Re: Web Standards version 2.3
Lars Wirzenius wrote:
>
> Bruce Perens:
> > "dwww" can get the name of the package from the directory, and the
> > description from whatever comes between <title> and </title> in the
> > Index.html
>
> That's an interesting idea. Minimum fuss for packages. I like it.
>
But in this way the package has anyway to provide a full index.html
file, and the title of a web page should be the shortest possible, thus
letting very little do say more than the name of the package.
With my proposal packages has to provide only a fragment of index,
leaving the freedom to add links to more structured html documents (a
specific index) in the directory.
We can set a limit to the fragment or, better, set up a policy and raise
bugs against verbose maintainers (like I am :-)
> A long index does need to be split into sections, but I guess
> that could be done based on which directory the package is in.
>
Even following the minimal structure that Bruce suggested, that index
will have hundreds of rows!
I suggest:
* to split the index into sections (or into subindexes)
options:
- let the package's maintainer choose (e.g. using the name of the
fragment) and putting all the packages that don't choose into a
"misc" or "others" section. A package could have more than one
fragment, for example a package that has documents in more than
one language can have its fragments installed into the proper
section of the index.
- Build an intermediate level based on the Section which the package
belongs is an actractive idea, but then we will have the most
visited documents under /auto/doc/doc/doc-linux !!
- we could let the doc packages in the same level with the other
sections (this increases the difficulty of the build)
* to leave the use of the Description first row from
/var/lib/dpkg/available only as a default for those packages that
don't have the index fragment, being those rows more a description
on how the package can be usefull, rather than explaining what's
inside the /usr/doc subdirs.
* to let a group of related packages (those that are splitted into
several binaries) installs a single entry in that index.
I know that these suggestions add complexity to the building of the
whole, but a well tested script can handle all this, letting enough
freedom to the packages to add or not the informations that they want,
without creating an unreadable monster page.
A long but well structured page is more readable than an half page
completely filled with only links.
Fabrizio
--
+-------------------------------------------------------------+
| e-mail: fpolacco@megabaud.fi fpolacco@debian.org |
| http://megabaud.fi/~fpolacco/ Join the UKI Linux Project! |
| fingerprint 70 1A 72 2D 2B C8 A5 63 7A C2 CC E0 2A 54 AE DA |
| finger: fpolacco@megabaud.fi fpolacco@master.debian.org |
+-------------------------------------------------------------+
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-REQUEST@lists.debian.org . Trouble? e-mail to Bruce@Pixar.com
Reply to: