Subject:
|
// and ** vs {} and [] (was: testing in rtl...)
|
Newsgroups:
|
lugnet.publish, lugnet.admin.nntp
|
Date:
|
Fri, 30 May 2003 00:38:20 GMT
|
Viewed:
|
19 times
|
| |
|
|
In lugnet.publish, William R. Ward wrote:
> I agree that FTX should translate {} and [] to // and ** when displaying in
> plain text, and I'm *not* joking.
The problem with that is that articles are stored in the news server in their
raw original format only. When they're displayed by the web interface, and
the FTX content is rendered into HTML for viewing on a web browser, it's done
so on-the-fly by a fairly complex set of Perl5 operations. Although it
wouldn't be the hardest thing in the world to convert {} and [] to // and **
when dispatching articles by e-mail, I sure don't like the idea of firing up
Perl to convert the content on-the-fly from the news server every time an
article is fetched. An alternative to on-the-fly converstion might be to
store two copies of articles on the server (one raw, one converted), but that
seems pretty gunky to me.
But all of the above presupposes that the conversion should even happen in the
first place, which I disagree with anyway. If the conversion did occur, what
would happen in replies? How would those get magicly converted back?
Let's approach the problem from a different direction: What if FTX simply
supported // and ** directly? I'll bet that if it did, a lot of people would
choose that over {} and [], and you'd be able to read them in Mozilla.
Now I'm tossing this idea out without having thoroughly thought through the
ramifications of parsing. The syntax of // and ** is much more subtle than
the syntax of {} and [].
Help convince me that // and ** could be parsed without horrible anomalies
and exception tables. Here is some food for thought...
Menelaus replied, "/How/ do I take your meaning? Am I to stay with ...
"Have you read Steven King's /Pet Semetary/?" she asked.
"What's the /Herald/'s position on this issue?"
Coffee with cream /and/or/ sugar.
"No, you're supposed to turn it /counter/clockwise!"
I think all of those except the last one could be handled with some clever
word-boundary detection. Some more:
* = .10 level (10%) ** = .05 level (5%) *** = .01 level (1%)
I think *Bill* was the one who mentioned this earlier.*
Dude, I love your mosaic! (*Wow!*)
Microsoft (*cough* bloat *cough*) Windows.
(*^_^*)
I suppose that defining a word boundary might start with dividing the ASCII
character set into 3 character classes...
1. Whitespace - tabs, spaces, newlines
2. Printable "punctuation" characters - .,:;!?'"()[]{}
3. Everything else - letters, numbers, and other symbols
...and then doing...
my $FOO = q{.,:;!?'"()[]{}};
my $BAR = join "", grep { !m/[\Q$foo\E\s]/ } map { chr($_) } (32..255);
s{\b([\Q$FOO\E]*)/([\Q$BAR\E])}{$1<I>$2}g;
s{([\Q$BAR\E])/([\Q$FOO\E]*)\b}{$1</I>$2}g;
s{\b([\Q$FOO\E]*)\*([\Q$BAR\E])}{$1<B>$2}g;
s{([\Q$BAR\E])\*([\Q$FOO\E]*)\b}{$1</B>$2}g;
...or something like that. But that doesn't handle (*Wow!*) or /Herald/'s
yet.
--Todd
|
|
Message has 2 Replies:
Message is in Reply To:
 | | Re: testing in rtl...
|
| (...) [...] (...) Mozilla does this too (the *bold* and _underline_ and /italics/ convention, and the smileys). But I don't use Mozilla on Lugnet. I agree that FTX should translate {} and [] to // and ** when displaying in plain text, and I'm *not* (...) (21 years ago, 29-May-03, to lugnet.publish)
|
31 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
