit" value="Evaluate Expression">
<p>
</form>
</font>
</p></td>
<td width="20%" valign="top"> </td>
</tr>
</tbody></table>
</body></html>
--------------------------
The Lisp code : (warning: really bad code but working)
This is written in CLISP and you would have to dump the Lisp image in
the CGI-BIN directory once first.
--------------------------
#!/usr/local/bin/clisp -M lispinit.mem
(let ((env "Content-type: text/html"))
(format t "~A~%" env)
(format t "~%"))
(setq *field-storage* (make-hash-table :test #'equalp))
(if (null *args*)
(setq *query-string* (read-line)
*request-method* "POST")
(setq *query-string* *args*
*request-method* "GET"))
(if (string= *request-method* "GET")
(setq *query-string*
(apply #'concatenate 'string
(mapcar #'(lambda (s)(concatenate 'string s " ")) *query-string*)))
(if (string= *request-method* "POST")
(setq *query-string* (substitute #\Space #\+ *query-string*))))
;; Split string and store in hash table
(defun str-to-hash (str hash sep1 sep2)
(let ((p1 (position sep1 str)))
(if p1
(progn
(let ((l1 (subseq str 0 p1))
(r1 (subseq str (1+ p1))))
(if l1
(let ((p2 (position sep2 l1)))
(if p2
(let ((l2 (subseq l1 0 p2))
(r2 (subseq l1 (1+ p2))))
(if (and l2 r2)
(setf (gethash l2 hash) r2))))))
(if r1
(str-to-hash r1 hash sep1 sep2))))
(progn
(if str
(let ((p2 (position sep2 str)))
(if p2
(let ((l2 (subseq str 0 p2))
(r2 (subseq str (1+ p2))))
(if (and l2 r2)
(setf (gethash l2 hash) r2))))))))))
;; Decoding url ecoding special characters
(defun url-decode (url)
(if (string= *request-method* "POST")
(let ((p (position #\% url)))
(if p
(progn
(let ((l (subseq url 0 p))
(c (subseq url (+ p 1) (+ p 3)))
(r (subseq url (+ p 3))))
(cond ((string-equal c "0D")
(progn
(let ((url (concatenate 'string l " " r)))
(url-decode url))))
((string-equal c "0A")
(progn
(let ((url (concatenate 'string l r)))
(url-decode url))))
((string-equal c "25")
(progn
(let ((url (concatenate 'string l "|25" r)))
(url-decode url))))
(t (progn
(let* ((h (string (character (parse-integer c :radix 16))))
(url (concatenate 'string l h r)))
(url-decode url)))))))
url))))
(defun convert-special (s)
(let ((p1 (position #\| s)))
(if p1
(progn
(let ((c (subseq s (+ p1 1)(+ p1 3))))
(if (string-equal c "25")
(progn
(let ((new-s (concatenate 'string (subseq s 0 p1) "%" (subseq s (+ p1
3)))))
(convert-special new-s)))
s))) s)))
(str-to-hash *query-string* *field-storage* #\& #\=)
(maphash #'(lambda (k v)
(setf (gethash k *field-storage*)
(convert-special (url-decode v)))) *field-storage*)
;; Hash table accessor
(defun get-value (fld)
(gethash fld *field-storage*))
(defun br ()
(princ "<br>"))
(defun p ()
(princ "<P>"))
;;--------------------------------------------------------------------------
---
(princ "<html>")
(princ "<head>")
(princ "<title>Lisp Engine</title>")
(princ "</head>")
(princ "<body>")
(princ "<table border=\"0\" cellspacing=\"0\" cellpadding=\"0\"
width=\"100%\">")
(princ "<tbody><tr>")
(princ "<td width=\"20%\" valign=\"top\"></td>")
(princ "<td width=\"60%\" valign=\"top\">")
(princ "<center>")
(princ "<h1>Lisp web interface</h1>")
(princ "</center>")
(princ "<p>")
(princ "<font face=\"Verdana, Arial, Helvetica, sans-serif\" size=\"2\">")
;;-------------------------
(princ "<b>Evaluated The S-Exp</b>")
(p)
(princ "S-expression = ")
(princ (get-value "expr"))
(p)
(princ "result = ")
(let ((exp (get-value "expr")))
(princ (eval (read-from-string exp))))
(p)
;;----------------
(princ "Go on ... you can continue.")
(princ "<p>")
(princ "<form name=\"input\" action=\"stdio3.lisp\" method=\"post\">")
(princ "<b>S-Expr :</b><br>")
(princ "<textarea name=\"Expr\" rows=\"15\" cols=\"70\"></textarea>")
(princ "<p>")
(princ "<input type=\"submit\" value=\"Evaluate Expression\">")
(princ "<p>")
(princ "</form>")
(princ "</font>")
(princ "</p></td>")
(princ "<td width=\"20%\" valign=\"top\"></td>")
(princ "</tr>")
(princ "</tbody></table>")
(princ "</body></html>")
;;--------------------------------------------------------------------------
-
;; Save image if there is any definition for persistent functions and
classes.
(if (string= (substring (car (read-from-string (get-value "expr"))) 0 3) (or
"SET" "DEF"))
(ext:saveinitmem "lispinit.mem"))
"JP Massar" <······@alum.mit.edu> wrote in message
·······················@netnews.attbi.com...
> Is there any code out there that, using Allegroserve
> (or otherwise) implements some form of Lisp Listener
> so that one can just point a browser at a web page
> and get a Read-Eval-Print loop, or fancier?
>
> (I have a few lines of code that implement a VERY simple
> such mechanism, and it could be extended; but I'm guessing
> there may be a number of such codes already existing, some
> available freely.)
>
> Thanks for info, pointers, etc.
>
>
You realize that this is horribly insecure, right?
I thought of doing this a while ago and even came to the point of
writing a Lisp reader so that I could have complete control of it. I
would intern all symbols in a sandbox package and would import CL
symbols into it on a case-by-case basis. I believe there was no way to
break out of it, but I'm not entirely sure either. I played with it for
a bit and got it to work, then put it aside.
If I ever get around to cleaning it up, I might post it. Admittedly,
the reader implementation is sucky, and might even be unnecessary.
--
; Matthew Danish <·······@andrew.cmu.edu>
; OpenPGP public key: C24B6010 on keyring.debian.org
; Signed or encrypted mail welcome.
; "There is no dark side of the moon really; matter of fact, it's all dark."
Yes, I know that
e.g, (ext:run-shell-command "rm -rf somedirectory")
It was just an experiment for something else. No production use or even
thoughts.
Take Care
"Matthew Danish" <·······@andrew.cmu.edu> wrote in message
···························@lain.cheme.cmu.edu...
> You realize that this is horribly insecure, right?
>
> I thought of doing this a while ago and even came to the point of
> writing a Lisp reader so that I could have complete control of it. I
> would intern all symbols in a sandbox package and would import CL
> symbols into it on a case-by-case basis. I believe there was no way to
> break out of it, but I'm not entirely sure either. I played with it for
> a bit and got it to work, then put it aside.
>
> If I ever get around to cleaning it up, I might post it. Admittedly,
> the reader implementation is sucky, and might even be unnecessary.
>
> --
> ; Matthew Danish <·······@andrew.cmu.edu>
> ; OpenPGP public key: C24B6010 on keyring.debian.org
> ; Signed or encrypted mail welcome.
> ; "There is no dark side of the moon really; matter of fact, it's all
dark."
I have a file where I stuff some of the Lisp code that gets posted here.
(I should do it more often.) Everything from small things like Thomas
Burdick's interesting use of packages for defclass namespacing, to
larger things like the URL decoder from OCID and the sandbox Matthew
Danish talks about here.
It would be nice if there were a searchable repository where it was easy
for people to post everything from little snippets, to medium-sized
routines, to larger working units. It is hard to comb through c.l.l
archives to pull out just the code snippets, especially by category.
There are times when that is exactly what is desired.
Any ideas on how such a repository would work and how it could be
integrated with the c.l.l discussions? Feature sets? Categorization?
Portability classifications or whatever? Search options?
If it were easy to make submissions, easy to refer to from a c.l.l post,
easy to search, and possibly easy to download/integrate code from the
repository, it could grow to be quite a resource within a relatively
short time.
Matthew Danish wrote:
> You realize that this is horribly insecure, right?
>
> I thought of doing this a while ago and even came to the point of
> writing a Lisp reader so that I could have complete control of it. I
> would intern all symbols in a sandbox package and would import CL
> symbols into it on a case-by-case basis. I believe there was no way to
> break out of it, but I'm not entirely sure either. I played with it for
> a bit and got it to work, then put it aside.
>
> If I ever get around to cleaning it up, I might post it. Admittedly,
> the reader implementation is sucky, and might even be unnecessary.
>
Jeff Caldwell <·····@yahoo.com> writes:
> I have a file where I stuff some of the Lisp code that gets posted
> here. (I should do it more often.) Everything from small things like
> Thomas Burdick's interesting use of packages for defclass namespacing,
> to larger things like the URL decoder from OCID and the sandbox
> Matthew Danish talks about here.
>
> It would be nice if there were a searchable repository where it was
> easy for people to post everything from little snippets, to
> medium-sized routines, to larger working units. It is hard to comb
> through c.l.l archives to pull out just the code snippets, especially
> by category. There are times when that is exactly what is desired.
>
> Any ideas on how such a repository would work and how it could be
> integrated with the c.l.l discussions? Feature sets? Categorization?
> Portability classifications or whatever? Search options?
>
> If it were easy to make submissions, easy to refer to from a c.l.l
> post, easy to search, and possibly easy to download/integrate code
> from the repository, it could grow to be quite a resource within a
> relatively short time.
>
One easy way to do it would be something like sourceforge,
web space(project page) with cvs and a cvs browser. It would
be good PR if this was in CL though. If it was a big hit that
could be a problem, someone needs to pay for connectivity and
hardware at a minimum.
openacs(openacs.org) might be another good model.
marc
Jeff Caldwell <·····@yahoo.com> writes:
> It would be nice if there were a searchable repository where it was
> easy for people to post everything from little snippets, to
> medium-sized routines, to larger working units.
Maybe these are relevant to this idea:
http://www.cliki.net/Anthology
http://examples.franz.com
Henrik Motakef <··············@web.de> writes:
> Jeff Caldwell <·····@yahoo.com> writes:
>
> > It would be nice if there were a searchable repository where it was
> > easy for people to post everything from little snippets, to
> > medium-sized routines, to larger working units.
>
> Maybe these are relevant to this idea:
>
> http://www.cliki.net/Anthology
> http://examples.franz.com
I like examples.franz.com one, looks kinda like the cpan web interface.
One of the nice things about a non wiki solution is that it prevents
people from adding new catagories, this can also be a bad thing.
One thing that I would liek to see is a more CPANish layout
for the directory tree, here is the top level:
[DIR] Parent Directory -
[DIR] 02_Perl_Core_Modules/ 31-Mar-2003 09:43 -
[DIR] 03_Development_Support/ 23-Mar-2003 02:41 -
[DIR] 04_Operating_System_Interfaces/ 17-Mar-2003 02:42 -
[DIR] 05_Networking_Devices_IPC/ 22-May-2002 16:40 -
[DIR] 06_Data_Type_Utilities/ 12-May-2003 01:44 -
[DIR] 07_Database_Interfaces/ 16-May-2003 06:15 -
[DIR] 08_User_Interfaces/ 25-Apr-2003 05:13 -
[DIR] 09_Language_Interfaces/ 04-Oct-2001 03:14 -
[DIR] 10_File_Names_Systems_Locking/ 22-Sep-2002 01:24 -
[DIR] 11_String_Lang_Text_Proc/ 15-May-2003 00:19 -
[DIR] 12_Opt_Arg_Param_Proc/ 23-Jan-2003 06:49 -
[DIR] 13_Internationalization_Locale/ 24-Mar-2003 09:11 -
[DIR] 14_Security_and_Encryption/ 11-Oct-2001 05:33 -
[DIR] 15_World_Wide_Web_HTML_HTTP_CGI/ 24-Apr-2003 06:44 -
[DIR] 16_Server_and_Daemon_Utilities/ 20-Mar-2003 13:42 -
[DIR] 17_Archiving_and_Compression/ 08-Apr-2001 14:08 -
[DIR] 18_Images_Pixmaps_Bitmaps/ 14-Apr-2003 08:13 -
[DIR] 19_Mail_and_Usenet_News/ 11-May-2003 23:45 -
[DIR] 20_Control_Flow_Utilities/ 23-Aug-2002 08:52 -
[DIR] 21_File_Handle_Input_Output/ 06-Jun-2002 01:42 -
[DIR] 22_Microsoft_Windows_Modules/ 30-Dec-1997 12:49 -
[DIR] 23_Miscellaneous_Modules/ 11-May-2003 12:44 -
[DIR] 24_Commercial_Software_Interfaces/ 27-Mar-2003 03:43 -
[DIR] 99_Not_In_Modulelist/ 09-May-2001 00:58 -
[DIR] 99_Not_Yet_In_Modulelist/ 02-Sep-2002 15:53 -
and here is 19_Mail_and_Usenet_News (just clicked on it)
[DIR] Parent Directory -
[DIR] IMAP/ 27-Dec-2001 08:13 -
[DIR] Mail/ 26-May-2003 11:15 -
[DIR] Mariachi/ 23-May-2003 06:44 -
[DIR] NNML/ 31-Dec-1997 18:29 -
[DIR] News/ 10-Feb-2003 05:25 -
[DIR] Sendmail/ 09-May-2002 19:30 -
Since the goal would be to have lots of stuff it
would seem to me a good idea to steal a working
directory tree for a layout, why reinvent the
wheel.
You would not copy everything, for example the
NNML dir would not apply to CL, but SMTP would.
marc
The number of entries on the Anthology page makes quite a statement.
(The number is zero, I believe.) People certainly aren't shy about
posting code, or at least code snippets, to c.l.l. I suspect it is the
format or facilities of the Anthologies page, more than people's
willingness to contribute, that accounts for the lack of examples.
Usenet discussions are organized by threads, which is an attribute of
openacs functionality, but there is no index categorizing the snippets
or examples and their associated discussions. They are good to see as
they whiz by and can be found if one happens to search for a lucky
keyword. I don't see how they can be systematically researched. Usenet
threads are temporally bound in that long-dead threads are less likely
to be posted against than are current threads. An indexed catalog of
code snippets and full examples seems more likely to be found, and thus
more likely to be commented upon, even long after the original posting.
The franz examples page seems to be for full, working examples, and not
also for code snippets and example techniques. Additionally, there does
not seem to be a place for ongoing discussions about each example.
The OpenACS (http://openacs.org) is perhaps close to what I had in mind.
If such a facility were available, I can't predict if it would be used.
It would be nice if it were in CL. I appreciate your feedback.
Marc Spitzer wrote:
> Henrik Motakef <··············@web.de> writes:
>
>
>>Jeff Caldwell <·····@yahoo.com> writes:
>>
>>
>>>It would be nice if there were a searchable repository where it was
>>>easy for people to post everything from little snippets, to
>>>medium-sized routines, to larger working units.
>>
>>Maybe these are relevant to this idea:
>>
>>http://www.cliki.net/Anthology
>>http://examples.franz.com
>
>
> I like examples.franz.com one, looks kinda like the cpan web interface.
> One of the nice things about a non wiki solution is that it prevents
> people from adding new catagories, this can also be a bad thing.
>
> One thing that I would liek to see is a more CPANish layout
> for the directory tree, here is the top level:
> [DIR] Parent Directory -
> [DIR] 02_Perl_Core_Modules/ 31-Mar-2003 09:43 -
> [DIR] 03_Development_Support/ 23-Mar-2003 02:41 -
> [DIR] 04_Operating_System_Interfaces/ 17-Mar-2003 02:42 -
> [DIR] 05_Networking_Devices_IPC/ 22-May-2002 16:40 -
> [DIR] 06_Data_Type_Utilities/ 12-May-2003 01:44 -
> [DIR] 07_Database_Interfaces/ 16-May-2003 06:15 -
> [DIR] 08_User_Interfaces/ 25-Apr-2003 05:13 -
> [DIR] 09_Language_Interfaces/ 04-Oct-2001 03:14 -
> [DIR] 10_File_Names_Systems_Locking/ 22-Sep-2002 01:24 -
> [DIR] 11_String_Lang_Text_Proc/ 15-May-2003 00:19 -
> [DIR] 12_Opt_Arg_Param_Proc/ 23-Jan-2003 06:49 -
> [DIR] 13_Internationalization_Locale/ 24-Mar-2003 09:11 -
> [DIR] 14_Security_and_Encryption/ 11-Oct-2001 05:33 -
> [DIR] 15_World_Wide_Web_HTML_HTTP_CGI/ 24-Apr-2003 06:44 -
> [DIR] 16_Server_and_Daemon_Utilities/ 20-Mar-2003 13:42 -
> [DIR] 17_Archiving_and_Compression/ 08-Apr-2001 14:08 -
> [DIR] 18_Images_Pixmaps_Bitmaps/ 14-Apr-2003 08:13 -
> [DIR] 19_Mail_and_Usenet_News/ 11-May-2003 23:45 -
> [DIR] 20_Control_Flow_Utilities/ 23-Aug-2002 08:52 -
> [DIR] 21_File_Handle_Input_Output/ 06-Jun-2002 01:42 -
> [DIR] 22_Microsoft_Windows_Modules/ 30-Dec-1997 12:49 -
> [DIR] 23_Miscellaneous_Modules/ 11-May-2003 12:44 -
> [DIR] 24_Commercial_Software_Interfaces/ 27-Mar-2003 03:43 -
> [DIR] 99_Not_In_Modulelist/ 09-May-2001 00:58 -
> [DIR] 99_Not_Yet_In_Modulelist/ 02-Sep-2002 15:53 -
>
> and here is 19_Mail_and_Usenet_News (just clicked on it)
>
> [DIR] Parent Directory -
> [DIR] IMAP/ 27-Dec-2001 08:13 -
> [DIR] Mail/ 26-May-2003 11:15 -
> [DIR] Mariachi/ 23-May-2003 06:44 -
> [DIR] NNML/ 31-Dec-1997 18:29 -
> [DIR] News/ 10-Feb-2003 05:25 -
> [DIR] Sendmail/ 09-May-2002 19:30 -
>
>
> Since the goal would be to have lots of stuff it
> would seem to me a good idea to steal a working
> directory tree for a layout, why reinvent the
> wheel.
>
> You would not copy everything, for example the
> NNML dir would not apply to CL, but SMTP would.
>
> marc
Jeff Caldwell <·····@yahoo.com> writes:
> The franz examples page seems to be for full, working examples, and
> not also for code snippets and example techniques. Additionally, there
> does not seem to be a place for ongoing discussions about each example.
It may not be perfect, but it would no doubt be better then anything
we have currently. The only real speed bump, I see, would be if Franz
made it an allegro specific archive. I have no problem with it, it
is their property and they can do with it what they want, but I use
cmu and other people use lispworks.
I still think a CPAN model for modules and applications would do a
world of good. It would give a great counter example for the idea
that no one uses CL, where did all this stuff come from.
>
> The OpenACS (http://openacs.org) is perhaps close to what I had in
> mind. If such a facility were available, I can't predict if it would
> be used. It would be nice if it were in CL. I appreciate your feedback.
You can loose stuff in openacs also, getting to the right place in file
storage can be tricky. Another problem is the dog food issue, if CL is
so great why are you using tcl for this? And it is a *big* system,
porting it would be a big and ugly job. Even if it was just the front
end, aolserver part.
marc
Marc Spitzer <········@optonline.net> writes:
> I still think a CPAN model for modules and applications would do a
> world of good. It would give a great counter example for the idea
> that no one uses CL, where did all this stuff come from.
Good thing that we already have two projects for this, CCLAN and
CLOCC. Problem solved, we can now all go on and actually write those
modules and applications.
The problem is not solved when a person has to fight a dead link on
Cliki and make their way to an out-of-date site that redirected them to
SourceForge only to find that even Vendor Neutral CCLAN is CMUCL-only,
which, of course, is not cross-platform. VN-CCLAN doesn't serve all the
participants in c.l.l.
I was initially interested in a place where code snippets could be
submitted and discussed in a forum which could categorize the
discussions and allow a repository to grow over time. Including
fully-working code as well as snippets is an extension of that idea. A
repository such as CLOCC, with only fully-working code and not
supporting on-going community commentary, is not a superset of what I
had in mind. Certainly what I had in mind would allow a section where
cclan and clocc could be discussed, commented upon, documented, and
linked. Automatic downloading and installation would be icing.
Maybe that's what Cliki is and it's just that nobody wants to do that.
Maybe that kind of a site isn't nearly as fun as usenet. Maybe Cliki is
the wrong tool for the job by being too freeform for people to view it
as friendly or useful for that purpose.
Henrik Motakef wrote:
> Marc Spitzer <········@optonline.net> writes:
>
>
>>I still think a CPAN model for modules and applications would do a
>>world of good. It would give a great counter example for the idea
>>that no one uses CL, where did all this stuff come from.
>
>
> Good thing that we already have two projects for this, CCLAN and
> CLOCC. Problem solved, we can now all go on and actually write those
> modules and applications.
Jeff Caldwell <·····@yahoo.com> writes:
> The problem is not solved when a person has to fight a dead link on
> Cliki and make their way to an out-of-date site that redirected them
> to SourceForge only to find that even Vendor Neutral CCLAN is
> CMUCL-only, which, of course, is not cross-platform. VN-CCLAN doesn't
> serve all the participants in c.l.l.
Though I worked on vn-cclan for a while, I don't want to defend it as
a good or even working project, because I believe that it's fairly
unworkable, at least on an unpaid, volunteer basis. That disclaimer
over, how on earth did you come to the conclusion that it's
CMUCL-only?
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
Christophe,
Thanks for your post. I'm blushing yet I still don't know where I went
wrong. I'll keep looking but I wanted to get back to you because this is
embarassing.
http://sourceforge.net/project/showfiles.php?group_id=28536&release_id=47297
is titled: Project: Comprehensive Common Lisp Archive Net: File List
I downloaded the only project file on that page, which is
clc-install-0.1.tar.gz. I expanded the archive. The README included the
comment:
"I've tested this with CMUCL 18c, and I would be surprised if it would
work with any other version."
The first instruction in the INSTALL file says to do ./clc-install.sh,
which is chock-full of references to cmucl and almost nothing else,
certainly no references to any other Lisp implementations.
I arrived at that download page by clicking on the "Latest File
Releases" Download link at:
http://sourceforge.net/projects/cclan
I arrived at that page from:
http://www.lichteblau.com/code/cclan-get/
which is titled "This is out of date" and includes a link which resolved
to the sourceforge one above.
I arrived at the lichteblau site from:
http://www.cliki.net/vn-cCLan
where
http://www.geocrawler.com/lists/3/SourceForge/12820/0/7981131/,
appears to be dead. The only other links on that page, other than ones
pointing within Cliki itself, are: 1) to a place that sets a mirror site
cookie, and 2) to "cclan-get", which goes to the lichteblau.com site
referenced above.
I followed the only Cliki link that actually took me somewhere where I
could get a file, found a site proclaiming itself out of date, followed
its link on to SourceForge where I found a page title "CCLAN",
downloaded the only file available on the CCLAN SourceForge project
page, and discovered the file is for CMUCL only.
I apologize for misrepresenting your work, and the work of others, on
CCLAN. I was, and remain, ignorant of the facts.
As of now, I am confused about why the SourceForge site is titled
"CCLAN" but the file is a Common Lisp Controller install only for CMUCL
only. I don't know enough about the packages to understand how it all
ties together.
Since the site http://ww.telent.net/cclan-choose-mirror only sets a
cookie, but does not redirect me to the sites named on that page, I will
now go to one of those sites by typing the link manually and try to get
a good copy of CCLAN itself.
If you can tell me how things really are, I volunteer to update the
Cliki page, although I would appreciate someone knowledgable going in
behind me to verify that my changes are accurate.
The choose-mirror page sets a cookie. What is the page, or pages, that
read the cookie and make use of it? I did not find any such links on the
vn-cclan cliki page. Should the vn-cclan Clicki page still link to the
out-of-date lichteblau site? What should be done with the non-responding
geocrawler link on the vn-cclan cliki page? Is there a replacement?
Thank you for bringing the fact that cclan is not cmucl-only to my
attention. I certainly appreciate the efforts of anyone who puts work
into making CL code available.
Jeff
Christophe Rhodes wrote:
> Jeff Caldwell <·····@yahoo.com> writes:
>
>
>>The problem is not solved when a person has to fight a dead link on
>>Cliki and make their way to an out-of-date site that redirected them
>>to SourceForge only to find that even Vendor Neutral CCLAN is
>>CMUCL-only, which, of course, is not cross-platform. VN-CCLAN doesn't
>>serve all the participants in c.l.l.
>
>
> Though I worked on vn-cclan for a while, I don't want to defend it as
> a good or even working project, because I believe that it's fairly
> unworkable, at least on an unpaid, volunteer basis. That disclaimer
> over, how on earth did you come to the conclusion that it's
> CMUCL-only?
>
> Christophe
[ disclaimer: opinions below are personal views only, mostly. ]
Jeff Caldwell <·····@yahoo.com> writes:
> Christophe,
>
> Thanks for your post. I'm blushing yet I still don't know where I went
> wrong. I'll keep looking but I wanted to get back to you because this
> is embarassing.
Agreed, and not just for you -- I think it's clear that this has been
afflicted by a certain amount of scope creep and link rot, and I'll
freely accept my share of the blame for that.
To start with, maybe I should outline my knowledge of the current
state of the art, with all the faults attendant, and then I'll try to
explain what's wrong with the presentation at the moment.
Current "vendor neutral" cclan's expected usage would look something
like this, I think:
* the user will acquire ASDF from somewhere; it may be distributed
with the lisp he wants to use, or it may not. If all else fails, it
can be pulled from CVS.
* the user will download some piece of software from a cCLan node,
maybe prompted by following a download link from CLiki. (an example
of what a cCLan node looks like, and what it contains, can be seen
at <http://ftp.linux.org.uk/pub/lisp/experimental/cclan/>, though
note that that's probably not a permanent location (see the
"experimental" in the URL :-).
* the user will untar it somewhere in his filesystem, and symlink[1]
the relevant asdf system description to a path that is on
asdf:*central-registry* (a customizable variable that controls some
of ASDF's behaviour).
* finally, the user will issue an asdf command to compile and load
the system; this is something like
(asdf:operate 'asdf:load-op :<name of system>). If the software
that has been downloaded is supported by the implementation that the
user is programming in, all is well; if not, then something will go
wrong. There isn't any guarantee of any given library working with
any given lisp implementation beyond those of the library authors.
I'm fairly sure that this process, apart from the [1] above which
obviously doesn't work on some non-Unix platforms, is vendor-neutral.
What might not be vendor-neutral are some of the libraries that are
distributed by the system, but if we restricted distribution to those
that worked on all lisps purporting to conform to ANSI, we might well
end up with the null set. :-/
> http://sourceforge.net/project/showfiles.php?group_id=28536&release_id=47297
>
> is titled: Project: Comprehensive Common Lisp Archive Net: File List
>
> I downloaded the only project file on that page, which is
> clc-install-0.1.tar.gz. I expanded the archive. The README included
> the comment:
>
> "I've tested this with CMUCL 18c, and I would be surprised if it would
> work with any other version."
>
> The first instruction in the INSTALL file says to do ./clc-install.sh,
> which is chock-full of references to cmucl and almost nothing else,
> certainly no references to any other Lisp implementations.
Right. This, I think, was an early attempt to port Debian's "common
lisp controller" to general Unixoid systems.
> As of now, I am confused about why the SourceForge site is titled
> "CCLAN" but the file is a Common Lisp Controller install only for
> CMUCL only. I don't know enough about the packages to understand how
> it all ties together.
Sure... there is currently some development happening in the
sourceforge CCLAN CVS; none of it has been "published" or "released"
in any real way, sadly.
> Since the site http://ww.telent.net/cclan-choose-mirror only sets a
> cookie, but does not redirect me to the sites named on that page, I
> will now go to one of those sites by typing the link manually and try
> to get a good copy of CCLAN itself.
The second time you choose a Download link (or once you've set a
cookie), say, from <http://www.cliki.net/Araneida>, you should
automatically start the download from whichever mirror you chose.
> If you can tell me how things really are, I volunteer to update the
> Cliki page, although I would appreciate someone knowledgable going in
> behind me to verify that my changes are accurate.
Thanks! That would be great, and I'll certainly be happy to review
what you write.
> The choose-mirror page sets a cookie. What is the page, or pages, that
> read the cookie and make use of it? I did not find any such links on
> the vn-cclan cliki page.
There aren't any on the vn-cclan page itself; libraries described on
CLiki that are downloadable though "vn" cCLan will have a download
link that should allow "one click" downloads, in theory at least.
> Should the vn-cclan Clicki page still link
> to the out-of-date lichteblau site? What should be done with the
> non-responding geocrawler link on the vn-cclan cliki page? Is there a
> replacement?
These are good questions. I don't have any immediate answers to them,
unfortunately.
> Thank you for bringing the fact that cclan is not cmucl-only to my
> attention. I certainly appreciate the efforts of anyone who puts work
> into making CL code available.
At this point, I think cCLan is more-or-less a distribution network
with a central point of contact (CLiki) and a standard way of
informing the lisp implementation of the availability of the package
(ASDF). As such, its focus is more on what it's distributing than on
any given implementation.
I hope that helps somewhat; I'll note that there are mailing lists for
cclan, where perhaps more people might be able to answer specific
questions (though don't hesitate to continue here should you think
that more applicable).
Cheers,
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
Jeff Caldwell <·····@yahoo.com> writes:
> I arrived at that download page by clicking on the "Latest File
> Releases" Download link at:
>
> http://sourceforge.net/projects/cclan
As you've correctly surmised, it's a stale archive relating to an
earlier attempt at an archive system. Sourceforge won't let me delete
it, but I have at least managed to hide it from view. Thanks for
pointing it out.
There's still a ton of stuff that would actually need doing to
resuscitate cclan, if there were the necessary level of interest in
doing so, but at least we might have saved the next person to stiumble
across it from wasting a few hours.
-dan
--
http://www.cliki.net/ - Link farm for free CL-on-Unix resources
Henrik Motakef <··············@web.de> writes:
> Marc Spitzer <········@optonline.net> writes:
>
> > I still think a CPAN model for modules and applications would do a
> > world of good. It would give a great counter example for the idea
> > that no one uses CL, where did all this stuff come from.
>
> Good thing that we already have two projects for this, CCLAN and
> CLOCC. Problem solved, we can now all go on and actually write those
> modules and applications.
I do not know, I have quickly looked at both. From what little I know
about both projects they are conceptually different from CPAN. CPAN is
really nice for "a la cart" shopping I do not download/install all of
CPAN, just the parts I want/need to get work done. Also from the cclan
page on cliki it states its goal is to get things working on at least
1 free version of CL, this seems to imply that commercial user should
go some where else for their needs. The vendor neutral CCLAN looks nice,
but it does not exist yet. I use freebsd so cclan does me no good.
Another thing is that CPAN lives in Perl not in the host os.
marc
On Wed, 28 May 2003 18:42:57 -0400, Matthew Danish
<·······@andrew.cmu.edu> wrote:
>You realize that this is horribly insecure, right?
>
Certainly. But I don't care.
What I was interested in was not having to reinvent all the code
that generates the HTML to provide a reasonable user interface for the
REP loop.
What the REP loop then allows to execute is a separate issue.
So really I'm talking about not a Lisp Listener, per se, but simply
a REP loop web interface for an arbitrary language.
I find it a bit strange that there isn't anything obvious out there.
Maybe I'm Google searching on the wrong terms but nothing seems to
come up that looks right.
>I thought of doing this a while ago and even came to the point of
>writing a Lisp reader so that I could have complete control of it. I
>would intern all symbols in a sandbox package and would import CL
>symbols into it on a case-by-case basis. I believe there was no way to
>break out of it, but I'm not entirely sure either. I played with it for
>a bit and got it to work, then put it aside.
>
>If I ever get around to cleaning it up, I might post it. Admittedly,
>the reader implementation is sucky, and might even be unnecessary.
>
JP Massar schrieb:
> Is there any code out there that, using Allegroserve
> (or otherwise) implements some form of Lisp Listener
> so that one can just point a browser at a web
Hallo
reading these texts I thought about what would an author say when you
snip around in his postings and present that elsewhere?
So, is it legal to take parts of news-posts and copy them without
acknowledgment of the posters?
stefan
"LIN8080" <·······@freenet.de> wrote in message ······················@freenet.de...
>
> reading these texts I thought about what would an author say when you
> snip around in his postings and present that elsewhere?
>
> stefan
Since I occasionally post code (and sometimes it even works!),
I thought I'd answer.
I'm happy to let people use the code snippets I write and post.
I'd be a little put off if someone took it and presented it as
their own, though. It doesn't take *that* much effort to
acknowledge where you got it from.
> So, is it legal to take parts of news-posts and copy them without
> acknowledgment of the posters?
No. It isn't.
Technically speaking, there is a presumed copyright and you can't
just snarf the code. Even if you *do* acknowledge the authors, you
still need their permission. (Actually, there are some `fair use'
considerations, but the right thing to do is ask.)
Besides, would *you* like it if someone did that? How hard is it
to just ask?
"Joe Marshall" <·············@attbi.com> writes:
> "LIN8080" <·······@freenet.de> wrote in message ······················@freenet.de...
> >
> > reading these texts I thought about what would an author say when you
> > snip around in his postings and present that elsewhere?
> >
> > stefan
>
> Since I occasionally post code (and sometimes it even works!),
> I thought I'd answer.
>
> I'm happy to let people use the code snippets I write and post.
> I'd be a little put off if someone took it and presented it as
> their own, though. It doesn't take *that* much effort to
> acknowledge where you got it from.
>
> > So, is it legal to take parts of news-posts and copy them without
> > acknowledgment of the posters?
>
> No. It isn't.
> Technically speaking, there is a presumed copyright and you can't
> just snarf the code. Even if you *do* acknowledge the authors, you
> still need their permission. (Actually, there are some `fair use'
> considerations, but the right thing to do is ask.)
>
> Besides, would *you* like it if someone did that? How hard is it
> to just ask?
In my opinion, code posted in newsgroups are more code snippets than
full fleshed applications or libraries. Even if a complete script is
posted, it's there more like an tutorial example and can be copied if
needed. Like when you copy the quick sort algorithm posted on the
black board by your teacher.
I would rather presume public domain!
But of course, IANAL, and ITTALSBKOTS.
--
__Pascal_Bourguignon__ http://www.informatimago.com/
----------------------------------------------------------------------
Do not adjust your mind, there is a fault in reality.
Pascal Bourguignon <····@thalassa.informatimago.com> writes:
> In my opinion, code posted in newsgroups are more code snippets than
> full fleshed applications or libraries. Even if a complete script is
> posted, it's there more like an tutorial example and can be copied if
> needed. Like when you copy the quick sort algorithm posted on the
> black board by your teacher.
>
> I would rather presume public domain!
You presume wrongly. See "10 Big Myths about copyright explained" at
<http://www.templetons.com/brad/copymyths.html>, with especial
reference to numbers 3 and 4
-dan
--
http://www.cliki.net/ - Link farm for free CL-on-Unix resources
This URL leads to a more succinct, application of copyright, too newsgroups.
http://help.prodigy.net/help/newsgroup_bb/newscopyright.html
Note the bullet just after "Fair Use:" starting with "An example..."
I wonder how copyright applies to Google's Data base, of past newsgroup
posts?
"Daniel Barlow" <···@telent.net> wrote in message
···················@noetbook.telent.net...
> Pascal Bourguignon <····@thalassa.informatimago.com> writes:
>
> > In my opinion, code posted in newsgroups are more code snippets than
> > full fleshed applications or libraries. Even if a complete script is
> > posted, it's there more like an tutorial example and can be copied if
> > needed. Like when you copy the quick sort algorithm posted on the
> > black board by your teacher.
> >
> > I would rather presume public domain!
>
> You presume wrongly. See "10 Big Myths about copyright explained" at
> <http://www.templetons.com/brad/copymyths.html>, with especial
> reference to numbers 3 and 4
>
>
> -dan
>
> --
>
> http://www.cliki.net/ - Link farm for free CL-on-Unix resources
"Greg" <·@G.com> writes:
> This URL leads to a more succinct, application of copyright, too newsgroups.
> http://help.prodigy.net/help/newsgroup_bb/newscopyright.html
> Note the bullet just after "Fair Use:" starting with "An example..."
>
> I wonder how copyright applies to Google's Data base, of past newsgroup
> posts?
If you're planning on making a decision to post or a decision to use a post,
pay a lawyer to answer this for you.
However, if you're just looking for an armchair understanding, I think
it's commonly understood that if a person writes to a venue, it is at
least understood that the venue itself is able to read the information,
that is, it is given some kind of limited license for the on-its-face use.
For example, if I write you a letter, I retain the copyright holder, but
you certainly have a right to have a copy of the letter I wrote you.
But you don't have a right to take my letters and publish them for profit.
Of course, reprinting text back and forth among messages within the forum
is again part of the process and doesn't "leave the venue", so I think
does count as "fair use" in that it's the intended way to create a sense
of conversational context within this forum.
Indeed, if I sell you a painting, that doesn't give you a right to reproduce
that painting. You have to buy that right from me separately!
To the extent that Google presents the information as a simple
historical archive of this newgroup, I suspect it's in mostly safe
waters, although this surprised me. When it first came out, and
perhaps still, one could argue that posts which preceded its (and
DejaNews's) existence were made with an expectation that obtaining
archives was hard to get. Now that's not really so. I think most
newsgroup posters publish in the knowledge that Google will be there,
and so are reasonably giving permission to have archives made
available for people who search there. In a sense, Google is "being"
the forum (just "interactively") in that it is respecting the original
venue structure, thread structure, presentation style, etc. (Whether
that means meta-search engines are allowed to find them, I'm not so
sure... And, as to the future, there are various advances in data
query that I can imagine making me want to reconsider some of this.)
Knowledge and ideas cannot be copyrighted, so you can use the ideas in a
post. However, the actual text is something that, as a strict matter
of law, you'd have to be careful in using.
HOWEVER, the form of expression is copyrightable, and if you copy the form
of my (or anyone's) expression for another venue, you're at risk.
FindLaw.com (an excellent resource in its own right) summarizes of the
court opinion on President Gerald Ford's Memoirs,
http://library.lp.findlaw.com/articles/file/00102/006976/title/Subject/topic/Intellectual%20Property%20Law_Copyright/filename/intellectualpropertylaw_1_233
In the summary, they note the court did the standard four-part fair use
analysis (e.g., see www.fairuse.com) and they [Findlaw] highlight this issue:
4. Market Effect - The Court stated that the market effect "is
undoubtedly the most important element of fair use." In analyzing
this factor the Court concluded that "[r]arely will a case of
copyright infringement present such clear-cut evidence of damage",
and that any inquiry into this factor must take into account any
damage to the original work as well as to any "harm to the market
for derivative works." Needless to say, the Court concluded that
this factor weighed against a finding of fair use.
Personally, I'd feel more comfortable using a suggested snippet of code
AS CODE than I would using it AS EXPLANATION, for example. This is for
several reasons: (1) your using my suggested code in your examples is
unlikely to undercut my ability to sell a book about that code, and (2)
there is often some hint in the text of a given post that at least the
person who is being spoken to is encouraged to use the code offered, or
something like it. HOWEVER, the use of the code AS EXPLANATION (i.e.,
with the person being spoken to assembling a cookbook) or as a library
(i.e, the person being spoken to assembling a library to share with others)
is both not the apparent reason for which the information is offered AND
is reasonably possible for a thinking person to realize might infringe the
economic prospects of the use of the text, and so tends to weigh negatively
in point 4 ("Market Effect") in the Fair Use Criteria. Note that this is
so EVEN IF the person assembling the stuff offers it free (because the
ill effects are not just "what the infringer made in money" but "what the
infringed was unable to make").
I and others here on this newsgroup often use this forum to test out
our presentation of ideas for formal publication. Were someone to be
picking up our writings and collecting/publishing them without
permission, that would undercut our ability to derive later economic
benefit from our own works. If you have done this or have any future
plans to do this, you should know that I believe this requires my
permission except on works that I have marked otherwise.
Additionally, independent of what a court might decide, if I and
certain other prominent posters here found that people were going to
be trying to collect our words in any organized way (cookbooks,
libraries, etc.) rather than writing their own text/code, I/we would
personally consider such actions HIGHLY antisocial. I/we would stop
feeling free to share our thoughts and would likely greatly curtail
posts or perhaps leave the group entirely. (I use "I/we" because I'm
typing for myself now, but I am fairly confident that there is a group
who would agree with me on this.)
On 16 Jun 2003 18:05:26 -0400, Kent M Pitman <······@world.std.com> wrote:
> "Greg" <·@G.com> writes:
>
[...]
> > I wonder how copyright applies to Google's Data base, of past newsgroup
> > posts?
[...]
> However, if you're just looking for an armchair understanding, I think
> it's commonly understood that if a person writes to a venue, it is at
> least understood that the venue itself is able to read the information,
Google gives the option of not archiving articles. The user has to add a
special header field, X-No-Archive or something similar, to such articles.
Paolo
--
Paolo Amoroso <·······@mclink.it>
hallo
Now in the thread obove there comes up the idea to make a compact short
form of the knowledge inside this newsgroup.
example from Jeff Caldwell (reference3 in my newsreader)
"It would be nice if there were a searchable repository where it was
easy for people to post everything from little snippets, to medium-sized
routines, to larger working units.
It is hard to comb through c.l.l archives to pull out just the code
snippets, especially by category. There are times when that is exactly
what is desired."
This is basicly truth. And who did not click around to find that
posting? It would be fine, to have a public place, where to store
important or helpfull text snippets.
There are some ways to do so:
Someone can sit down and sort the postes by hand. I guess this needs
some months. Doing so, I ask about copyrights (to protect that someone).
Someone can snip relevant parts and put them sorted in a nice html frame
set (maybe with a javascript search routine). This can be a great help
for every lisp user or newbie. (extra stands the paradise for optimizing
code in Lisp)
There is available software that can reduce digital texts up to 30%.
Other programms (the data miner clones) will be able to return nearly
everything you ordered, but most need the hands of a specialist.
The problem seems clear. Through the years this newsgroup collect lisp
specific informations. One can analyse the questions over the years, the
point where things will repeated is near and the faq is big (and not the
right place for everything). How would this look in a few years?
And sure, some postings are realy good, means worth to store them
somewhere for quick access. Also thought about the writers, how will
they do their posts, when they know they become famous or historic?
(what is the right word in english?). So my thoughts circles around...
but thanks for writing your opinions.
stefan
(gui)
> restart
"Greg" <·@G.com> writes:
> This URL leads to a more succinct, application of copyright, too newsgroups.
> http://help.prodigy.net/help/newsgroup_bb/newscopyright.html
> Note the bullet just after "Fair Use:" starting with "An example..."
Yes. "short excerpt", "not claiming ownership", "not using
commercially" and "does not diminish the value". That's a _long_ way
from public domain.
-dan
--
http://www.cliki.net/ - Link farm for free CL-on-Unix resources