On 1/29/07, Yann Forget yann@forget-me.net wrote:
I think galleries are useful. I do not think they should be deleted. Instead, I think we should make a new namespace called "gallery" and move all the galleries there.
Well I understand, but I am not sure what is the benefit expected here with the deletion of many pages from main namespace, and I don't see how it will be achieved.
We would automatically mass move all the pages with galleries to the gallery namespace. Deletion could also be done automatically.
Commons is generally very badly referenced. I think this is mainly because of the category system (maybe developers could give more hints here). For "Mohandas Gandhi" in Google Images, you won't find any images directly from Commons. That's very surprising seeing that Commons is now the biggest source of free (as in beer) images of Gandhi.
The biggest factor for this is that most search engines will not index pages with names which look like image names, for example http://commons.wikimedia.org/wiki/Image:Foo.jpg. They would index Image:Foo but our image pages are not named like that.
Since most of the time when other sites (including Wikimedia's own) link to commons they link to image pages, we do not gain 'googlejuice' from those links.
There are a lot of things we could do to enhance the popularity of commons, but the page name issue really should be solved first.
It's in the long term plans for mediawiki to support filenames which are unrelated to the file type.... but even using that will require massive renames on commons. Does anyone have any suggestions? There are a lot of possibilities.
Another question is.. are we ready to handle an increase in public visibility?
Gregory Maxwell a écrit :
On 1/29/07, Yann Forget yann@forget-me.net wrote:
I think galleries are useful. I do not think they should be deleted. Instead, I think we should make a new namespace called "gallery" and move all the galleries there.
Well I understand, but I am not sure what is the benefit expected here with the deletion of many pages from main namespace, and I don't see how it will be achieved.
We would automatically mass move all the pages with galleries to the gallery namespace. Deletion could also be done automatically.
I understand what you want to do, but what benefit you expect from such a change.
Commons is generally very badly referenced. I think this is mainly because of the category system (maybe developers could give more hints here). For "Mohandas Gandhi" in Google Images, you won't find any images directly from Commons. That's very surprising seeing that Commons is now the biggest source of free (as in beer) images of Gandhi.
The biggest factor for this is that most search engines will not index pages with names which look like image names, for example http://commons.wikimedia.org/wiki/Image:Foo.jpg. They would index Image:Foo but our image pages are not named like that.
Since most of the time when other sites (including Wikimedia's own) link to commons they link to image pages, we do not gain 'googlejuice' from those links.
There are a lot of things we could do to enhance the popularity of commons, but the page name issue really should be solved first.
It's in the long term plans for mediawiki to support filenames which are unrelated to the file type.... but even using that will require massive renames on commons. Does anyone have any suggestions? There are a lot of possibilities.
Another question is.. are we ready to handle an increase in public visibility?
Good question. Yes, this should be addressed fist obviously.
Regards,
Yann
On 1/29/07, Yann Forget yann@forget-me.net wrote:
We would automatically mass move all the pages with galleries to the gallery namespace. Deletion could also be done automatically.
I understand what you want to do, but what benefit you expect from such a change.
I would expect to see a substantial decrease in junk page creation. I would expect this for two reasons: #The easiest and most attractive location (namespace zero) for junk page creation could be closed to new page creation. #Confusion about the gallery namespace would be removed if it had an explicit name.
Of course, these gains might be possible in other ways... For example, if we shut off non-talk page creation by anons I would expect similar gains, although many of the junk pages are created by named accounts.
On 30/01/07, Gregory Maxwell gmaxwell@gmail.com wrote:
There are a lot of things we could do to enhance the popularity of commons, but the page name issue really should be solved first.
It's in the long term plans for mediawiki to support filenames which are unrelated to the file type.... but even using that will require massive renames on commons. Does anyone have any suggestions? There are a lot of possibilities.
I would imagine some renaming thing would be part of the transition, don't you? Even if it was just Foo.jpg to Foo-jpg or 'Foo jpg'.
I am pretty sure the file storage transition thing will also allow image renaming as a consequence.
The limited details I have found about it are at http://www.mediawiki.org/wiki/MediaWiki_roadmap ("Expand FileStore to all image storage for filename reform, better cache/mirroring ability") and http://www.mediawiki.org/wiki/FileStore .
Another question is.. are we ready to handle an increase in public visibility?
IMO that's a pretty clear "of course not".
Anyway, just wanted to say that I strongly oppose disallowing anon-page creation, at least until SUL has been implemented. After that, I will re-evaluate my position.
Just delete the crud. It's not a disaster, it's just the fact in the life of a completely open wiki.
cheers Brianna
I am pretty sure the file storage transition thing will also allow image renaming as a consequence.
The limited details I have found about it are at http://www.mediawiki.org/wiki/MediaWiki_roadmap ("Expand FileStore to all image storage for filename reform, better cache/mirroring ability") and http://www.mediawiki.org/wiki/FileStore .
I wouldn't expect it before a year after SUL...
It's in the long term plans for mediawiki to support filenames which are unrelated to the file type.... but even using that will require massive renames on commons. Does anyone have any suggestions? There are a lot of possibilities.
I would imagine some renaming thing would be part of the transition, don't you? Even if it was just Foo.jpg to Foo-jpg or 'Foo jpg'.
I don't see why 'massive renames' would be needed (ie. people renaming images). The software treating Image:Foo as the former Image:Foo.png wouldn't be *so* difficult. The script would only have problems with Image:Foo.png and Image:Foo.jpg
Problem with search engines is that they are being too inteligent. Posible solutions:
*Use <a type="text/html" href=".../Image:Foo.png"> **I wasn't expecting it, but the a element *supports* the type attribute. **Validates as XHTML 1.0 Transitional **Search engines probably don't support it, but we are providing the tags.
*Append .html to image pages. **URLs will look ugly /wiki/Image:Foo.png.html
*Use index.php URLs **Either ***/w/index.php?title=Image:Foo.png ***/w/index.php/Image:Foo.png **Not-so-nice urls. **/w/ URLs are currently blocked on robots.txt **Squids would cache less.
Platonides
On 31/01/07, Platonides Platonides@gmail.com wrote:
Problem with search engines is that they are being too inteligent. Posible solutions: *Use <a type="text/html" href=".../Image:Foo.png"> **I wasn't expecting it, but the a element *supports* the type attribute. **Validates as XHTML 1.0 Transitional **Search engines probably don't support it, but we are providing the tags.
Run it past someone suitable at Google, Yahoo and MSN, let Slashdot know so everyone else knows?
The image search engines would undoubtedly want to index our image pages; if we can distinguish image pages ending in .jpg from actual JPEGs in a proper and conformant way, that should I expect be just what they would need to work with.
- d.
Good idea. Who is going to bug the devs?
On 1/31/07, David Gerard dgerard@gmail.com wrote:
On 31/01/07, Platonides Platonides@gmail.com wrote:
Problem with search engines is that they are being too inteligent. Posible solutions: *Use <a type="text/html" href=".../Image:Foo.png"> **I wasn't expecting it, but the a element *supports* the type attribute. **Validates as XHTML 1.0 Transitional **Search engines probably don't support it, but we are providing the tags.
Run it past someone suitable at Google, Yahoo and MSN, let Slashdot know so everyone else knows?
The image search engines would undoubtedly want to index our image pages; if we can distinguish image pages ending in .jpg from actual JPEGs in a proper and conformant way, that should I expect be just what they would need to work with.
- d.
Commons-l mailing list Commons-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/commons-l
On 1/31/07, David Gerard dgerard@gmail.com wrote:
On 31/01/07, Platonides Platonides@gmail.com wrote:
Problem with search engines is that they are being too inteligent. Posible solutions: *Use <a type="text/html" href=".../Image:Foo.png"> **I wasn't expecting it, but the a element *supports* the type attribute. **Validates as XHTML 1.0 Transitional **Search engines probably don't support it, but we are providing the tags.
Run it past someone suitable at Google, Yahoo and MSN, let Slashdot know so everyone else knows?
The image search engines would undoubtedly want to index our image pages; if we can distinguish image pages ending in .jpg from actual JPEGs in a proper and conformant way, that should I expect be just what they would need to work with.
Cute. I like, but it may be akin to moving the earth.
Alternatively, we could also just find some character that we don't accept in file names, but which is valid in URLs, and make it a synonym for . in image page names.. then change our software to link to that form.
I would suggest "/" as a character that we will not be permitting in file names.. (probably not even after the filesystem decoupling of the names)
So http://commons.wikimedia.org/wiki/Image:Ferrofluid_large_spikes.jpg would also be http://commons.wikimedia.org/wiki/Image:Ferrofluid_large_spikes/jpg
I suspect this could be done with a very minor change to mediawiki. (If it's image namespace convert all / to . ... and in image page links convert . to /) ... The question is if we, and more importantly, Brion, think it's just too ugly. :)
On 31/01/07, Gregory Maxwell gmaxwell@gmail.com wrote:
On 1/31/07, David Gerard dgerard@gmail.com wrote:
Run it past someone suitable at Google, Yahoo and MSN, let Slashdot know so everyone else knows? The image search engines would undoubtedly want to index our image pages; if we can distinguish image pages ending in .jpg from actual JPEGs in a proper and conformant way, that should I expect be just what they would need to work with.
Cute. I like, but it may be akin to moving the earth.
Oh, I dunno. "Hi, I'm Greg Maxwell, a developer on MediaWiki, the software that runs Wikipedia. I want to add something to MediaWiki so that our image pages get indexed by Google Images. I expect you'd like that too. How's this idea ... [...] What works best for you, in a standards compliant manner?" Worth a try!
- d.