Re: [Monotone-devel] Documentation Texinfo XML to Wiki converter

[hendrik]

On Thu, Feb 26, 2009 at 02:45:03PM +0100, Philipp Gröschler wrote:
> Thomas Keller schrieb:
> > The whole xslt thing is basically a test balloon how well we can convert
> > the current texinfo source to wiki markup and if this works out well if
> > we use this instead of plain HTML files for the website (or even switch
> > completly to a wiki manual, if people agree). The reason for this step
> > is because makeinfo is rather limited when it comes to HTML output, you
> > cannot rearrange things or add certain needed header / footer parts
> > easily to it, _but_ there is an XML output available which theoretically
> > could be transformed to anything we want. And exactly this is Philipp's
> > work.
> 
> Thanks for clearing this out ;-) My last mail wasn't really what one
> would call creative or informational.
> 
> Again, if there are questions, please first take a look at the README
> file in the branch I committed. There I tried to sum up my intentions.
> Well, at least the important ones.
> 
> Zack Weinberg schrob:
> > What is *your* vision for this thing?  I don't really understand what
> > you have in mind so it's hard for me to evaluate.  I think we would
> > prefer to keep the master copy of the manual as part of the main
> > sources rather than opening it up to wiki-based editing, but that
> > doesn't mean support for more different rendering formats is
> > uninteresting -- far from it.
> 
> My vision is to get involved in the project in a way that I can help
> within my limited spare time/abilities.
> 
> The current state of the package is very experimental, just to show a
> way how it *could* be done. Besides the possibility of completely
> switching over from the Techinfo Manual to Wiki at some day, I rather
> thought of a continous one-way process. The documentation would still be
> maintained with Techinfo and new versions would be converted to Wiki
> format. The respective resulting parts of the Wiki should then be read
> only to prevent concurrent modifications which could not easily be
> brought back into the Techinfo part.

If text can be modified both on wiki and on original sources, conversion 
has to be a two-way process, probably moderated.

> But that is not my decision. I wanted to show a way how it could be
> done, and if there's reasonable demand then I will keep working on it.
> So much for the "vision" part ;-)
> 
> Daniel Carosone schrieb:
> > The great thing about this work is it (begins to) breaks the coupling
> > between purpose and style, which means content can be used for
> > multiple purposes regardless of style, in turn meaning that
> > "unification" doesn't get confused with "markup conversion".
> > 
> > So, really, yay, and yay again.
> 
> I carefully consider this as a positive response ;-)

I first encountered this distinction way back in the days that SGML was 
an experiment by the American Associatin of Publishers (AAP).  
Separating the author's job from the book designer's job was the whole 
point.  Since then there's been an explosion of variations on the 
original design, most of them completely contrary to the original 
purpose.  XML is one of the few that *is* in the right direction, though 
overly complex.

By the way, "AAP" is the Dutch word for "monkey" or "ape", yielding much 
mirth in Amsterdam, wher I was at the time.

> 
> Yes, XSLT is a nice way to break the border between plain markup and
> presentation. I learned to use it about five years ago and I was
> astounded by its possibilities. For quite a while I used it for a self
> written XML-based servlet engine.
> 
> But besides the magic one could do with XSLT there are also a lot of
> limitations. The way a stylesheet works and is processed could drive you
> easily into madness. At least it has done so a few times with me. And
> some things just do not work.
> 
> Example: I wanted to give the user the possibility to choose the output
> format converter at runtime. The first intended way was to get a command
> line parameter into a XSLT variable, which then contains the complete
> file name of the converter stylesheet and then use this variable with an
> include instruction. So to speak, for dynamically chosing the included
> converter stylesheet. That one went wrong. XSLT stylesheets are first
> compiled completely and then interpreted. There's no way for such
> dynamic and runtime dependent behaviour without external programed logic.

I forget.  Is XSLT the one that has a syntax resembling Scheme?

-- hendrik

> 
> 
> Ah, so much text. I think I should stop here.
> 
> Greetings!
> 
> Philipp
> 
> 
> _______________________________________________
> Monotone-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/monotone-devel