I run my personal blog using blosxom. It works and I'm mostly happy with it. After a long period of stagnation there is some work on the tool, which is good.
I recently tried the new revised version and it didn't like my blog's "flavour" data. I had a look at the code and I didn't like it - well not enough to try and fix it.
I can keep my old Blosxom install which does work perfectly well. I could try something else (e.g. Bryar) - but there are not that many light-weight non database blogs about, or I could write my own.
I'm really tempted to write my own, not as a fork of Blosxom, rather a fresh application that can use the same filesystem data, similar templates and possibly a few extra features. My plan would be to create a OO module with a very simple wrapper app that could be a mod_Perl module or CGI script.
Re:i quite like chronicle
ajt on 2008-11-25T14:12:57
I know Steve but had forgotten Chronicle - I'll have to have another look.
Oddly enough I just rewrote blosxom last week. I got tired of patching because it didn't quite exactly match 100% what I wanted it to do. Considering the blog I use it on is static, with relatively few posts, it was easy enough to cobble up something that works just for me.
Incidently I (and you, apparently) followed exactly the process MJD described in Blosxom Sucks: first write plugins, then hack the code, then rewrite it.
In any case, I kinda like blosxom, except for the need to have so many different templates. One for the page and one for the story should be enough, the rest can be taken care of through CSS I believe. Also I never seem to have the right information available when I need it (the number of posts per section/tag for example), so a 2 pass process, first gather info then generate HTML, would be nicer I think.
Re:E tù...
ajt on 2008-11-25T16:21:00
I really like Blosxom's simple design but with simplicity comes the complexity of hacking it to make it work correctly... On my system I wrote my own "flavours" from scratch and hacked the RSS generator so that it creates RSS feeds with valid dates. Other than that I've not hacked it that much...
I'm currently playing with ikiwiki which is a functionally excellent wiki compiler and I'm wondering if the "compile" design is a good alternative for a blogging engine. You have entries on the filesystem in blosxom format, MarkDown or any other format there is a module that can parse it, this then generates static xhtml files in a per "type" tree and a per "date" tree. Static html serves much faster than dynamic and a cron job runs the compiler every now and then if any new entries have been written. You can even integrate it into a VCS as ikiwiki does.
Looking a ikiwiki's code also gives me goose bumps - it doesn't look very "Perl Best Practices" either...
Re:E tù...= encoding fun!
mir on 2008-11-25T17:37:35
I love the way blosxom is coded, with long lists of statements separated by 'and' or 'or'. That's hell to debug, but fun to read.
I am a big fan of the "compile" design. I lets me create (and check twice) posts on one machine, generate the blog, check some more, then rsync it to the live server. Of course it's easier when you don't want to handle comments...
Re:E tù...= encoding fun! - or not!
ajt on 2008-11-25T18:42:52
I like the compile design of ikiwiki which is what made me think of making my own blog compiler that used blosxom style files but generated static xhtml, which is then published via rsync/ssh. Like you I'm not interested in comments or a web form to create fancy content, I just want a simple to use blogging engine.
The original code is "compact" and while it's nice that some is looking after it, it's hardly idiomatic Perl and that's probably not a good idea on something like a once popular CGI blogging application.
