On Tue, Jul 12, 2011 at 12:18, Richard Quadling <rquadl...@gmail.com> wrote: > On 12 July 2011 09:30, Hannes Magnusson <hannes.magnus...@gmail.com> wrote: >> On Thu, Jul 7, 2011 at 11:47, Richard Quadling <rquadl...@gmail.com> wrote: >>> On 7 July 2011 08:23, Hannes Magnusson <hannes.magnus...@gmail.com> wrote: >>>> What other options do we have? >>> >>> Adding role="xxx" to <refentry> where xxx is one of a predefined set >>> of values (class, interface, function, method, etc.), would seem to be >>> one way of going. >> >> >> The problem with that approach is that we only use role="noversion" >> for some refentries... >> >> -Hannes >> > > Would treating role="" for like you can treat class="" in HTML? > > i.e. > > role="noversion function" > > As I understand things, role is essentially free text and reader > interpreted. We can put any meta data in there that we want and it is > up to us to process it.
I suppose. OK. You have my vote to make the changes, doesn't look like any disagreed? We need to somehow document all these custom things we are doing at some point :] If at some point we decide to leave PhD (for whatever) it could become horribly difficult. -Hannes
