[tex-live] [LONG] Improving TeX package classification and the associated documentaion

[Florent Rougon]

Oliver Bandel <oliver at first.in-berlin.de> wrote:

> If there would be a simple tool that asks for the values
> for each key, and the user types in the text, the file could be created
> automatically, and no typos could destroy the validity of the
> data.
> Simple input-masks or text-based question-and-answer
> scripts should be written in a short time.
> If people were encouraged to use these tools, then this
> minimizes the syntax-error-problems drastically.

Just the same with XML. It can be easily generated by a program.
Syntax errors can come when the file is manually edited afterwards, but
maybe the new upload CTAN procedure could have checks for the file
syntax to minimize such problems, and reject the upload with a nice
explanatory message in case the syntax is invalid...

> More experienced users might use the texteditor and therefore
> would be the biggest source of problems, when doing typos, but a
> syntax-checker tool, or something that writes new entries to the
> local catalogue - and always should check the syntax - could help
> here also.

Yup.

> I have not read the RFC-2822 but the Linux-Software-Map (LSM)
> uses a simple format since a long time, and I see no reason
> why this should not be possible for TeX-stuff also:
>
>    http://lsm.execpc.com/LSM.README

Better stick to standards when they are adapted to your needs. Your LSM
stuff is modelled after RFC-2822, but slightly different (tolerates
empty lines):

,----[ http://lsm.execpc.com/LSM.README ]
| The general syntax is now similar to that of netnews headers, i.e., a
| keyword (at the beginning of a line), a colon (with no spaces separating
| it from the keyword), and text that may be continued to the next line by
| starting each continuing line with whitespace (space, tab).  Empty lines
| are also allowed (unlike netnews).
`----

There are modules in the Python standard library[1] to parse both XML
and RFC-2822-conformant files, but not "modified RFC-2822". Most
probably the same for Perl.

> BTW: there are not only XML-parsers out there, tehere are also tools
> like lex and yacc available for most languages, and so, a simple
> format could be parsed easily.

These tools are useful to parse custom languages, but I really don't
think our needs are so complex to justify their use.

  [1] i.e., they are shipped with Python itself.

-- 
Florent