On Fri, 23 May 2003, Matthew A. Schneider wrote: (01)
> Nice work on the transclusions. (02)
Thank you. (03)
> I've been toying with transclusions for sometime myself. I've been working
> on a method to easily scrape content from any given Web page - on the fly
> of course. This work is related to ideas I have for quotation in any part.
> Until everyone is publishing XML, it seems duct tape and baling wire will
> have to do. (04)
Yes, it's frustrating to have things like Xpointer looming on the
horizon but pretty much unavailable in any real sense. My
solution screen scrapes using the perl LWP module. I suspect what
I have to do is considerably more easy to do since I'm pulling
from pages that have the purple numbers already parsed in. (05)
> This link,
>
>http://www.sasites.com/suse/apache/files/psScrape.php?url=http://www.burningchrome.com:8000/~cdent/wiki.cgi?FoodTest&line=13,
> will scrape a paragraph off your FoodTest page. Change your paragraph and
> my scrape updates (upon refresh). (06)
Cool, nice work. (07)
Sometime tonight I'm hoping to push out an early draft of a why
purple numbers are cool document. I'm going to be using samples
from both PurpleWiki and PurpleSlurple. (08)
One thing that is becoming clear as I work on the document is the
degree to which "well, it may be not quite right, but it works
now" plays a part in all this. It is true that it is not quite
right, but if it is possible to demonstrate the value of the
tasks eased by the tool, then the tool will eventually improve,
or be replaced by the right stuff. (09)
Another thing that's clear is that when Xpointer does come, there
will need to be something like the link handles that
PurpleNumbers provide. They provide a strong sense of "you can do
something here" which is not strictly necessary with Xpointer but
without which, how will you know what to do? (010)
> It should be fairly simple to transclude/include this. In fact, when I was
> looking at the FoodTest page this evening, it occurred to me that this
> could be achieved by letting the Wiki parse a new tag (just as you proposed
> for pointing URLs in PurpleWiki to PurpleSlurple; i.e.,
> [P:http://www.burningchrome.com/ burningchrome through purple slurple],
> which by the way I thing is a GREAT idea :) I suspect you are doing this or
> something similar for your transclusions? (011)
The transclusions are parsed wiki tags that look like this: (012)
[t<NID>] (013)
where <NID> is replaced by the NID that is being transcluded.
This opens up some serious usability issues: it's hard to get
that NID, especially with the newer long, obscure and obscured
NIDs that I've been using. (014)
--
Chris Dent
cdent@blueoxen.org (015)
--
This message is archived at: (016)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=Pine.LNX.4.30.0305252130110.30013-100000@hot.burningchrome.com (017)
|