On 04/16/2010 09:36 PM, Bill Hart wrote: > On 16 April 2010 21:32, Antony Vennard <antony.venn...@gmail.com> wrote: >> >> >> On 04/16/2010 09:15 PM, Bill Hart wrote: >>> On 16 April 2010 20:54, Antony Vennard <antony.venn...@gmail.com> wrote: >>>> Alrighty, sounds good to me, just checking 'cause you've mentioned >>>> "off-list" support a lot... >>> >>> Yes, lots of "off-list" support, which I encourage people to put >>> "on-list" where possible please! >>> >> >> Excellent :D >> >>>> >>>> Yes, Django is my current favourite web framework. I looked at a lot of >>>> PHP frameworks but I couldn't get excited about them and most enforce >>>> MVC and strict url parsing: http://bsdnt.org/class/function/argument -> >>>> class{ function (args} } which isn't massively flexible. Django is happy >>>> either way. >>>> >>>> So I'd need a server with Python installed, preferably 2.6+. mod_wsgi is >>>> also the easiest way of integrating with apache which is what I've been >>>> doing - none of this fancy nginx/lighttpd stuff. Database can be >>>> anything supported - SQLite, MySQL, PostgreSQL. I've heard good stuff >>>> about the latter. >>>> >>> >>> Selmer has mod_wsgi, python 2.6.2, python_django, postgresql >>> installed. I've just sent you a username and password to log in. >> >> Thanks, I'll have a look and start building something up. >> >>> >>>> So that'd basically be the idea - to build a "lite" CMS using Django so >>>> anyone, not just web devs, can add say "news updates" or "version >>>> releases" and modify repository urls, contributor details etc WITHOUT >>>> digging through HTML. It won't be a fully fledged CMS - new content >>>> types will need someone to hack on Django. >>>> >>>> The beauty of this is we can build tools to suit. Trac looks pretty good >>>> and is also python but you could easily re-implement it. >>> >>> Hmm. Trac is pretty sophisticated. I'd be surprised if you could just >>> reimplement it. >>> >>> Also bear in mind I know nothing whatsoever about CMS's. I once used >>> Drupal and found it impossible to figure out. Website stuff is not my >>> thing at all. >> >> Probably not in one hit, but over time. The alternative would be to >> merge it somehow with Django. I've never looked at it from a source >> perspective, but I imagine there's all sorts of interesting combinations >> of trac+django. Their both being python means we can pull trac info into >> Django, too. I expect at least a level of compatibility and interoperation. >> >>> >>>> Name a tool and >>>> we can probably create it quickly enough. Upload a zip file? Submit a >>>> patch? Send a question to the mailing list? Add a sponsor? Add a test >>>> result? Publish a new test matrix? All done relatively easily. >>>> >>>> The rest would be "branding" via CSS and static media such as images, >>>> tarballs and whatever. >>>> >>>> So, I like django. I can however do PHP too if anybody really wants >>>> that. I've never used Ruby but I've heard good things about Rails and >>>> Sinatra. >>>> >>>> Thoughts? >>> >>> Let's keep it pretty simple for now. I personally "get" Ruby. But it >>> has not gained as much of a following as say Python. So I've not put >>> much time into learning it. Moreover, I know very little about rails. >>> I've never personally used PHP, but it is a great language from what I >>> know of it. I've never heard of Sinatra. >> >> Django is a framework on which you build web apps - you basically have a >> copy of its libraries in /usr/lib/python-xx/site-packages on the >> PYTHON_PATH and you create a few files that Django would like to be >> there (or else you tell it something different) and it all magically >> works. The PHP frameworks are conceptually similar in that you have a >> library of PHP classes which are "included" into your current site, >> although it doesn't work quite as cleanly as django, which is on the >> python path. >> >> Python path is comparable to path for executables, LD_LIBRARY_PATH or >> the Java Classpath. It's good. >> >> See djangoproject.com and there's a free "Django book" out there too, >> both of which are really good resources. >> >> CMSes themselves I have never been entirely happy with - there's always >> something not-quite-to-my-liking and customisation is always just out of >> reach. With Django, we don't need a full CMS product (like drupal), we >> can build just the dynamic content we like the idea of and code the rest >> in as html/css. The other thing I forgot to mention is templating - you >> only create one or two base templates and the rest inherit from that, >> minimising the amount of html you have to write which is a massive bonus >> as far as I'm concerned. >> >> I'm no web guru but I know enough to get by - to be honest though, most >> web dev is just a little bit dull. The real fun stuff is written in C... > > That's right, which is why I would say keep it simple for now. Better > to have your coding effort spent on writing great C code for us, > rather than reinventing trac. :-) > > Looking forward to seeing what emerges on the website front. What we > have for MPIR is currently a pain to maintain and I've never had the > time to learn anything other than html and css (the latter of which I > didn't use for the MPIR website). > > Bill. >
Right, I'll build something basic for now - probably minimally themed until we've agreed on stuff like branding and a name... maybe it'll actually stay that way, who knows. I'll write something simple and upload it to my staging server over the weekend/early next week so people can look, comment and complain then we have something simple to put live soon-ish. Primary goal is easy maintenance of whatever content is on the pages so that "releasing" etc doesn't take as much html editing as it does C code. Antony -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-de...@googlegroups.com. To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en.