Diwaker Gupta wrote:
On Tuesday 24 January 2006 4:09 pm, David Crossley wrote:
The new http://projects.apache.org/ is now
available. Forrest is not yet listed.
Do we need to change our DOAP file [1]
or maybe transform into a different syntax?
[1] http://forrest.apache.org/doap.xml
Will it still serve the existing purpose
or do we need to maintain two files?
I had created the aforementioned file to facilitate indexing in O'Reilly's
CodeZoo [2]. The current doap.xml is not really a DOAP file -- its actually
an ATOM feed, that contains an embedded DOAP spec as an entry. My bad in not
naming it appropriately.
I've not looked at the final code for projects.a.o but the original code
acepted the DOAP as either a DOAP or an ATOM feed of a DOAP. I hope that
this feature has been kept.
So what we need to do is:
o create a proper DOAP file for Forrest. There's even a web-based form to help
creating/updating it [3]
May not be necessary, see above, but also see below.
There are also a number of RDF additions to the DOAP format in the file
requried by projects.a.o. So we need to add those, hopefully they have
been documented at projects.a.o (not had time to look yet). If not then
I guess a look over the httpd DOAP will show the full range of
additional info since David Reid (who wrote the scripts for this) is an
HTTPD dev.
o make sure the embedded DOAP in the Atom feed (rename the current doap.xml to
atom.xml?) remains in sync with the actual DOAP.
If it is necessary then we should add a pipeline to the projectinfo
plugin that creates the Atom feed from the DOAP file.
In fact, it may be a good idea to do it this way anyway, then we are
able to provide both formats. Maybe we should ewven think about
providing a pattern that strips out all the additional rdf info just in
case it messes with external applications (I don't think it wil though).
Ross
