"Jonas Drewsen" <[email protected]> wrote in message news:[email protected]... > On 16/03/11 16.13, Regan Heath wrote: >> On Wed, 16 Mar 2011 01:13:33 -0000, jasonw <[email protected]> wrote: >> >>> One problem is the large amount of obsolete data ( >>> http://www.dsource.org/projects/dmdfe ) >>> >>> Dsource is The place for D projects. The problem with dsource is if >>> you're a serious professional and need professional quality libraries >>> and tools, dsource does nothing in the way of supporting these types >>> of users. The sections are filled with small hobby projects such as >>> http://www.dsource.org/projects/libcalc. What I'm looking for is >>> somehing that emphasizes the names of "important" projects. For >>> example standard parallel/concurrency/server/socket/vfs libraries are >>> a first class priority. It takes a day to browse through the list of >>> mediocre crap. >> >> I was browsing dsource the other day and I wanted to be able to sort >> projects by last update date or something, to find the ones which were >> being currently maintained. It would certainly be useful to sort by a >> category like [alpha] [beta] [stable] etc as well. I think dsource is >> the correct place to put any/all of our 'crap' but it just needs to be >> easier to sort and find the things you're interested in, at any one >> time. i.e. what if you were looking for a project to lend a hand to, no >> use finding one which is pretty much [stable] and complete. >> >> R > > An improvement that would really help a lot and probably be quite simple > to do is to let the dsource server serve the project listing as a cached > page. Currently it takes forever to load the project listing which only > changes once per month or so. >
It would be far faster even if it just generated the page on the server for every request. Currently, the page has JS that requests an XML DB of all the projects from the server, then the JS parses the XML and generates the list. A real mess. It's like posting a screenshot by taking a picutre with a camera phone, emailing it to youself, printing it, and then scanning it. The new upcoming version of the site doesn't do any of that mess, and works *much* faster (even on known-to-be-slow FF2), but who knows when it'll finally get up...
