From: LIN8080 <li...@fr...> - 2003-05-22 05:17:04
|
Hallo In Impnotes.html - 1 564 205 - version 2.30.1 with: last modified 2003-05-15 I do some layout changes. The important are: - strip all span-tags, all div-tags and all class-ids in case one download this file for off-line browsing the *.css file gets lost. Therefor this tags make no sence (until they where included in the head-area of this file). - changed all Hyper-Spec online-links to local hard drive c:\hypers~1\... in most cases when one looks inside this file, this happens off-line. The number of Hyper-Spec links is massive and often one needs to consult Hyper-Spec pages for more infos. - file-size is now 849 822 bytes - other modifications change the table-layout to an uniform look replaced some dl, dt, dd-tags to ul, li tags (the rest will follow) replaced all " to " &(qoute); (see "-//W3C//DTD HTML 4.01//EN") in source set all links in a new line removed all h1, h2, h3... tags. Inside this file all is important, except some big nice headlines, hm? .... - else you can download the modified file as impmod2.zip - 160 321 bytes at (hope it works) http://people.freenet.de/LIN8080/Impmodi2.zip done is partI this ends with Chapter 27. In partII -Extensions- no table is visible because the classids are removed. Planed is a version with less tables (this will speed up load-time) - Please send comments stefan |
From: Hoehle, Joerg-C. <Joe...@t-...> - 2003-05-22 07:39:31
|
Hi, LIN8080 wrote: >In Impnotes.html - 1 564 205 - version 2.30.1 >with: last modified 2003-05-15 >I do some layout changes. The important are: Uh oh. This seems quite unfortunate to me since impnotes.HTML is not the place to work on. The HTML is automatically generated from a impnotes.XML (Docbook XML). So instead of improving mechanically generated HTML, I believe the only viable way is to change the transformation of XML (Docbook) into HTML. Sam already plays with such transformations, since CLISP typically comes with one huge impnotes file, while clisp.cons.org also has a split version (every chapter in a separate file), which is more user-friendly for external reference (URLs). Comparing various impnotes.html files (clisp-2.27, 2.28 etc.), you'll see quite some layout changes Sam did. See the CVS online repository http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/clisp/clisp/doc/ I fear that you wasted some time working on .html. You'll probably have to learn about DSL, e.g. see doc/impnotes.dsl, which influences how impnotes.XML gets translated into impnotes.HTML, or you propose another way of turning docbook XML into html. I never looked into dsl myself. I still have to learn docbook (and still am no friend, to say the least, of the verbosity of XML and thus docbook). Furthermore, impnotes.html from clisp 2.30 or 2.30.1 is almost outdated as soon as it appears. The up-to-date source is always in CVS. But Sam certainly accepts patches to released versions of CLISP (even though he always begs for patches against CVS, at least for source code). >- file-size is now 849 822 bytes I'm pretty sure there's room for improvement. E.g. impnotes for 2.27 was roughly this size, while 2.28 was >1MB. Some content was added, but also some of the HTML generated differently. My personal opinion is also that there's a bit too much of cross-linking to CLHS. Every time e.g. CHARACTER appears in text, I see no need to have that link to CLHS. When I browse impnotes, I always have to check the status line to see "is this a link to another interesting section or is this yet another boring link to CLHS?". >- strip all span-tags, all div-tags and all class-ids > in case one download this file for off-line browsing the *.css file >gets lost. Sam added .css file to the distribution (or put it in all clisp.cons directories where it was needed, I don't remember exactly) when I reported that problem here. So people can either o download this .css file from where they found impnotes.html o disable CSS (works in Netscape for me). >- changed all Hyper-Spec online-links to local hard drive >c:\hypers~1\... That really needs customization :-) This is done at install time for those who build CLISP from source. Maybe there should be a small lisp program to do this at any time. (defun change-impnotes-clhs-references (impnotes-file clhs-url) #) Regards, Jorg Hohle. |
From: Sam S. <sd...@gn...> - 2003-05-22 14:06:10
|
> * In message <9F8582E37B2EE5498E76392AEDDCD3FE04700420@G8PQD.blf01.telekom.de> > * On the subject of "[clisp-list] working on Implementations Notes layout" > * Sent on Thu, 22 May 2003 09:39:21 +0200 > * Honorable "Hoehle, Joerg-Cyril" <Joe...@t-...> writes: > > I fear that you wasted some time working on .html. You'll probably > have to learn about DSL, e.g. see doc/impnotes.dsl, which influences > how impnotes.XML gets translated into impnotes.HTML, or you propose > another way of turning docbook XML into html. I never looked into dsl > myself. forget DSSSL. XSL is the way to go. Note that impnotes in <http://cvs2.cons.org/ftp-area/clisp/snapshots/impnotes/> and <http://cvs2.cons.org/ftp-area/clisp/snapshots/impnotes.html> are generated with XSL while the released version <http://clisp.cons.org/impnotes.html> and <http://clisp.cons.org/impnotes/> are generated using DSSSL. the framework is not yet in the CVS because I am also trying to get the CLISP man page into Docbook/XML so that both groff and html version would be generated from the same XML sources. > I still have to learn docbook (and still am no friend, to say the > least, of the verbosity of XML and thus docbook). who is?! > My personal opinion is also that there's a bit too much of > cross-linking to CLHS. Every time e.g. CHARACTER appears in text, I > see no need to have that link to CLHS. think of the word CHARACTER as an object, with many properties (font, bg, fg, link &c). An object has the same properties everywhere. > When I browse impnotes, I always have to check the status line to see > "is this a link to another interesting section or is this yet another > boring link to CLHS?". When you read impnotes "from cover to cover", you can get used to the extensive linking. When you use it as a reference, you appreciate that every term in the fragment you are interested in has an appropriate link. > >- changed all Hyper-Spec online-links to local hard drive > >c:\hypers~1\... > That really needs customization :-) > This is done at install time for those who build CLISP from source. > Maybe there should be a small lisp program to do this at any time. > (defun change-impnotes-clhs-references (impnotes-file clhs-url) #) 1. change the entity &clhs; in cl-ent.xml to point to your favorite CLHS location. 2. regenerate clhs-ent.xml from the index file at that location using CLOCC/CLLIB/clhs.lisp (<http://www.podval.org/~sds/data/cllib.html>) 3. edit references in cl-ent.xml to &clhs; by hand. -- Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> nobody's life, liberty or property are safe while the legislature is in session |
From: LIN8080 <li...@fr...> - 2003-05-25 22:13:35
|
"Hoehle, Joerg-Cyril" schrieb: > So instead of improving mechanically generated HTML, I believe the only viable way is to change the transformation of XML (Docbook) into HTML. ... > I fear that you wasted some time working on .html. You'll probably have to learn about DSL, e.g. see doc/impnotes.dsl, which influences how impnotes.XML gets translated into impnotes.HTML, or you propose another way of turning docbook XML into html. I never looked into dsl myself. I still have to learn docbook (and still am no friend, to say the least, of the verbosity of XML and thus docbook). Aha. Now xml is like ??? for me, so I began to look around. First I read some docs and this leads me to some questions. ie: why using xml when sitting on an interpreter? or is it possible to use the xml component svg (scalable vector graphics) and draw some lines via a browser with cl? or what needs more time: to learn xml and do a workthrough in docbooks or replace some details with a nice editor? hm. Next I found some older zip files (01.03.00) named db152, db3x317, dbx106, docbk31 and the imp-xml. Extracting this and take a closer look - hey is that also generated stuff? (it looks like). Anyway. I think I finished what I start (work inside the impnotes.html). Meanwhile I have a look to learn something about creating dtds. > Sam added .css file to the distribution (or put it in all clisp.cons directories where it was needed, I don't remember exactly) when I reported that problem here. So people can either > o download this .css file from where they found impnotes.html > o disable CSS (works in Netscape for me). Ok. Thanks, found 2 css files. I write 3 lines into the head-section of the html. Nice: I set all link text to size:11pt; (default is 12pt) and text-decoration:none;. This removes the line under the link and the mostly capital link-letters are a bit smaller (all links have different color, thats enough). Now it is better readable (for me). > (defun change-impnotes-clhs-references (impnotes-file clhs-url) #) Great. Thats what I keep in mind. ---- Please check out: in section 1.4 in the overview ... 6 symbols missing ...(out of 978...) is this number still ok? Or remove the number, otherwise update it every time a new symbol is added. on some places there is -td align="left"-, this is a default value and can be removed. (maybe some exotic browsers need this?) nearly 20 times there is the string '::=='. This is hardly visible in the defaul font (Arial?). Also the ':' is not clearly in the text. For my case I changed to the courier font, this looks ok (but there seems to be a bug in Netscape 4.x: after every table-tag with -width="xx"- the font-rules are gone and default is used. Please check this out) stefan |
From: Hoehle, Joerg-C. <Joe...@t-...> - 2003-05-22 16:20:06
|
Hi, >Von: Sam Steingold [mailto:sd...@gn...] >forget DSSSL. >XSL is the way to go. Oh boy, yet another verbose ugliness! I recently browsed the book "Web-publishing with LaTeX" or some such from Michael Goossens and Sebastian Rahtz. It had complete examples of XSL for bibtex: Boy, what an unreadable nightmare compared to the clean, concise and powerful SXML pre-post-order transformers that I once wrote in Scheme using Oleg's tiny code. Please look up http://pobox.com/~oleg/ftp/Scheme/SXML.html SXML transformers could IMHO be ported to CL in an afternoon. XML parsers in CL seem to be lurking around cliki. BTW, I used that to generate validated HTML from validatable templates, using Kawa (a Scheme environment & compiler to JVM bytecode). Regards, Jorg Hohle. |
From: Sam S. <sd...@gn...> - 2003-05-22 17:49:04
|
> * In message <9F8582E37B2EE5498E76392AEDDCD3FE04700614@G8PQD.blf01.telekom.de> > * On the subject of "[clisp-list] working on Implementations Notes layout" > * Sent on Thu, 22 May 2003 18:19:52 +0200 > * Honorable "Hoehle, Joerg-Cyril" <Joe...@t-...> writes: > > >Von: Sam Steingold [mailto:sd...@gn...] > >forget DSSSL. > >XSL is the way to go. > > Oh boy, yet another verbose ugliness! yep, big time! I want to use XSL and DocBook/XML not because I like these abominations, but because they are standard. The same reason applies to why I want to switch from *.d to *.c in the CLISP sources. We need to lower the barrier to entry, which means using the standard tools whenever possible. Can we concoct a beautiful concise format from which we could generate html, pdf, ps, tex, groff &c? Yes! Should we? NO! > XML parsers in CL seem to be lurking around cliki. I wrote one (CLOCC/CLLIB/xml.lisp) -- Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> Winword 6.0 UNinstall: Not enough disk space to uninstall WinWord |
From: Michael K. <kr...@co...> - 2003-05-22 20:38:43
|
Moin Sam Steingold, > > Oh boy, yet another verbose ugliness! > yep, big time! I did a lot of GML,SGML and XML but my home pages are now driven by CGI::pWiki, because its much more natural to write the content and let the computer worry about how to understand the layout. what about using Perldoc or pWiki like documentation. So we have sparse coding, and readable documentation. Take a look at some of my URLs with a ?plain as extension to show the plain file on the disk. e.g. on [1] on text, lists, wiki links, tables and even ascii art. you might run into =location entries, when asking for ?plain, just follow them. CGI::pWiki is implemented as a CGI to implement an apache handler for *.html, so if the browser calls for *.html there might be a =location or purl redirection. clispWiki should be written in CLisp, and be a part of the source as it has to process the documentation into html and/or latex. It should be a subset of CLiki, to batch process CLiki pages. We would than just need to convince the admin of CLiki for CVS upload areas, and we can form a clisp documentation coding standard, by a rough consense of running code, that is part of standard clisp source tree ;) > I want to use XSL and DocBook/XML not because I like these > abominations, but because they are standard. they are recommendations not standards. UN/EDIFACT would be a document standard, but better dont think about using it for documentations (; > We need to lower the barrier to entry, which means using the standard > tools whenever possible. I often had the fealing that standards and like UN/EDIFACT or ISO 9001 are used to raise the barrier to entry ;) > Should we? > NO! i can hardly agree - but i wont find time to write a clispWiki right now ;( sorry for wasting bandwidth with vaporware - but you are invited to cut'n'paste the regexp's of CGI::pWiki in any other GPLed code ;) > > XML parsers in CL seem to be lurking around cliki. > I wrote one (CLOCC/CLLIB/xml.lisp) *aehm* whats the base URL ? peace on your way, Michael [1] : http://open360.copyleft.de/OS360/OS_360.html?plain (intro text, lists and wiki links) or http://open360.copyleft.de/OS360/IPL_into_MVT.html?plain (typical PerlDoc like documentation) or http://www.clamv.iu-bremen.de/CLAMV/TeachingLab.html?plain (subtitles, tables and text) or http://rom.copyleft.de/Maps/Southern_Part_of_Midgaard.html?plain (tables and ascii art) -- mailto:kr...@co... UNA:+.? 'CED+2+:::Linux:2.4.18'UNZ+1' http://www.xml-edifact.org/ CETERUM CENSEO WINDOWS ESSE DELENDAM |
From: Sam S. <sd...@gn...> - 2003-05-22 20:45:13
|
> * In message <200...@ba...> > * On the subject of "Re: [clisp-list] working on Implementations Notes layout" > * Sent on Thu, 22 May 2003 23:19:20 +0200 > * Honorable Michael Koehne <kr...@co...> writes: > > clispWiki should be written in CLisp, and be a part of the source as > it has to process the documentation into html and/or latex. It > should be a subset of CLiki, to batch process CLiki pages. We would > than just need to convince the admin of CLiki for CVS upload areas, > and we can form a clisp documentation coding standard, by a rough > consense of running code, that is part of standard clisp source tree > ;) go ahead and write that. > > > XML parsers in CL seem to be lurking around cliki. > > I wrote one (CLOCC/CLLIB/xml.lisp) > *aehm* whats the base URL ? <http://clocc.sf.net/> <http://www.podval.org/~sds/data/cllib.html> > peace on your way, I guess I missed it. -- Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> Even Windows doesn't suck, when you use Common Lisp |
From: Nikodemus S. <tsi...@cc...> - 2003-05-23 12:33:19
|
On 22 May 2003, Sam Steingold wrote: > I want to use XSL and DocBook/XML not because I like these > abominations, but because they are standard. > > We need to lower the barrier to entry, which means using the standard > tools whenever possible. Just a datapoint: I had some notes lying about clarifying the Impnotes about modules and FFI, but never got around to doing it, because I never found the energy to learn DocBook. IMHO DocBook is not exactly a barrier-lowerer, unless you happen to know it already. Cheers, -- Nikodemus |
From: Sam S. <sd...@gn...> - 2003-05-23 14:23:26
|
> * In message <Pin...@ke...> > * On the subject of "Re: [clisp-list] working on Implementations Notes layout" > * Sent on Fri, 23 May 2003 15:32:47 +0300 (EEST) > * Honorable Nikodemus Siivola <tsi...@cc...> writes: > > Just a datapoint: I had some notes lying about clarifying the Impnotes > about modules and FFI, but never got around to doing it, because I never > found the energy to learn DocBook. would you rather learn an entirely new, CLISP-specific, syntax and semantics (instead of a fairly well known XML syntax and a rather widely used DocBook semantics)? > IMHO DocBook is not exactly a barrier-lowerer, unless you happen to > know it already. many people know and use it already and more are willing to learn it because of that. -- Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> Your mouse pad is incompatible with MS Windows - your HD will be reformatted. |
From: Nikodemus S. <tsi...@cc...> - 2003-05-23 15:05:43
|
On 23 May 2003, Sam Steingold wrote: > would you rather learn an entirely new, CLISP-specific, syntax and > semantics (instead of a fairly well known XML syntax and a rather widely > used DocBook semantics)? Nope, you're quite right there. Ideally I'd like a format I can edit with a reasonable degree of confidence by "just guessing" and looking at the existing text. DocBook *almost* does that, but not quite. But as I said, that was just a datapoint. The lowest barrier of entry would probably be provided by plaintext, I guess -- but that's not really an option... Cheers, -- Nikodemus |
From: Sam S. <sd...@gn...> - 2003-05-23 16:15:06
|
> * In message <Pin...@ke...> > * On the subject of "Re: [clisp-list] working on Implementations Notes layout" > * Sent on Fri, 23 May 2003 17:34:03 +0300 (EEST) > * Honorable Nikodemus Siivola <tsi...@cc...> writes: > > The lowest barrier of entry would probably be provided by plaintext, I > guess -- but that's not really an option... It is. if your change is _small_ __AND__ _important_, you can just send a plain text version and I will add markup myself. -- Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> If brute force does not work, you are not using enough. |