2015-02-15 0:38 GMT+01:00 Vladimir Panteleev via Digitalmars-d < digitalmars-d@puremagic.com>:
> On Saturday, 14 February 2015 at 23:06:20 UTC, Mathias Lang wrote: > >> We'll get the best of both >> world, keeping the possibility to output any target out of the specs, but >> writting website-specific part in a proper framework. >> > > And the advantages of this are..? > > Sorry, but I don't think this is going to fly. > > You are proposing switching parts of the the website to Diet templates. > This is one more DSL that contributors will need to know. We will have two > different macro systems for building one website, and a ton of dependencies > to boot. > > Furthermore, this is probably the worst possible moment for such a > proposal - you are suggesting that we move even more weight on something > that is, right now, not merely unmaintained, but broken, and with no one > wanting to fix it. Even if you complete your work, and assuming it gets > accepted, who is going to maintain it? > > Right now, it would take a very strong argument to prove that we need to > migrate MORE parts to Vibe when the existing ones barely hold up. > One of the point of the vision document was to emphasize Vibe.d, and I'm trying to do just that. The advantages, from the top of my head, are: - Dogfooding: we have a web framework, and we're using it for our website. What message does it send ? - Simplified build process: on the long run we'll be able to reduce, if not ditch, the makefiles, and get the site up and running with a simple command, no matter what platform you're on. - Feedback on both Vibe.d and dub. - More capabilities in the website: currently, anything dynamic has to be done as a separate part (i.e. the forums), or in php (bugstats). In the future such work could be integrated directly into the website. It could also be made more dynamic, which opens a lot of possibilities, from simple improvement (detecting the user's system and pushing the proper download link to the top of the list), to more advanced one (providing Martin Nowak's REPL, or better downloads / site consultation information). A good point with this proposal is that it's not distruptive to the existent: DDOC allows us to make this change smoothly, so if the change gets pulled, it won't change anything to you, or anyone else that doesn't want to use it. I expect some resistance, but I believe using it will be the best way to prove it's usefulness.