Re: Some thoughts on a new publication approach

On 29/10/2013 15:17, Simon Sapin wrote:
> Ideally, each spec would have a link to a web-browsable view of its
> source repository, where every revision ever published can be viewed and
> the files served with a proper Content-Type header.
>
> For example:
>
> https://dvcs.w3.org/hg/csswg/raw-file/db9b855680ef/css-syntax/Overview.html
>
> Richard, would that help?
>
>
> Admittedly, this needs to be more discoverable. Also, unfortunately,
> GitHub insist of sending everything as text/plain for its "raw" view.
> http://rawgithub.com/ helps, but we may want something more robust.

I don't really understand all the technical ins and outs, but this 
sounds promising.

I guess that, at the very least, I'd want editors to make regular 
hearbeat snapshots, that we could latch onto. (It's *much* more 
effective, btw, for the i18n group to review specs *before* they reach LC.)

If, however, there is some way to view all commits in a way that is 
easily discoverable, and permanently addressable with a easy to obtain 
URL, we may actually be able to do away with the need for heartbeat 
snapshots, and provide much more useful granularity for picking a review 
target or finding deleted text.

We'd certainly need an easy to locate list of commits for a given spec, 
in chronological order, with links that produce files with the proper 
Content-Type header and associated files. We get this out of the box 
with CVS or with wikis: would it be difficult with github?

(Hoping i'm making some sense)
RI

Received on Tuesday, 29 October 2013 16:58:00 UTC