[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Request: Ticker-2-Mail



On Mon, Jul 12, 1999 at 12:53:07AM +0200, Colin Marquardt wrote:
> * Craig Sanders <cas@taz.net.au> writes:
> 
> > ideally i'd like to convince the freshmeat guys that it would be
> > cool to be able to subscribe to a daily machine-parsable (XML
> > perhaps, or just plain old key=val with \n\n to separate records)
> > update which could then be fed into your own 'localmeat' :-)
> > database.
>
> Uhmm, do you mean something different than the current newsletter?

yep. i mean a format that is specifically designed to be unambiguously
parseable, and without any layout/formatting cruft.

(i.e. this is the same reason why unix commands have such terse output
- so you don't have to trim off verbose cruft when you want to pipe it
into another program)

> One example article is (the | are added by me, of course):

that example shows the most common format, but there are minor
variations. in particular, not every field is present in every article
and some longer fields (description, changes) occasionally consist of
multiple paragraphs. i hacked a work around for the multiple paragraphs
but i'm not happy with it - it's parsable, but only by guesswork not
specification.

variations mean extra code to handle the special-cases, which tends to
make the code more complicated - harder to read, harder to debug, harder
to extend. (this is partly technical, and partly aesthetic of course).

automated data exchange should be done by a defined protocol, not by
guesswork.

a machine readable format would have every field listed even if the
content was blank, and there would be a fool-proof way of identifying
continuation paragraphs (e.g. the way Description: fields are handled in
debian package control files). 

craig

--
craig sanders


Reply to: