Actually, // and ** are at least as clear, and are most definitely parsable by a fixed-lookahead context-free grammar - even an unaugmented LL(k) grammar could probably handle it. <i> and <b> are unambiguous, but ugly and language-dependent. MediaWiki's current behavior "fixes" many of the issues with its ambiguous bold/italics representation with little ad-hoc DWIM-type behavior. It works, but cannot be represented by a CFG and is difficult to extend.
Another benefit of changing the double single quotes and triple single quotes to // and ** respectively is that it would be a small step to making MediaWiki markup more Creole-compatible (www.wikicreole.org). Also, it seems like a conversion script for just these two elements would not be that difficult to write. What could be potential complications?
Chuck
wikitech-l-bounces@lists.wikimedia.org schrieb am 14.02.2007 10:42:47:
Actually, // and ** are at least as clear, and are most definitely parsable by a fixed-lookahead context-free grammar - even an
unaugmented
LL(k) grammar could probably handle it. <i> and <b> are unambiguous,
but
ugly and language-dependent. MediaWiki's current behavior "fixes" many of the issues with its ambiguous bold/italics representation with
little
ad-hoc DWIM-type behavior. It works, but cannot be represented by a
CFG
and is difficult to extend.
Another benefit of changing the double single quotes and triple single quotes to // and ** respectively is that it would be a small step to making MediaWiki markup more Creole-compatible (www.wikicreole.org). Also, it seems like a conversion script for just these two elements would not be that difficult to write. What could be potential complications?
The users.
On 2/14/07, Chuck Smith chuckssmith@gmail.com wrote:
Another benefit of changing the double single quotes and triple single quotes to // and ** respectively is that it would be a small step to making MediaWiki markup more Creole-compatible (www.wikicreole.org). Also, it seems like a conversion script for just these two elements would not be that difficult to write. What could be potential complications?
In terms of the pure syntax?
1) C++-style comments use //. 2) URLs use // (and probably not all will be linked). 3) Lists use **. 4) ** is often used for footnote-style things, once * has been used (although this usage may not be common on WMF projects). 5) Probably there are some weird languages out there that use these somehow.
All of those are, however, fairly easily surmountable with a well-designed initial pass to clean them up (nowikiing where needed, adding spaces after line-initial ** outside of nowiki/pre, flagging URL-y things to be looked at), at least in theory. The basic problem is, as Christopher says, getting users who are used to '''? to start using (**|//). It would be frustrating, and we'd get invalid markup being put into articles regularly for months. And what would the benefit be? Putting a Band-Aid on a missing limb of parseability? Compatibility with wikisyntax that we definitely don't need and that none of our import sources use? Not worth it, not by a long shot.
On Wednesday 14 February 2007 10:42:47 Chuck Smith wrote:
Another benefit of changing the double single quotes and triple single quotes to // and ** respectively is that it would be a small step to making MediaWiki markup more Creole-compatible (www.wikicreole.org). Also, it seems like a conversion script for just these two elements would not be that difficult to write. What could be potential complications?
I don't want to disrupt you but I definitely like the ' syntax and I am not willing to stop using it.
Why?
'' get used especially for highlighting of single words words or short sentences of other people. So it is used very much for quick and dirty "inline"-quoting. You probably can argue that " '' are hard to distinguish and that a proper quote needs " (well proper quotes use a bunch of other characters not on our standard keyboards ;-) beside italic.
Another point is that * syntax clashes with lists. Lists with * are a really intuitive thing. Do not mix them with something else (for example what about italic in a list, something I really use often). Something like ** [space] * is not going to work. Never.
So although the current mediawiki syntax has its disadvantages, the core syntax (I don't talk about templates) is very much balanced and is causing very few problems withing our limited common standard character set.
And of course Usenet doesn't count. It is about exchanging arguments and not about creating articles.
Arnomane
On Thu, Feb 15, 2007 at 06:08:11PM +0100, Daniel Arnold wrote:
Another point is that * syntax clashes with lists. Lists with * are a really intuitive thing. Do not mix them with something else (for example what about italic in a list, something I really use often). Something like ** [space] * is not going to work. Never.
I keep hearing this asserted, but with nothing other than personal, anecdotal opinions as evidence.
My personal perception is that I don't expect
* This is an element ** This is a second level *element with a bold phrase* ** *This is another second level element, all bold* * One more element
to be at all difficult for people to cope with. The lists I see marked up in my limited work on en.W seem to go about 60/40 having the space.
Anyone with the full corpus on a machine who can check that statistic for real?
And, why is is to difficult, Daniel, for people to have to put a space in after a list item marker? That seems much less ambiguous than
'''''this is'' a weird phrase''' '''''this one''' is too''
to *me*. In list markup, it's reasonable to eat the whitespace character, which it is *not* reasonable for a parser to do in the bold/italics collision case.
Cheers, -- jra
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
[blah blah blah]
Everybody, please don't bother discussing alternative bold and italic markup. That's not on the table right now.
- -- brion vibber (brion @ pobox.com / brion @ wikimedia.org)
On 2/16/07, Brion Vibber brion@pobox.com wrote:
Everybody, please don't bother discussing alternative bold and italic markup. That's not on the table right now.
Which was my original question: has it absolutely been ruled out. Answer: yes.
Thanks, Steve
Brion Vibber wrote:
[blah blah blah]
Everybody, please don't bother discussing alternative bold and italic markup. That's not on the table right now.
Hi Brion,
thanks for this message. However, please may I cheekishly take the liberty to point out that it could have been worded less authoritatively? You see, while it is of course perfectly within your rights and freedoms to leave the bold/italic mark-up the way it is, I am finding it somewhat objectionable that you are implying that we should all stop discussing it. Perhaps a better way to respond would be something like: "Bold and italic mark-up is not on the table right now, so I am going to ignore the rest of this thread." This will tell the participants that the discussion is futile, but does not imply that it is in some way "forbidden".
Thanks, Timwi
Timwi wrote:
Brion Vibber wrote:
[blah blah blah] Everybody, please don't bother discussing alternative bold and italic markup. That's not on the table right now.
thanks for this message. However, please may I cheekishly take the liberty to point out that it could have been worded less authoritatively?
I was under the impression that Brion is *paid* to give authoritative answers.
Or have I misunderstood horribly?
HTH HAND
Phil Boswell wrote:
Timwi wrote:
Brion Vibber wrote:
[blah blah blah] Everybody, please don't bother discussing alternative bold and italic markup. That's not on the table right now.
thanks for this message. However, please may I cheekishly take the liberty to point out that it could have been worded less authoritatively?
I was under the impression that Brion is *paid* to give authoritative answers.
Authoritative over the MediaWiki codebase, yes, and I fully respect that authority. But I don't think he should have authority over the set of permissible discussions on the mailing list.
Timwi
Would it add anything to this debate if I said that, in discussion with publishers, OLPC has found vehement objection to the ''' and '' bold/italic syntax, and strong support for using * and / instead?
- Eric Astor
Daniel Arnold wrote:
On Wednesday 14 February 2007 10:42:47 Chuck Smith wrote:
Another benefit of changing the double single quotes and triple single quotes to // and ** respectively is that it would be a small step to making MediaWiki markup more Creole-compatible (www.wikicreole.org). Also, it seems like a conversion script for just these two elements would not be that difficult to write. What could be potential complications?
I don't want to disrupt you but I definitely like the ' syntax and I am not willing to stop using it.
Why?
'' get used especially for highlighting of single words words or short sentences of other people. So it is used very much for quick and dirty "inline"-quoting. You probably can argue that " '' are hard to distinguish and that a proper quote needs " (well proper quotes use a bunch of other characters not on our standard keyboards ;-) beside italic.
Another point is that * syntax clashes with lists. Lists with * are a really intuitive thing. Do not mix them with something else (for example what about italic in a list, something I really use often). Something like ** [space] * is not going to work. Never.
So although the current mediawiki syntax has its disadvantages, the core syntax (I don't talk about templates) is very much balanced and is causing very few problems withing our limited common standard character set.
And of course Usenet doesn't count. It is about exchanging arguments and not about creating articles.
Arnomane
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Eric Astor wrote:
Would it add anything to this debate if I said that, in discussion with publishers, OLPC has found vehement objection to the ''' and '' bold/italic syntax, and strong support for using * and / instead?
Not really. Changing MediaWiki syntax isn't considered an option at the moment, so this thread is pointless.
Ivan Krstić wrote:
Eric Astor wrote:
Would it add anything to this debate if I said that, in discussion with publishers, OLPC has found vehement objection to the ''' and '' bold/italic syntax, and strong support for using * and / instead?
Not really. Changing MediaWiki syntax isn't considered an option at the moment, so this thread is pointless.
Sorry - I thought as much, but wrote my e-mail before seeing Brion's official response.
wikitech-l@lists.wikimedia.org