Support for Google Sitemaps

[Martijn Pieters <mj@…>]

Google can index websites more efficiently and more frequently when a site provides information about the site in the form of a Google Sitemap; a centralized description of site content including information about change frequency and last update time.

Trac already has such information at the ready; all it needs to provide is the XML file itself and perhaps offer a few other niceties such as a sitemap index and automatic resubmission when pages change.

Trac could be extended to provide the following features:

Sitemap

 http://www.google.com/webmasters/sitemaps/docs/en/protocol.html

The sitemap lists every wiki page, roadmap milestone, changeset and ticket, with last-modification date. Perhaps the homepage of a trac project can be marked with a higher priority, and changesets can be marked with a changefreq tag set to 'never', while the timeline could marked as changing 'always'.

The spec leaves it open an what URL to present the map; it needs to be registered explicitly with Google the first time. However, it may be that in the future the file will be autodiscovered much like robots.txt is; by trying the /sitemap.gz URL. As such Trac should probably offer the sitemap under that URL; gzipping the generated content is easy enough.

Sitemaps Index

 http://www.google.com/webmasters/sitemaps/docs/en/protocol.html#sitemapLocation

In a multiproject configuration using TracModPython the ModPythonHandler could provide for a /sitemaps.gz URL that lists the sitemaps for all the Trac projects it knows about using a sitemapindex document.

Automatic resubmission

 http://www.google.com/webmasters/sitemaps/docs/en/submit.html#ping

Google allows for automatic notification of changes to sitemaps using a very simple HTTP GET API. Trac could automatically call this API whenever it learns of a change (wiki edits, ticket activity, etc.). This should be configurable, defaulting to 'off' as sitemaps have to be registered explicitly first (although there is no penalty for calling the API without having registered).

[Pedro Algarvio <ufs@…>]

I was going to ask about robots.txt, and then I found this ticket.

Guess this is the way to go.

I'm listening.

[Koen Werdler <werdlerk@…>]

This plugin made by  datenimperator creates a sitemap.xml:  http://www.trac-hacks.org/wiki/GoogleSitemapPlugin

I'm not sure if this is the most recent version since the developer has his own repository at his own website -  http://voidblog.de/wiki/GoogleSitemap .

[Koen Werdler <werdlerk@…>]

Correction; the links on trac-hacks.org already point to his own site.