For years it was slightly bugging me that the "Most wanted articles" function is deactivated. On en.wp, the last cache was generated in October 2009.
Well, I cooked an alternative. Listing all non-existing articles that have at least 10 links from article namespace, updated daily:
http://toolserver.org/~magnus/most_wanted.php
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Cheers, Magnus
On 6 February 2011 22:19, Magnus Manske magnusmanske@googlemail.com wrote:
http://toolserver.org/~magnus/most_wanted.php Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Yes, I wonder who the German military buff is ... hitting "random" helps avoid that sort of thing :-)
Another link for people who want to realise they're not that bored after all: http://en.wikipedia.org/wiki/Wikipedia:WikiProject_Missing_encyclopedic_arti...
The low-hanging fruit is gone, now there's something slightly resembling work to be done.
- d.
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about.
Can *anyone*, even in principle, do something about that? It really bugs me that the "what links here" function doesn't distinguish between links arising from templates (often not directly relevant) and links directly from the article wiki-text. If the answer is something to do with parsers, please do explain!
Carcharoth
On 07/02/11 10:56, Carcharoth wrote:
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about.
Can *anyone*, even in principle, do something about that? It really bugs me that the "what links here" function doesn't distinguish between links arising from templates (often not directly relevant) and links directly from the article wiki-text. If the answer is something to do with parsers, please do explain!
Yes, it's possible. It was necessary to register links from templates in the pagelinks table so that when a page is deleted or created, the HTML caches can be updated so that the link colour will change. With a schema change and some parser work, it would be possible to flag such links so that they are optional in "what links here".
-- Tim Starling
It could be possible to only count real links and not template links by doing this: *List all the links from "What links here" from "Example_article" *Enter to the ones that are in the Template: mainspace *Count the number of links in "What links here" in each template *Then number of real links = total links - links from templates
Maybe someone could code something like that. Also the program should check add all the "what links here" from each template in a list, and delete repeated articles (because one article may have two templates both linking at other article)
On Sun, Feb 6, 2011 at 10:34 PM, Tim Starling tstarling@wikimedia.org wrote:
On 07/02/11 10:56, Carcharoth wrote:
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about.
Can *anyone*, even in principle, do something about that? It really bugs me that the "what links here" function doesn't distinguish between links arising from templates (often not directly relevant) and links directly from the article wiki-text. If the answer is something to do with parsers, please do explain!
Yes, it's possible. It was necessary to register links from templates in the pagelinks table so that when a page is deleted or created, the HTML caches can be updated so that the link colour will change. With a schema change and some parser work, it would be possible to flag such links so that they are optional in "what links here".
-- Tim Starling
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On Mon, Feb 7, 2011 at 1:34 AM, Tim Starling tstarling@wikimedia.org wrote:
On 07/02/11 10:56, Carcharoth wrote:
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about.
Can *anyone*, even in principle, do something about that? It really bugs me that the "what links here" function doesn't distinguish between links arising from templates (often not directly relevant) and links directly from the article wiki-text. If the answer is something to do with parsers, please do explain!
Yes, it's possible. It was necessary to register links from templates in the pagelinks table so that when a page is deleted or created, the HTML caches can be updated so that the link colour will change. With a schema change and some parser work, it would be possible to flag such links so that they are optional in "what links here".
That would be wonderful. It might even get me to create a bugzilla account to vote for a bug if there is one open on this...(of course, one problem is still that some templates are relevant to article content and some are not - the ones that generate distracting links are the navigational ones that tend to be at the bottom of pages, the footer templates - and I'm not sure if infobox links would count as template links or not - they are generated from parsing of a template parameter, but don't appear in the template itself, unlike the footer navboxes).
[In case anyone is confused, an example is the massive footer templates that can lead to Nobel prize winners decades apart linking to each other, or diverse topics within a broad area linking to each other, though only through templates and not in the text. Oh, and some links appear in both footer templates, infoboxes, and the article 'text'. Not sure how that is handled.]
Carcharoth
agreed. The footer templates are the biggest source of linkage bloat. the templates are useful, and we need some way of keeping track of what should be in them when we add or delete articles, but they make working with what links here for any practical purpose extremely difficult. They'd be much more helpful if they were separated.
On Sun, Feb 6, 2011 at 9:52 PM, Carcharoth carcharothwp@googlemail.com wrote:
On Mon, Feb 7, 2011 at 1:34 AM, Tim Starling tstarling@wikimedia.org wrote:
On 07/02/11 10:56, Carcharoth wrote:
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about.
Can *anyone*, even in principle, do something about that? It really bugs me that the "what links here" function doesn't distinguish between links arising from templates (often not directly relevant) and links directly from the article wiki-text. If the answer is something to do with parsers, please do explain!
Yes, it's possible. It was necessary to register links from templates in the pagelinks table so that when a page is deleted or created, the HTML caches can be updated so that the link colour will change. With a schema change and some parser work, it would be possible to flag such links so that they are optional in "what links here".
That would be wonderful. It might even get me to create a bugzilla account to vote for a bug if there is one open on this...(of course, one problem is still that some templates are relevant to article content and some are not - the ones that generate distracting links are the navigational ones that tend to be at the bottom of pages, the footer templates - and I'm not sure if infobox links would count as template links or not - they are generated from parsing of a template parameter, but don't appear in the template itself, unlike the footer navboxes).
[In case anyone is confused, an example is the massive footer templates that can lead to Nobel prize winners decades apart linking to each other, or diverse topics within a broad area linking to each other, though only through templates and not in the text. Oh, and some links appear in both footer templates, infoboxes, and the article 'text'. Not sure how that is handled.]
Carcharoth
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On 7 February 2011 04:10, David Goodman dggenwp@gmail.com wrote:
agreed. The footer templates are the biggest source of linkage bloat. the templates are useful, and we need some way of keeping track of what should be in them when we add or delete articles, but they make working with what links here for any practical purpose extremely difficult. They'd be much more helpful if they were separated.
Currently, in whatlinkshere, links from redirects are shown indented as subsets of the redirect, something like:
* Article1 * Article2 * Article3 * Redirect * * Article4 * * Article5 * Article6 * Article7
Something like this for template links would be an ideal solution, I think - it'd let us see the template links, which are still useful for various purposes, whilst keeping them distinct from "real" links. I am not sure how much more of a logistic hassle this would be, though...
On Sun, Feb 6, 2011 at 2:19 PM, Magnus Manske magnusmanske@googlemail.comwrote:
For years it was slightly bugging me that the "Most wanted articles"
function is deactivated. On en.wp, the last cache was generated in October 2009.
Well, I cooked an alternative. Listing all non-existing articles that have at least 10 links from article namespace, updated daily:
http://toolserver.org/~magnus/most_wanted.php
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Cheers, Magnus
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
Thanks for doing this Magnus! It's fascinating and really helpful to see trends in the list.
On 06/02/2011, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Is that necessarily a problem in this case?
We still will have pages, that are potentially clickable from a large number of articles, that don't exist. Even if they're linked from a template, they're still likely to be fairly important to be in a widely used template.
And google would probably tend to consider these articles as important if they existed.
Cheers, Magnus
On 07/02/2011 15:38, Ian Woollard wrote:
On 06/02/2011, Magnus Manskemagnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Is that necessarily a problem in this case?
We still will have pages, that are potentially clickable from a large number of articles, that don't exist. Even if they're linked from a template, they're still likely to be fairly important to be in a widely used template.
Yes, it is a practical problem, in that priority will tend to be given to redlinks that are in navboxes enthusiastically applied to pages, rather than potential articles referred to in text (which is a much better test of "wantedness"). It is hard to care so much about such redlinks - the navbox in question may well be doing no more than mirroring a category, and we don't care that much about "articles missing from categories". Put it this way: if the navbox were a list, it would contribute once to counting such a redlink, which wouldn't distort the priorities in the same fashion.
Charles
On Mon, Feb 7, 2011 at 3:53 PM, Charles Matthews charles.r.matthews@ntlworld.com wrote:
Put it this way: if the navbox were a list, it would contribute once to counting such a redlink, which wouldn't distort the priorities in the same fashion.
Indeed. And navbox footer templates are effectively just lists/categories transcluded at the bottom of articles. Templates, lists and categories are all ways of organising the same information, but they interact in different ways with the software.
http://en.wikipedia.org/wiki/Wikipedia:Categories,_lists,_and_navigation_tem...
Carcharoth
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
I took a closer look, and realised what you mean.
http://en.wikipedia.org/wiki/Template:KCwithOL
That template alone accounts for hundreds of links. Uncollapse the sections to see how many there are. Maybe there should be a limit to the number of links allows in a single template?
Carcharoth
On 7 February 2011 17:46, Carcharoth carcharothwp@googlemail.com wrote:
http://en.wikipedia.org/wiki/Template:KCwithOL
That template alone accounts for hundreds of links. Uncollapse the sections to see how many there are. Maybe there should be a limit to the number of links allows in a single template?
Are we not risking throwing the baby out with the bathwater, here? whatlinkshere is a useful tool, but so are large templates - limiting the latter to benefit the former seems a bit of a zero-sum game.
On 7 February 2011 19:26, Andrew Gray andrew.gray@dunelm.org.uk wrote:
On 7 February 2011 17:46, Carcharoth carcharothwp@googlemail.com wrote:
http://en.wikipedia.org/wiki/Template:KCwithOL That template alone accounts for hundreds of links. Uncollapse the sections to see how many there are. Maybe there should be a limit to the number of links allows in a single template?
Are we not risking throwing the baby out with the bathwater, here? whatlinkshere is a useful tool, but so are large templates - limiting the latter to benefit the former seems a bit of a zero-sum game.
Indeed. If a tool turns out inadequate, we should aim to fix the tool rather than hamper the content to make the tool's life easier. That template's a *monster*, but it's an interesting exercise in new ways of presenting WIkipedia content, and is quite usable for the reader.
- d.
On Mon, Feb 7, 2011 at 7:26 PM, Andrew Gray andrew.gray@dunelm.org.uk wrote:
On 7 February 2011 17:46, Carcharoth carcharothwp@googlemail.com wrote:
http://en.wikipedia.org/wiki/Template:KCwithOL
That template alone accounts for hundreds of links. Uncollapse the sections to see how many there are. Maybe there should be a limit to the number of links allows in a single template?
Are we not risking throwing the baby out with the bathwater, here? whatlinkshere is a useful tool, but so are large templates - limiting the latter to benefit the former seems a bit of a zero-sum game.
Oh, if there was a way of separating out the links from the templates from the other links, I'd have no problems with it. But at the moment, try going through the list here and tell me which come from the template and which come from elsewhere?
http://en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Fritz_Bayerl...
In extreme cases, I've come across articles linked only from templates and not from anywhere else. They *should* be linked from other articles (i.e. from the actual text of other articles), but the template links swamp things, meaning you can't see the problem.
Carcharoth
On Sun, Feb 6, 2011 at 10:19 PM, Magnus Manske magnusmanske@googlemail.com wrote:
For years it was slightly bugging me that the "Most wanted articles" function is deactivated. On en.wp, the last cache was generated in October 2009.
Well, I cooked an alternative. Listing all non-existing articles that have at least 10 links from article namespace, updated daily:
http://toolserver.org/~magnus/most_wanted.php
Many of these links are due to templates, which I can do little about. I hope it will still be useful to some.
Could you do a version of this where the tool also lists the number of links from Template namespace? This is namespace 10, as opposed to namespace 0 (articles). And then allow an option (sorting or eliminating) to only look at "wanted" articles with no links from template namespace?
For example:
http://en.wikipedia.org/w/index.php?title=Special%3AWhatLinksHere&target...
That is 'British films of 2011', with 2 entries in templates.
http://en.wikipedia.org/w/index.php?title=Special%3AWhatLinksHere&target...
But Richard Kruse is not linked from templates, and the links appear to be mostly from lists.
Carcharoth