On Mon, 16 Dec 2002, Garold (Gary) L. Johnson wrote: (01)
[purple numbers in email] (02)
> It appears that conversion for archiving is done after the message is saved
> for sending to the list. I am hoping to see the transformation done *before*
> the message is mailed. (03)
I've been meaning to jump into this conversation because I
noticed there seemed to be some confusion on this particular
point. (04)
Now that you've made it clear that you want the transformation
before delivery that raises a couple of questions: (05)
1. What are your expectations for what is delivered to the
participants?
1. A text/html message that is the original message translated?
2. A URL pointing to the transformed message at some web site?
3. The original message with a URL pointing to the transformed
message? (This is being done)
4. A multipart message combinging the original and 1 (06)
2. In what format do the participants compose?
1. Straight text
2. Quotable Printable text
3. HTML
4. Wiki formatted text
5. Accept some or all of the above and cope as well as
possible on the receiving end? (07)
In both 1 and 2 large constraints are placed on the participants
to meet requirements for composing and reading messages. (08)
Assume we decided that we were going to parse incoming text as
WikiText markup and send out messages as text/html WikiHTML
messages with purple numbers and store these in the archive. We'd
get the purple numbers in the messages but only be able to fully
_use_ them in the messages if using an HTML capable mail reader.
One that allowed the right click copy-shortcut trick that is so
handy. (09)
Would we parse everything as WikiText or only messages that made
some indication that that is what was wanted? If that indication
was not provided would the entire messasge be considered a nowiki
block and get only one purple number? (010)
The diversity of clients and people motivates the desire to
handle as much as possible on the backend in the archive, but
then how do you get people to use the archive? It's the same old
thing: people like the comfort of home; of their email client. (011)
To make an experiment like this one work in a true bootstrapping
sense we need to adjust our habits and see where that gets us.
What direction would be best? (012)
I think we need one that enforces, to some degree, proper
behavior. That is, we want to magnify opportunity in some areas
by limiting them in others. So, for example, we can force more
purpling by following one of two courses: (013)
- send out email messages that are just the archive URL and
instead of quoting, people reference purple numbers. This would
probably motivate eventual development of transcluding on the
mailing list server side.
- do WikiText parsing in the mailing list software, rejecting
messages that fail to parse, sending out WikiHTML messages and
storing those html messages in the archive (014)
(there are plenty of other examples, these two just stand out) (015)
Part of the code to do WikiText parsing on email messages already
exists. Intergrating it, especially in a way that handles errors
well, would be challenging but certainly not impossible. (016)
--
Chris Dent
cdent@blueoxen.org (017)
--
Archives: http://collab.blueoxen.net/forums/tools-yak/
Wiki: http://collab.blueoxen.net/ (018)
|