Hello, I wondered if any of the programmers out there could help me. When I first started contributing to wikipedia, I took it upon myself to add all the country info in the CIA World Factbook. This was before Larry hashed out all the consequences of subpages, and so I just followed the established model instead of creating a new one. Now all the countries of the world except for a few of the A's are in subpage format, and between all the individual entries and the [[countries of the world]] page, that's a lot of influence towards creating subpages. I've worked on converting the pages, but the process is slow and tedious. So I wanted to know if it would be possible to write a script to convert those pages, and if anyone out there has the time and inclination to do so. I'd do it myself except my last programming experience was in BASIC. <g>
thanks,
kq
p.s. I really don't know how long it might take to write this script, so if it's some ridiculous amount of time, just forget about it.0
That could be done, but we (I) would have to know what to convert, and into what format.
Would that be [[Egypt/History]] => [[Egypt (History)]]? Or rather => [[History of Egypt]]? Same for Geography, or different? And so on.
I suggest making this a function that is available only for sysops in the sidebar. I am certain that Larry will be most eager to deconstruct some subpages ;)
A similar function is planned for moving/renaming pages, including subpages (for example, to move them to the wikipedia namespace).
Should we attempt to convert all links to these pages automatically, or should we use #REDIRECTs? The former would be more elegant, but also more complicated, while the latter is rather simple but crude.
Magnus
-----Original Message----- From: wikipedia-l-admin@nupedia.com [mailto:wikipedia-l-admin@nupedia.com]On Behalf Of koyaanisqatsi@nupedia.com Sent: Saturday, February 23, 2002 5:10 PM To: wikipedia-l@nupedia.com Subject: [Wikipedia-l] subpages
Hello, I wondered if any of the programmers out there could help me. When I first started contributing to wikipedia, I took it upon myself to add all the country info in the CIA World Factbook. This was before Larry hashed out all the consequences of subpages, and so I just followed the established model instead of creating a new one. Now all the countries of the world except for a few of the A's are in subpage format, and between all the individual entries and the [[countries of the world]] page, that's a lot of influence towards creating subpages. I've worked on converting the pages, but the process is slow and tedious. So I wanted to know if it would be possible to write a script to convert those pages, and if anyone out there has the time and inclination to do so. I'd do it myself except my last programming experience was in BASIC. <g>
thanks,
kq
p.s. I really don't know how long it might take to write this script, so if it's some ridiculous amount of time, just forget about it.0
Magnus Manske wrote:
Would that be [[Egypt/History]] => [[Egypt (History)]]? Or rather => [[History of Egypt]]? Same for Geography, or different? And so on.
The current work, at least for the /History pages, is to use the form "History of Egypt".
I've started to rename some /History pages in this fashion. I also make sure I change all old links to [[Egypt/History]] into links to [[History of Egypt]], so that the "pages that link here" link will work as expected. (It would help if links to redirects were rendered in a different color.) I also wikify any years and dates that would be useful, and add a context at the top of the article, something like this:
This is the [[history]] of [[Egypt]]. See also the [[history of Africa]] and the [[history of present-day nations and states]].
See also http://www.wikipedia.com/wiki/History_of_present-day_nations_and_states and http://www.wikipedia.com/wiki/user:LA2
I think we should have subpage functionality (or modify the namespace functionality) to help with pages such as the September 11 pages. The current situation is a step back from the earlier software.
It really isn't. It's a strong step forward, and I'm very glad we've done it.
On Sun, 24 Feb 2002 kband@www.llamacom.com wrote:
I think we should have subpage functionality (or modify the namespace functionality) to help with pages such as the September 11 pages. The current situation is a step back from the earlier software.
[Wikipedia-l] To manage your subscription to this list, please go here: http://www.nupedia.com/mailman/listinfo/wikipedia-l
From: kband@www.llamacom.com
I think we should have subpage functionality (or modify the namespace functionality) to help with pages such as the September 11 pages. The current situation is a step back from the earlier software.
I disagree. It will lead to more self-contained articles about things like the hijackers, which is what we want in this encyclopedia.
-- Jan Hidders
From: kband@www.llamacom.com
I think we should have subpage functionality (or modify the namespace functionality) to help with pages such as the September 11 pages. The current situation is a step back from the earlier software.
I disagree. It will lead to more self-contained articles about things like the hijackers, which is what we want in this encyclopedia.
What exactly do you mean? As the one who made pretty much all the Sept. 11 pages, I can tell you that the subpage functionality made it much easier--I certainly wouldn't have been able to deal with all of the information in there if I didn't make some kind of subpagery. I'm certainly not tied to the particular implementation of the UseModWiki, but the ability to generate pages tied into another page without having to type the entire title every time is extremely useful.
-TC
Please--let's not get into this again. We haven't even given a subpageless Wikipedia the old college try yet.
Larry
I just realized we don't need to forbid the use of colons in titles; unlike the subpage "/", the only time colons are used as a special character are with the limited number of namespaces.
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
-TC
From: kband@www.llamacom.com
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then the first colon is always the end of the namespace.
-- Jan Hidders
From: kband@www.llamacom.com
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then the first colon is always the end of the namespace.
That's perfect.
At 09:56 AM 2/25/02 -0600, kband wrote:
From: kband@www.llamacom.com
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then the first colon is always the end of the namespace.
That's perfect.
Would this require us to explicitly put "content:" in front of every link?
If so, I doubt it would work--people would forget, because it's not intuitive.
From: "Vicki Rosenzweig" vr@redbird.org
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then
the
first colon is always the end of the namespace.
That's perfect.
Would this require us to explicitly put "content:" in front of every link?
Not necessarily. I would propose that a link without a namespace should default to the namespace that it is in. That's how namespaces usually work anyway. It means for example that if you create a link on your home page without a namespace it becomes a page in the user: namespace. (Actually I still believe that every user should have his or her own namespace, but that is another issue.)
-- Jan Hidders
The original software *did* stay within the current namespace if no other was explicitly given. So, "[[test]]" on the page "[[user:Magnus Manske]]" did link to "[[user:test]]" instead of "[[test]]". At that time, I still planned a "stable:" namespace for approved articles, which would automatically link only to other stable articles.
The "blank" namespace was (and still should be) accessible through "[[:Main Page]]" (note the leading colon).
The main switch for that should still be somewhere in wikiTextEn.php.
Magnus
-----Original Message----- From: wikipedia-l-admin@nupedia.com [mailto:wikipedia-l-admin@nupedia.com]On Behalf Of Jan Hidders Sent: Monday, February 25, 2002 6:26 PM To: wikipedia-l@nupedia.com Subject: Re: [Wikipedia-l] colons
From: "Vicki Rosenzweig" vr@redbird.org
There is a simpeler solution. The actual contents of
Wikipedia get the
namespace "content:". If everything is prefixed with a
namespace then the
first colon is always the end of the namespace.
That's perfect.
Would this require us to explicitly put "content:" in front of
every link?
Not necessarily. I would propose that a link without a namespace should default to the namespace that it is in. That's how namespaces usually work anyway. It means for example that if you create a link on your home page without a namespace it becomes a page in the user: namespace. (Actually I still believe that every user should have his or her own namespace, but that is another issue.)
-- Jan Hidders
[Wikipedia-l] To manage your subscription to this list, please go here: http://www.nupedia.com/mailman/listinfo/wikipedia-l
From: "Magnus Manske" Magnus.Manske@epost.de
The original software *did* stay within the current namespace if no other was explicitly given. So, "[[test]]" on the page "[[user:Magnus Manske]]" did link to "[[user:test]]" instead of "[[test]]". At that time, I still planned a "stable:" namespace for approved articles, which would automatically link only to other stable articles.
It's already there? So all we would have to do is take care that the content: namespace doesn't show up in the title? (Plus change some code that assumes that any title that contains a ":" is not content.) Can I, can I, can I? Please? :-) As far as I can see this actually makes the code simpeler because titles will always have the same format.
The "blank" namespace was (and still should be) accessible through
"[[:Main
Page]]" (note the leading colon).
Yup, works on my site.
-- Jan Hidders
On Mon, 25 Feb 2002, Jan Hidders wrote:
From: "Magnus Manske" Magnus.Manske@epost.de
The original software *did* stay within the current namespace if no other was explicitly given. So, "[[test]]" on the page "[[user:Magnus Manske]]" did link to "[[user:test]]" instead of "[[test]]". At that time, I still planned a "stable:" namespace for approved articles, which would automatically link only to other stable articles.
It's already there? So all we would have to do is take care that the content: namespace doesn't show up in the title? (Plus change some code that assumes that any title that contains a ":" is not content.) Can I, can I, can I? Please? :-) As far as I can see this actually makes the code simpeler because titles will always have the same format.
Let's consider some arguments though on both sides. Let's think it through!
Larry
From: "Larry Sanger" lsanger@nupedia.com
Let's consider some arguments though on both sides. Let's think it through!
Don't worry. I won't change any code unless I have a feeling there is a consensus about this. But by now I have programmed most of the SQL that deals with this and I think I have a reasonable insight in what the consequences are from a technical point of view. From the users point of view not much changes (unless we want to) except that the namespace-feature that keeps links by default within the namespace is turned back on.
And there should at least be one good counter-argument, shouldn't there? :-)
-- Jan Hidders
From: "Larry Sanger" lsanger@nupedia.com
Let's consider some arguments though on both sides. Let's think it through!
Don't worry. I won't change any code unless I have a feeling there is a consensus about this. But by now I have programmed most of the SQL that deals with this and I think I have a reasonable insight in what the consequences are from a technical point of view. From the users point of view not much changes (unless we want to) except that the namespace-feature that keeps links by default within the namespace is turned back on.
And there should at least be one good counter-argument, shouldn't there? :-)
Yup, there are good counterarguments. The central one is that because the main wikipedia holds the vast majority of information on the site, most links are into the main site, no matter what section the page is in.
Note also that if the change were turned on, all current links would need to be automagically changed.
That said, I'd think it not unreasonable to make the switch, inasmuch as we have the test case of meta.wikipedia.com, where to link to the main site you have to put wp: before the link, which is essentially what we're talking about.
But this should be lower priority than fixing the many, many bugs the current software has.
-tc
From: kband@www.llamacom.com
Yup, there are good counterarguments. The central one is that because the main wikipedia holds the vast majority of information on the site, most links are into the main site, no matter what section the page is in.
Yes, so?
Note also that if the change were turned on, all current links would need to be automagically changed.
No, it wouldn't. I've explained this a couple of postings ago. Because we would turn the switch that redirects links in a namespace that have no specified namespace to the same namspace. We only need to change the page titles and that can be done with a single SQL statement.
But this should be lower priority than fixing the many, many bugs the current software has.
We are working on that. Any bug you would like to go first?
-- Jan Hidders
From: "Jan Hidders" hidders@uia.ua.ac.be
From: kband@www.llamacom.com
Note also that if the change were turned on, all current links would need to be automagically changed.
No, it wouldn't. I've explained this a couple of postings ago. Because we would turn the switch that redirects links in a namespace that have no specified namespace to the same namspace. We only need to change the page titles and that can be done with a single SQL statement.
Hold it, that's not completely true. The links from non-content namspaces to the content namspace would indeed be broken. But that can be remedied with a PHP script that replaces these links.
-- Jan Hidders
From: kband@www.llamacom.com
Yup, there are good counterarguments. The central one is that because the main wikipedia holds the vast majority of information on the site, most links are into the main site, no matter what section the page is in.
Yes, so?
Note also that if the change were turned on, all current links would need to be automagically changed.
No, it wouldn't. I've explained this a couple of postings ago. Because we would turn the switch that redirects links in a namespace that have no specified namespace to the same namspace. We only need to change the page titles and that can be done with a single SQL statement.
This is pretty wikitech, but e.g. on a user page, say there's a link to [[mind control]]. I don't want that to default to [[user:mind control]].
But this should be lower priority than fixing the many, many bugs the current software has.
We are working on that. Any bug you would like to go first?
Here's a short list.
#1 would be restoring the old wikipedia edit history, but I suspect that won't ever happen, so:
#2 www.wikipedia.com homepage doesn't recognize a user as logged in if you link to it from another site or type in the url; only once you go to a www.wikipedia.com/wiki/* page does it recognize you.
#3 Can't edit a redirected entry
#4 Edit conflict on re-direct to nonexistent page
#5 Edit conflict on using back button to edit some more
I recommend setting up a bugzilla.wikipedia.com to make life easier.
--tc
#1 would be restoring the old wikipedia edit history, but I suspect that won't ever happen, so:
That's up to Jimbo (who is pretty busy right now updating the "real" site;)
#2 www.wikipedia.com homepage doesn't recognize a user as logged in if you link to it from another site or type in the url; only once you go to a www.wikipedia.com/wiki/* page does it recognize you.
That's an apache setting, also for Jimbo;)
#3 Can't edit a redirected entry
That's long fixed, the software will be installed today
#4 Edit conflict on re-direct to nonexistent page
Yes, that seems to be still in...
#5 Edit conflict on using back button to edit some more
I thought about that one since the first versions, but I can't figure out how to avoid that. Maybe someone else???
Magnus
#5 Edit conflict on using back button to edit some more
I thought about that one since the first versions, but I can't figure out how to avoid that. Maybe someone else???
The answer is to implement edit conflict testing like it did in UseModWiki. The timestamp shouldn't come from the userside HTML, but from the server. In addition, the software should know that the same user can't conflict with himself.
--tc
On lun, 2002-02-25 at 13:19, Magnus Manske wrote:
#2 www.wikipedia.com homepage doesn't recognize a user as logged in if you link to it from another site or type in the url; only once you go to a www.wikipedia.com/wiki/* page does it recognize you.
That's an apache setting, also for Jimbo;)
This works fine for me. Is the problem specific to certain user accounts?
-- brion vibber (brion @ pobox.com)
From: kband@www.llamacom.com
This is pretty wikitech, but e.g. on a user page, say there's a link to [[mind control]]. I don't want that to default to [[user:mind control]].
Ah, ok. That might be a good point, although I myself wouldn't mind it. The other side of the coin is that I now have to type [[user:Jan Hidders/Sandbox]] instead of [[Jan Hidders/Sandbox]].
#5 Edit conflict on using back button to edit some more
I'm not so sure that's a bug. You can always click on the edit link again if you want to edit some more. I know we could simply arrange things so that a user can never conflict with him or her-self, but that might lead to mistakes like starting to re-edit the penultimate version.
-- Jan Hidders
From: kband@www.llamacom.com
This is pretty wikitech, but e.g. on a user page, say there's a link to [[mind control]]. I don't want that to default to [[user:mind control]].
Ah, ok. That might be a good point, although I myself wouldn't mind it. The other side of the coin is that I now have to type [[user:Jan Hidders/Sandbox]] instead of [[Jan Hidders/Sandbox]].
But you wouldn't have to do that if subpage functionality were turned on for the user pages. Then all you'd have to type is [[/Sandbox]].
#5 Edit conflict on using back button to edit some more
I'm not so sure that's a bug. You can always click on the edit link again if you want to edit some more. I know we could simply arrange things so that a user can never conflict with him or her-self, but that might lead to mistakes like starting to re-edit the penultimate version.
Oh, it's definitely a bug. It was a very useful feature in the UseModWiki software, since it takes much longer to load the edit page from the server than from the local browser cache (and on a smaller but still cumulatively significant point, it's a little bit faster to find the back button or hit the back-page command than to scroll/find the "edit page link").
There's no need to keep users from conflicting with themselves; it's just a matter of fixing the way timestamping is handled. It shouldn't be browserside.
--tc
From: kband@www.llamacom.com
From: kband@www.llamacom.com
This is pretty wikitech, but e.g. on a user page, say there's a link to [[mind control]]. I don't want that to default to [[user:mind control]].
Ah, ok. That might be a good point, although I myself wouldn't mind it.
The
other side of the coin is that I now have to type [[user:Jan Hidders/Sandbox]] instead of [[Jan Hidders/Sandbox]].
But you wouldn't have to do that if subpage functionality were turned on for the user pages. Then all you'd have to type is [[/Sandbox]].
True. But if I'm thinking of future namespaces such as "stable" I'd prefer a consistent behaviour over all namespaces. It is in my experience important that the behaviour of the system is governed by a few clear and simple rules with as little exceptions as possible. It not only makes it easier to build a correct implementation, but also makes it easier to explain to the user.
There's no need to keep users from conflicting with themselves; it's just a matter of fixing the way timestamping is handled. It shouldn't be browserside.
Both the begin-time and the commit-time of the edit are determined server-side at the moment. As I said, changing the behaviour would probably mean that a user never conflicts with him or herself.
-- Jan Hidders
On Mon, 25 Feb 2002, Jan Hidders wrote:
From: kband@www.llamacom.com
Yup, there are good counterarguments. The central one is that because the main wikipedia holds the vast majority of information on the site, most links are into the main site, no matter what section the page is in.
Yes, so?
Well, it's not true for the wikipedia: namespace, though it's probably true for the user: namespace and the talk: namespace.
Larry
At 09:56 AM 2/25/02 -0600, kband wrote:
From: kband@www.llamacom.com
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then the first colon is always the end of the namespace.
That's perfect.
Would this require us to explicitly put "content:" in front of every link?
If so, I doubt it would work--people would forget, because it's not intuitive.
It's not just not intuitive, it makes the user's life harder, and technology should work the other way around--put the burden on the programmers, not the users (editors).
Yeah, I realized this too a little later. It's immaterial from a user perspective what happens behind the stage; the problem is to match user practices with parsability.
Which gets us back to setting up standards for parsable colon usage, which is probably a mostly Wikitech-L issue, if we agree that it's worth it to try to find a way to make colons a (by and large) non-reserved character in titles (and I think we do).
-TC
While I have to agree that it would be nicer to have the use of colons in titles, I don't think it would be better to have "content:" preceding every Wikipedia title. Colonless page titles could be automatically converted to content: titles (so that [[foo]] would be saved automatically as [[content:foo]]), but it would make the system more complicated and more importantly it would make the titles and the URLs uglier.
This is really more of a technical issue than a policy one. I mean, from a policy standpoint, it's great to have the use of colons in titles. If there aren't any technical objections, we should do it. But, of course, there might be technical objections. (Cc'ing wikitech-l.)
Larry
Thus instead of the ":" being a reserved character anywhere in a title, only "user:", "talk:", "wikipedia:" etc. need to be reserved. Any other uses of colons should be fine. This will let us have entries for books with standard formatting of the subtitle (e.g., "The Muggles: A Tale of Woe") or other natural uses of the colon.
There is a simpeler solution. The actual contents of Wikipedia get the namespace "content:". If everything is prefixed with a namespace then the first colon is always the end of the namespace.
That's perfect.
wikipedia-l@lists.wikimedia.org