Alan Burlison wrote: > Which web app, the ogb/arc one? of course - we're not talking about modifying your auth site...
> This whole scheme is way to dangerous as far as I'm concerned. Why? Seriously. If the site were ROR based, or perl or python, there would be no compile step. Once a schema is in place, most of the changes will be either cosmetic (css/html focus) or new feature additions. Being able to have the website "in" a hg repository means that anyone can clone it, develop locally, test locally (within the bounds of whatever auth sandbox you provide) and, when satisfied (and if authorized!), push back to the ARC- site-preintegration-gate with an RTI. If that gate were connected to, say "http://arc.os.o/test", the submitter could even test things out. Once the gatekeeper(s?) are happy, they could then push from the pre-gate into the live-gate, thus updating the real site. All without having to give everyone in the community root access to the server. Nope, this isn't enterprise datacenter/MIS level protocol, but in my mind, there is absolutely no need for that level of paranoia. Yet. Maybe never. Make it easy for people to contribute, make it easy to see results, and maybe that plocher guy won't have to do all the work :-) -John
