hi,
i was looking for an option to turn off case sensitive linking, cause [[Building]] and [[building]] and [[BuilDing]] should be the same in our wiki. does this exist?
-solo.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
solo turn wrote: | i was looking for an option to turn off case sensitive linking, cause | [[Building]] and [[building]] and [[BuilDing]] should be the same in | our wiki. does this exist?
No. The normal behavior is to force the first letter of a title to uppercase, which makes [[Building]] and [[building]] match. There is no full case-insensitivity option at present.
- -- brion vibber (brion @ pobox.com)
Actually what I'd love to see is a way to automatically lower case a reference.
I don't know how may times I've typed something like.
The [[Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Framis Bulkhead|framis bulkhead]]
or
The [[Globulator:Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Globulator:Framis Bulkhead|framis bulkhead]]
Just like [[Globulator:Farbulator Overflow Valve|]] seems to get saved as [[[[Globulator:Farbulator Overflow Valve|Farbulator Overflow Valve]] it would be very nice to have a shortcut for lowercasing the expansion after the |
On Sun, 30 Jan 2005 22:35:36 -0800, Brion Vibber brion@pobox.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
solo turn wrote: | i was looking for an option to turn off case sensitive linking, cause | [[Building]] and [[building]] and [[BuilDing]] should be the same in | our wiki. does this exist?
No. The normal behavior is to force the first letter of a title to uppercase, which makes [[Building]] and [[building]] match. There is no full case-insensitivity option at present.
- -- brion vibber (brion @ pobox.com)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFB/dG4wRnhpk1wk44RAo1eAKDDFFm4jF3/OD9tv8QnWeKSAhfpaQCfaywT nU5uJLwqIETgMUt1UGfIBiU= =SdQp -----END PGP SIGNATURE----- _______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
The solution to this is to only use lowercase letters except in proper names. ie, the names of the two articles would be "Farbulator overflow valve" and "Framis bulkhead". Now if there was a band called Framis Bulkhead, you could have two seperate articles: one for the band and one for the mounting location.
On Mon, 31 Jan 2005 13:39:11 -0500, Rick DeNatale rick.denatale@gmail.com wrote:
Actually what I'd love to see is a way to automatically lower case a reference.
I don't know how may times I've typed something like.
The [[Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Framis Bulkhead|framis bulkhead]]
or
The [[Globulator:Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Globulator:Framis Bulkhead|framis bulkhead]]
Just like [[Globulator:Farbulator Overflow Valve|]] seems to get saved as [[[[Globulator:Farbulator Overflow Valve|Farbulator Overflow Valve]] it would be very nice to have a shortcut for lowercasing the expansion after the |
On Sun, 30 Jan 2005 22:35:36 -0800, Brion Vibber brion@pobox.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
solo turn wrote: | i was looking for an option to turn off case sensitive linking, cause | [[Building]] and [[building]] and [[BuilDing]] should be the same in | our wiki. does this exist?
No. The normal behavior is to force the first letter of a title to uppercase, which makes [[Building]] and [[building]] match. There is no full case-insensitivity option at present.
- -- brion vibber (brion @ pobox.com)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFB/dG4wRnhpk1wk44RAo1eAKDDFFm4jF3/OD9tv8QnWeKSAhfpaQCfaywT nU5uJLwqIETgMUt1UGfIBiU= =SdQp -----END PGP SIGNATURE----- _______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
On Tue, 1 Feb 2005 15:52:25 -0500, Jamie Bliss astronouth7303@gmail.com wrote:
The solution to this is to only use lowercase letters except in proper names.
On Mon, 31 Jan 2005 13:39:11 -0500, Rick DeNatale rick.denatale@gmail.com wrote:
Actually what I'd love to see is a way to automatically lower case a reference.
I don't know how may times I've typed something like.
The [[Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Framis Bulkhead|framis bulkhead]]
if there are many users with different spelling behaviour it looks more like a "workaround" than a "solution" ;) if you are allowed to type it wrong, users will type it wrong.
an option to ignore case would solve our problems. in which place/file would it fit best?
-solo.
I agree with Solo. I can help to make necessary changes. It doesn't really make sense to have two articles with different letter cases when they mean the same thing. I think this should be included in the newer versions of Mediawiki, too.
-----Original Message----- From: mediawiki-l-bounces@Wikimedia.org [mailto:mediawiki-l- bounces@Wikimedia.org] On Behalf Of solo turn Sent: Thursday, February 03, 2005 3:40 PM To: MediaWiki announcements and site admin list Subject: Re: [Mediawiki-l] case insensitive links
On Tue, 1 Feb 2005 15:52:25 -0500, Jamie Bliss astronouth7303@gmail.com wrote:
The solution to this is to only use lowercase letters except in proper
names.
On Mon, 31 Jan 2005 13:39:11 -0500, Rick DeNatale rick.denatale@gmail.com wrote:
Actually what I'd love to see is a way to automatically lower case a
reference.
I don't know how may times I've typed something like.
The [[Farbulator Overflow Valve|farbulator overflow valve]] was mounted on the [[Framis Bulkhead|framis bulkhead]]
if there are many users with different spelling behaviour it looks more like a "workaround" than a "solution" ;) if you are allowed to type it wrong, users will type it wrong.
an option to ignore case would solve our problems. in which place/file would it fit best?
-solo.
Muzaffer Ozakca wrote:
I agree with Solo. I can help to make necessary changes. It doesn't really make sense to have two articles with different letter cases when they mean the same thing. I think this should be included in the newer versions of Mediawiki, too.
It's non-trivial; the entire system assumes that normalized titles are binary comparable.
-- brion vibber (brion @ pobox.com)
On Thu, 3 Feb 2005 21:40:25 +0100, solo turn soloturn@gmail.com wrote:
an option to ignore case would solve our problems. in which place/file would it fit best?
One problem that occurs to me with this is that if everything in the code treats titles as case insensitive, where do you define what the "normalised" version of the title is - e.g. how do you display it when the user goes to the page? You could have "uppercase first, rest lower", but then you'd get things like "Microsoft windows xp", which is just wrong; alternatively, you could have "uppercase first of each word", but that would look odd in other circumstances, particularly if you didn't special case words like "of", "the", "and", etc (which, in turn, would be a nightmare for l10n...)
It's been suggested [bug 469*] that (because even the current system "guesses" that you want the first letter capitalised, causing a few nuisances, such as "H2g2") there should be markup to override the display of the title - but nobody's actually coded it yet. If they did, you could probably come up with a capitalisation default that was "good enough for most cases", but it would mean people would have to go round overriding titles - so they might just as well go round correcting badly capitalised links (as Jamie says, either "Framis Bulkhead" is a proper name, or it shouldn't be written like that anyway).
* http://bugzilla.wikimedia.org/show_bug.cgi?id=496
One problem that occurs to me with this is that if everything in the code treats titles as case insensitive, where do you define what the "normalised" version of the title is - e.g. how do you display it when the user goes to the page? You could have "uppercase first, rest lower", but then you'd get things like "Microsoft windows xp", which is just wrong; alternatively, you could have "uppercase first of each word", but that would look odd in other circumstances, particularly if you didn't special case words like "of", "the", "and", etc (which, in turn, would be a nightmare for l10n...)
It's been suggested [bug 469*] that (because even the current system "guesses" that you want the first letter capitalised, causing a few nuisances, such as "H2g2") there should be markup to override the display of the title - but nobody's actually coded it yet. If they did, you could probably come up with a capitalisation default that was "good enough for most cases", but it would mean people would have to go round overriding titles - so they might just as well go round correcting badly capitalised links (as Jamie says, either "Framis Bulkhead" is a proper name, or it shouldn't be written like that anyway).
How about the following resolution: 1. Use some newly defined markup for the page title, if present 2. Search the first paragraph/n words for a match to the link name ignoring case, in bold. If found, use that as the page name. 3. Default to the case of the link name (including redirect links).
I don't yet know enough about the guts of the code to say if this is easy, hard, or impossible. But it seems like it could work conceptually.
One question is how to deal with already existing pages where the name differs only in case.
-Rich Holton
__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
On Fri, 4 Feb 2005 06:50:16 -0800 (PST), Rich Holton rich_holton@yahoo.com wrote:
How about the following resolution:
- Use some newly defined markup for the page title, if present
- Search the first paragraph/n words for a match to the link name
ignoring case, in bold. If found, use that as the page name.
Not an easy step, but possible I guess. Of course, it assumes, as does the rest of this, that you can guarantee a clean case-insensitive search (think different character encodings, etc) - and also that the title as written in the page is *punctuated* identically to that of the title. Of course, a page like [[acid]] would end up with its title displayed in all lowercase that way, which isn't necessarily a good thing, but it would lend some degree of consistency I guess.
- Default to the case of the link name (including redirect links).
Note that there is no way of knowing what *link* you came in by, only what *URL* - it might not have been an internal link at all. But yes, you could use that. Again, things like "acid", referred to in the middle of sentences, wouldn't become "Acid" as they do atm.
I don't yet know enough about the guts of the code to say if this is easy, hard, or impossible. But it seems like it could work conceptually.
An additional thought to throw into the ring is how to deal with the title in *generated* text, like Special:Watchlist, Special:Recentchanges, etc. Perhaps the 'display title' would have to be stored in the database, and updated whenever the page was edited. Come to think of it, that would make the checks a lot more efficient, too, because they could just be skipped on most views.
And that all assumes that Brion's comment about assumptions inherent in the code doesn't prove too huge an obstacle.
One question is how to deal with already existing pages where the name differs only in case.
That would be a problem if, say, the English Wikipedia switched to using the feature, but I think that's unlikely to happen for all sorts of reasons (it would have such far-reaching consequences, that pages with only case difference would the least of our worries). For smaller existing installations, it should be possible to craft a SQL query to find the titles which need changing, and then follow the backlinks to change references to them too.
-----Original Message----- On Thu, 3 Feb 2005 21:40:25 +0100, solo turn soloturn@gmail.com wrote:
an option to ignore case would solve our problems. in which place/file would it fit best?
One problem that occurs to me with this is that if everything in the code treats titles as case insensitive, where do you define what the "normalised" version of the title is - e.g. how do you display it when the user goes to the page? You could have "uppercase first, rest lower", but then you'd get things like "Microsoft windows xp", which is just wrong; alternatively, you could have "uppercase first of each word", but that would look odd in other circumstances, particularly if you didn't special case words like "of", "the", "and", etc (which, in turn, would be a nightmare for l10n...)
I think an article title should be however the first creator assigns it to be. If the first person gives "Microsoft windows xp" it should be titled that way until somebody moves it to the correct title. However, all references (links) to "Microsoft windows xp", "microsoft Windows XP" should point to the only page existing. We can't assume people know the correct capitalization all the time, even then they will make mistakes. So, search in the database should be case insensitive but the correct capitalizations should be achieved manually.
It's been suggested [bug 469*] that (because even the current system "guesses" that you want the first letter capitalised, causing a few nuisances, such as "H2g2") there should be markup to override the display of the title - but nobody's actually coded it yet. If they did, you could probably come up with a capitalisation default that was "good enough for most cases", but it would mean people would have to go round overriding titles - so they might just as well go round correcting badly capitalised links (as Jamie says, either "Framis Bulkhead" is a proper name, or it shouldn't be written like that anyway).
Maybe instead of this, the same approach above can be taken. Nothing is automatically capitalized and if there are mistakes somebody else corrects it. Isn't it how the wiki concept works, anyways.
-- Rowan Collins BSc [IMSoP] _______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
On Fri, 4 Feb 2005 11:01:38 -0500, Muzaffer Ozakca mozakca@indiana.edu wrote:
I think an article title should be however the first creator assigns it to be. If the first person gives "Microsoft windows xp" it should be titled that way until somebody moves it to the correct title. However, all references (links) to "Microsoft windows xp", "microsoft Windows XP" should point to the only page existing. We can't assume people know the correct capitalization all the time, even then they will make mistakes. So, search in the database should be case insensitive but the correct capitalizations should be achieved manually.
I guess my general opinion is that any reference to "microsoft windows xp" is incorrect and needs fixing *anyway*, so why bother making the link work? I know in many cases it's not as clear-cut as that, but on any site that's aiming for a reasonable consistency of style, even a phrase like "list of software design companies" will have a "correct" capitalisation.
I guess I should have thought that the obvious system is to have a particular capitalisation stored as the title, with only *links* being case insensitive - and hope that someone will spot that it needs moving. But how you could ever track down incorrectly typed instances under such a system, I'm not sure - links to "microsoft windows xp" would be buried in "what links here" along with all the correct references. At least under the current system you can create a redirect, or move the page leaving a redirect behind, and then the incorrect pages can be identified as linking to that redirect.
In other words, I think such a system would just encourage sloppy style, whereas MediaWiki (due to its primary role of running Wikipedia et al) is largely designed to aid in authoring "professional-looking" content (I would claim the very existence of [[free links]] as opposed to CamelCase, as well as things like seperation of discussion pages, as examples of this design goal).
-----Original Message-----
I guess my general opinion is that any reference to "microsoft windows xp" is incorrect and needs fixing *anyway*, so why bother making the link work? I know in many cases it's not as clear-cut as that, but on any site that's aiming for a reasonable consistency of style, even a phrase like "list of software design companies" will have a "correct" capitalisation.
Just a quick comment here, even if "microsoft windows xp" is incorrect it still works. It will create a link to an empty page. Then somebody has to notice it and correct it. By the time, somebody else could have typed some text and there are two versions of the same thing.
I guess I should have thought that the obvious system is to have a particular capitalisation stored as the title, with only *links* being case insensitive - and hope that someone will spot that it needs moving. But how you could ever track down incorrectly typed instances under such a system, I'm not sure - links to "microsoft windows xp" would be buried in "what links here" along with all the correct references. At least under the current system you can create a redirect, or move the page leaving a redirect behind, and then the incorrect pages can be identified as linking to that redirect.
I think, we are talking about the same thing. As long as links are case insensitive, it shouldn't really matter how it is stored in the database.
In other words, I think such a system would just encourage sloppy style, whereas MediaWiki (due to its primary role of running Wikipedia et al) is largely designed to aid in authoring "professional-looking" content (I would claim the very existence of [[free links]] as opposed to CamelCase, as well as things like seperation of discussion pages, as examples of this design goal).
(Btw, sorry, for copying e-mails to your personal e-mail address)
-- Rowan Collins BSc [IMSoP] _______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
On Fri, 4 Feb 2005 11:54:23 -0500, Muzaffer Ozakca mozakca@indiana.edu wrote:
Just a quick comment here, even if "microsoft windows xp" is incorrect it still works. It will create a link to an empty page. Then somebody has to notice it and correct it. By the time, somebody else could have typed some text and there are two versions of the same thing.
That's a good point - although at least an experienced user would know that the page might exist at another capitalisation. In fact, the simplest solution to this would be to add some logic to the non-existent page display: in addition to the [[MediaWiki:Newpagetext]] warning users to do this themselves, the software could search for pages with the same name but different capitalisations - if it found one, it could add an extra message, such as:
NOTE: You are about to create a page called "microsoft windows xp", but there is already a page called [[Microsoft Windows XP]]. Please make sure you are not creating a duplicate of this page before proceeding. If you followed a link in another article, you might want to edit that page to refer to the correct name; if you think this is likely to be a common mistake (or if there are already many pages [[Special:Whatlinkshere/{{PAGENAME}}|linking to this title]]) you might also consider creating a [[Help:Redirect|redirect]] to the correct title.
While this would add the burden of an automatic search every time a red link was followed, this would logically be less than having to search case-insensitively for *every* link.
Meanwhile, my point stands about how you would find and correct links which had been typed wrong if they were all treated as references to the same page - there's no obvious mechanism for using "what links here" to see which pages use "[[microsoft windows xp]]" and which "[[Microsoft Windows XP]]" if those titles are not treated as distinct.
But perhaps I should be more open-minded and offer solutions as well as problems. If the links tables were all id->name pairs (as seems sensible, and could perhaps be brought in along with unifying them w/ a link_type field), the references in the database could conserve the exact capitalisation of the link. Assuming an efficient case-insensitive DB query is possible elsewhere, it would then be no stretch for Special:Whatlinkshere to look up all variant capitalisations at once. Then, it could sort its output by how the link was typed, and those which represented "incorrect" capitalisations could be identified, and the pages corrected.
(Btw, sorry, for copying e-mails to your personal e-mail address)
*shrug* it happens. I filter on the subject field anyway, so I barely notice. :)
On Tue, 1 Feb 2005 15:52:25 -0500, Jamie Bliss astronouth7303@gmail.com wrote:
The solution to this is to only use lowercase letters except in proper names. ie, the names of the two articles would be "Farbulator overflow valve" and "Framis bulkhead". Now if there was a band called Framis Bulkhead, you could have two seperate articles: one for the band and one for the mounting location.
I agree with your main point - if the term is correctly written with capitals, it should be written with capitals; it not, the article title should match how it's written.
I would, however, advise against having two titles differing only in capitalisation, because this seems to screw up browser histories, at least in Mozilla & Firefox - to demonstrate, use an existing browser window/tab to go to http://en.wikipedia.org/wiki/Acid and then http://en.wikipedia.org/wiki/ACID; now click your "back" button, and be confused... [Now I see that this doesn't happen in IE, I guess it's a moz bug, but it's still confusing, and there may be similar problems elsewhere and in other browsers]
Rowan Collins wrote:
I would, however, advise against having two titles differing only in capitalisation, because this seems to screw up browser histories, at least in Mozilla & Firefox - to demonstrate, use an existing browser window/tab to go to http://en.wikipedia.org/wiki/Acid and then http://en.wikipedia.org/wiki/ACID; now click your "back" button, and be confused... [Now I see that this doesn't happen in IE, I guess it's a moz bug, but it's still confusing, and there may be similar problems elsewhere and in other browsers]
I can confirm this behavior in Firefox 1.0 (Mac OS X). Please file a bug report at http://bugzilla.mozilla.org
-- brion vibber (brion @ pobox.com)
And it doesn't matter if the non-pretty forms of the uris are used.
I substituted http://en.wikipedia.org/wiki/index.php?title=Acid and http://en.wikipedia.org/wiki/index.php?title=ACID
and Firefox 1.0 on Linux exhibits the same odd behavior.
On Fri, 04 Feb 2005 11:41:51 -0800, Brion Vibber brion@pobox.com wrote:
Rowan Collins wrote:
I would, however, advise against having two titles differing only in capitalisation, because this seems to screw up browser histories, at least in Mozilla & Firefox - to demonstrate, use an existing browser window/tab to go to http://en.wikipedia.org/wiki/Acid and then http://en.wikipedia.org/wiki/ACID; now click your "back" button, and be confused... [Now I see that this doesn't happen in IE, I guess it's a moz bug, but it's still confusing, and there may be similar problems elsewhere and in other browsers]
I can confirm this behavior in Firefox 1.0 (Mac OS X). Please file a bug report at http://bugzilla.mozilla.org
-- brion vibber (brion @ pobox.com)
MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
where could i replace this "first letter capitalization" with a "force to lowercase"? this would have the same effect, and therefor it would not need to be case insensitive.
On Sun, 30 Jan 2005 22:35:36 -0800, Brion Vibber brion@pobox.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
solo turn wrote: | i was looking for an option to turn off case sensitive linking, cause | [[Building]] and [[building]] and [[BuilDing]] should be the same in | our wiki. does this exist?
No. The normal behavior is to force the first letter of a title to uppercase, which makes [[Building]] and [[building]] match. There is no full case-insensitivity option at present.
- -- brion vibber (brion @ pobox.com)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFB/dG4wRnhpk1wk44RAo1eAKDDFFm4jF3/OD9tv8QnWeKSAhfpaQCfaywT nU5uJLwqIETgMUt1UGfIBiU= =SdQp -----END PGP SIGNATURE----- _______________________________________________ MediaWiki-l mailing list MediaWiki-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/mediawiki-l
mediawiki-l@lists.wikimedia.org