[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [sacampbe@mercator.math.uwaterloo.ca (Sue Campbell)] Re: Timestamp



> > >    copyright
> > >    date it got into distribution
> > >    changelog
> 
> > The last three can be extracted by mechanical means.  The first three
> > can not (yet).
> 
> The original source location and can theoretically retrived from the
> copyright file. (policy 5.6) However, in real life we would need to
> standardisize it, if we want a script to extract the info...

I've got some scripts to extract the copyrights from the deb files en-mass
in my home directory on master (/debian/home/jim). 

> About the same amount of work would be to creat a new file (maybe called
> webinfo), that would include the extra info wanted for the web pages.
> (Homepage, upstream author, source retriaval point).
> 
> Webinfo should be included in /usr/doc, so that users can find the info on
> their own page and so that dwww can support the new file format.

dwww already uses .dwww-index and menu files.  dhelp uses .dhelp files.

I am working on a replacement format for these files.  It will be SGML/XML
based and I will come up with a DTD and some parsers/HTML generating tools
for that.  I want to define a standard format that any upstream author can
use.

I imagine there might be two sets of files installed, or even more - one 
would be from the upstream author, and one would be from the Debian
maintainer.  These would be parsed via an SGML parser (probably Jade), 
stored in a `grove', and stylesheets would be used to generate HTML pages.
Actually, I'd like to load in the /var/lib/dpkg/available and status files,
plus a bunch of other things (LSM data, RFC's, SGML files by "editors", etc)
into the `grove' as well.  It would be truly amazing what you would be able
to with stylesheets with all this structured data in place.  There are
also querying languages that can work against these `grove' things to
enable really, really smart search engines.

The dwww program itself would just be a couple of simple CGI scripts that
just maneuver into place the appropriate style sheets, converters, and
querying engines.

It's going to be a while however, since I'm a complete and total SGML
newbie -- and there might need to be some additional development on some
of the tools.

Cheers,

 - Jim


Attachment: pgpeWgDcJpVxO.pgp
Description: PGP signature


Reply to: