There's been considerable discussion on wikien-l of rel=nofollow on external links, and how and why those don't apply to the interwiki map.
(This is not wikien-l asserting dominion over meta and the interwiki map :-) It's because en:wp is most of the reason wikipedia.org has a stupidly high Google page rank, and so the SEO spammers whine that it's our job to make the spammers look good to Google, and never mind us or our editors or readers. But anyway.)
It was suggested (and I concur) that if our page rank is so all-fired powerful, that it be turned to the benefit of Free Content, like ourselves. So Jonathan Stokely posted a suggested rewording (http://lists.wikimedia.org/pipermail/wikien-l/2007-May/070344.html), which I reworded a bit and placed here:
http://meta.wikimedia.org/wiki/Talk:Interwiki_map#Inclusion_criteria_clarifi...
Please go there and/or discuss it here.
Precis on the controversy:
http://www.techcrunch.com/2007/04/28/wikipedia-special-treatment-for-wikia-a... (an idiotic TechCrunch story) http://blog.valuewiki.com/2007/04/29/time-to-overhaul-or-abolish-the-interwi... (upset ValueWiki blog post) http://blog.valuewiki.com/2007/04/30/quietly-stepping-down-from-my-high-hors... (slightly embarrassed ValueWiki followup blog post) http://davidgerard.co.uk/notes/2007/04/30/seo-spammers-and-googlemancers/ (my post) http://lists.wikimedia.org/pipermail/wikien-l/2007-April/ (nearer the end) http://lists.wikimedia.org/pipermail/wikien-l/2007-May/ (start)
- d.
On 5/2/07, David Gerard dgerard@gmail.com wrote:
(This is not wikien-l asserting dominion over meta and the interwiki map :-) It's because en:wp is most of the reason wikipedia.org has a stupidly high Google page rank, and so the SEO spammers whine that it's our job to make the spammers look good to Google, and never mind us or our editors or readers. But anyway.)
"The InterWiki Map exists to allow a more efficient syntax for linking between wikis, and thus promote the cooperation and proliferation of wikis and free content. Sites considered for inclusion should probably (1) provide clear and relevant usefulness to the Wikimedia projects (2) be trusted not to encourage spam links being added to the Wikimedia projects (3) be free content (under a Commons-acceptable license) (4) be a wiki (5) be reasonably developed."
If we're really going to rethink what should constitute an IW link, I'd suggest adding "(6) be free from advertising".
"Sites included in the InterWiki Map are considered by the Wikimedia community to be trusted not to encourage spam links being added to Wikimedia projects, and thus "nofollow" is removed from InterWiki links."
I strongly disagree with this, as it confuses the purpose of the IW links. If the only criterion for an IW link was number 2, then it might be reasonable, but this isn't the purpose of IW links.
To quote Anthere: "Either all links (external and interwiki) should be nofollow. Or none of them should be. Or we can work on white lists together for both types of links."
Anthony
On 02/05/07, Anthony wikilegal@inbox.org wrote:
To quote Anthere: "Either all links (external and interwiki) should be nofollow. Or none of them should be. Or we can work on white lists together for both types of links."
As she noted in a followup, that was her speaking as user:Anthere rather than ex cathedra as Chair of the Board of the Wikimedia Foundation. Jimbomancy is bad enough ...
- d.
On 5/2/07, David Gerard dgerard@gmail.com wrote:
On 02/05/07, Anthony wikilegal@inbox.org wrote:
To quote Anthere: "Either all links (external and interwiki) should be nofollow. Or none of them should be. Or we can work on white lists together for both types of links."
As she noted in a followup, that was her speaking as user:Anthere rather than ex cathedra as Chair of the Board of the Wikimedia Foundation. Jimbomancy is bad enough ...
Of course. As Board Chair she has no say anyway, just a vote if the board wishes to pass a resolution.
I still think her quote is quite appropriate. If we're going to turn off nofollow for certain links, a good candidate would be IMDB, for instance. But they can't get an IW link per the criteria you set up, as they're not a wiki. So if we want to whitelist sites, let's create a whitelist for all types of links.
But really I think we should stop doing the job of the search engines.
Anthony
David Gerard wrote:
On 02/05/07, Anthony wikilegal@inbox.org wrote:
To quote Anthere: "Either all links (external and interwiki) should be nofollow. Or none of them should be. Or we can work on white lists together for both types of links."
As she noted in a followup, that was her speaking as user:Anthere rather than ex cathedra as Chair of the Board of the Wikimedia Foundation. Jimbomancy is bad enough ...
I've always been happy to fulfill requests from individual members of the Wikimedia management team. Jimmy has encouraged this. I've done jobs for various members of the board, plus Brad. Of course most of those things were uncontroversial. I'm not blind to the issues here, see my comments on [[m:Talk:Interwiki map]].
-- Tim Starling
David Gerard wrote:
There's been considerable discussion on wikien-l of rel=nofollow on external links, and how and why those don't apply to the interwiki map.
(This is not wikien-l asserting dominion over meta and the interwiki map :-) It's because en:wp is most of the reason wikipedia.org has a stupidly high Google page rank, and so the SEO spammers whine that it's our job to make the spammers look good to Google, and never mind us or our editors or readers. But anyway.)
It was suggested (and I concur) that if our page rank is so all-fired powerful, that it be turned to the benefit of Free Content, like ourselves. So Jonathan Stokely posted a suggested rewording (http://lists.wikimedia.org/pipermail/wikien-l/2007-May/070344.html), which I reworded a bit and placed here:
http://meta.wikimedia.org/wiki/Talk:Interwiki_map#Inclusion_criteria_clarifi...
Please go there and/or discuss it here.
Precis on the controversy:
http://www.techcrunch.com/2007/04/28/wikipedia-special-treatment-for-wikia-a... (an idiotic TechCrunch story) http://blog.valuewiki.com/2007/04/29/time-to-overhaul-or-abolish-the-interwi... (upset ValueWiki blog post) http://blog.valuewiki.com/2007/04/30/quietly-stepping-down-from-my-high-hors... (slightly embarrassed ValueWiki followup blog post) http://davidgerard.co.uk/notes/2007/04/30/seo-spammers-and-googlemancers/ (my post) http://lists.wikimedia.org/pipermail/wikien-l/2007-April/ (nearer the end) http://lists.wikimedia.org/pipermail/wikien-l/2007-May/ (start)
As I stated in the WikiEn thread, the main issue here is the perception of conflict of interest (or the reality of it, according to some).
David's post suggests that we make it so that only links to site supporting free content be exempt from the nofollow attribute. This has some advantages, including being consistent with the goals of the project, and not be apparently arbitrary.
Some issues remain with this approach. Of course, we'd need to agree to some definition of "free content." Would no-commercial-use sites be included? Another issue is the situation where parts of a site are "free content" by whatever definition we use, and other parts are not.
Finally, it does not truly address the perception of conflict of interest.
Now, we can't hope to eliminate everything that someone might consider to be a conflict of interest. It's a question of judgment; cost to benefit, where neither are quantifiable.
As an aside: Given the volume of complaining that the nofollow policy has generated from the SEO community, the so-called "link juice" from Wikipedia must be valuable. Have we considered the possibility of this as a revenue stream?
-Rich
On 02/05/07, Rich Holton richholton@gmail.com wrote:
Some issues remain with this approach. Of course, we'd need to agree to some definition of "free content." Would no-commercial-use sites be included? Another issue is the situation where parts of a site are "free content" by whatever definition we use, and other parts are not.
I was assuming "free content" per http://freedomdefined.org/ , which happens to be the definition we use. (And was written by Erik Moeller.)
Finally, it does not truly address the perception of conflict of interest.
True.
Given the volume of complaining that the nofollow policy has generated from the SEO community, the so-called "link juice" from Wikipedia must be valuable. Have we considered the possibility of this as a revenue stream?
Hah! Sell it? Note, not selling interwiki mappings.
- d.
On 5/2/07, Rich Holton richholton@gmail.com wrote:
As I stated in the WikiEn thread, the main issue here is the perception of conflict of interest (or the reality of it, according to some).
David's post suggests that we make it so that only links to site supporting free content be exempt from the nofollow attribute. This has some advantages, including being consistent with the goals of the project, and not be apparently arbitrary.
But it's obviously just retrofitting these consistent goals to fit the status quo. If we were really going to come up with a system from scratch, to say which links should be nofollowed and which shouldn't, I don't think anyone would suggest the 5 criteria that have been listed.
As an aside: Given the volume of complaining that the nofollow policy has generated from the SEO community, the so-called "link juice" from Wikipedia must be valuable. Have we considered the possibility of this as a revenue stream?
I thought about that, and I came to the conclusion that if Wikimedia derived any significant revenue from such a thing the search engines wouldn't like it. As such they'd probably find a way to either take away the benefit from the links or else lower the rank of Wikipedia itself.
Anthony
Anthony wrote:
On 5/2/07, Rich Holton richholton@gmail.com wrote:
As an aside: Given the volume of complaining that the nofollow policy has generated from the SEO community, the so-called "link juice" from Wikipedia must be valuable. Have we considered the possibility of this as a revenue stream?
I thought about that, and I came to the conclusion that if Wikimedia derived any significant revenue from such a thing the search engines wouldn't like it. As such they'd probably find a way to either take away the benefit from the links or else lower the rank of Wikipedia itself.
I'd be surprised if they don't already special-case Wikipedia, nofollow or not. Nofollow is merely advisory, and a search engine is free to completely ignore it, if they feel ignoring it, or treating it in some other unspecified manner, would improve their search results.
-Mark
On 5/2/07, Delirium delirium@hackish.org wrote:
Anthony wrote:
On 5/2/07, Rich Holton richholton@gmail.com wrote:
As an aside: Given the volume of complaining that the nofollow policy has generated from the SEO community, the so-called "link juice" from Wikipedia must be valuable. Have we considered the possibility of this as a revenue stream?
I thought about that, and I came to the conclusion that if Wikimedia derived any significant revenue from such a thing the search engines wouldn't like it. As such they'd probably find a way to either take away the benefit from the links or else lower the rank of Wikipedia itself.
I'd be surprised if they don't already special-case Wikipedia, nofollow or not. Nofollow is merely advisory, and a search engine is free to completely ignore it, if they feel ignoring it, or treating it in some other unspecified manner, would improve their search results.
I'd be very surprised if they already special-case Wikipedia, as they've said before that they don't do that sort of thing, and it's much cleaner from a design standpoint to not fill up code with special-cases. They also allegedly asked Jimbo to turn nofollow on, which would imply that such a move had an actual effect.
It would make a lot more sense if Google just found a way to detect sites like Wikipedia that were likely to contain user-submitted content, and apply an algorithm intelligent enough to separate the wheat from the chaff (using the length of time the link appeared in the text, for instance). But if they did that, why bother Jimbo about whether or not Wikipedia has enabled nofollow?
Anthony
wikimedia-l@lists.wikimedia.org