Wed Sep 12 01:47:46 CEST 2007
parsing a ramblings file seems to be pretty straightforward. need to
fix some small organization problems for 'wikifying' my syntax. i
guess i've been able to avoid most parsing problems using lisp and
forth.. time to read some theory.
anyways. one important node for indexing is to make the article url
stable, such that if i edit the original ramblings file, all
references stay unless the article is deleted.
the only thing that's reliable is the date. since i'm using standard
'date' output, this can be parsed back into some simpler date rep.
just was watching
and one of the remarks david weinberger is making is that it is simply
too expensive to delete everything. he also mentioned some
'meta-wikipedia' where experts link to a subset of frozen articles
that they can endorse.
made me thing of these ramblings files. instead of cleaning them up, i
can just make a decent article here and there and archive the
lot. then if i want to link to some historic thought, i just
can.. that's why i do need a proper indexing system. the exact dates
are probably good enough. i don't even need a project name.
in fact, i can index everything i run into by standard date
encoding.. maybe that's the 'pool' i've been thinking about?