Mythical New Back Links Implementation

Twic I currently (23/3/2003) uses a fairly dumb implementation of the Back Links function. Apart from anything else, this Back Links Implementation is very, very slow. Tom Anderson has been talking about writing a new, faster, one for as long as anyone can remember, and, when discussing the management of the wiki, will often make statements like "when the new Back Links Implementation arrives ..." as if they meant "tomorrow and for ever after ...", when in truth they probably mean something more like "at some point in the distant future ...".

For the record, i now plan to do some major refactoring of the code before i finally build the Mythical New Back Links Implementation; at present, everything is in one gargantuan module file, which is way too big to fit into my head all at once, and is thus very hard to work on. I will split it into smaller modules (one for playing with files, one for core logic, one for making HTML, etc), so that i can understand it again. Then, then, i'll actually get round to it. I expect to have time to do this in early to mid March 2003; all my round tuits are currently being sucked into two major time vortices, which should die down around then. -- TA

Oh, and the reasons the Mythical New Back Links Implementation is such a big bit of work are (a) that it touches a lot of parts of the code at once, (b) requires one major bit of the code to work in a very subtly different way which is nonetheless hard to achieve and (c) will require some cunning in rolling out, as a few thousand link files will need to be created (requiring a complete parse of every page file) in the time between deploying the new code and anyone editing or asking for backlinks. -- TA

Sun, 23 Mar 2003 22:10:50 GMT Front Page Recent Changes Message Of The Day