Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality: * We're reasonably sure CirrusSearch's language support is better than the current search. [1] * CirrusSearch indexes expanded templates. * CirrusSearch indexes articles within a few seconds of when they are changed. Articles that contain a changed template take longer but they are also updated. * Most of the special search syntax is the same. You can read the syntax here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Hi Nik, Thanks for introducing an option to participate on this testing.
I am Nasir, one of the administrators of Bengali WIkipedia. I would love to volunteer to test this new search feature in my native wiki, Bengali WIkipedia. Please let me know when it will be enabled. So that i can inform my other community members to participate on this testing.
thanks
*-- **Nasir Khan Saikat* http://profiles.google.com/nasir8891 www.nasirkhn.com
On Fri, Oct 18, 2013 at 1:49 AM, Nikolas Everett neverett@wikimedia.orgwrote:
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than the
current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikimedia Sverige and se.wikimedia.org is happy to make search better. Do we need to make anything else to get it enabled or will this email make do?
*Med vänliga hälsningar, Jan Ainali*
Verksamhetschef, Wikimedia Sverige http://se.wikimedia.org/wiki/Huvudsida 0729 - 67 29 48
2013/10/17 Nikolas Everett neverett@wikimedia.org
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than the
current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Hi Nikolas,
One of the grievances I have with the current search system is that it's incapable of offering suggestions with matches in the middle of the title (e.g. when searching for "X", also find "County X (some other words)" if such an article exist). Is CirrusSearch capable of this?
Thanks, Strainu
2013/10/17 Nikolas Everett neverett@wikimedia.org:
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than the
current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Short answer: it's going to but doesn't yet. I'd love to know more about _exactly_ what you mean.
Long answer:
A version of this feature was requested by someone beta testing CirrusSearch on their private wiki. I implemented it as an option turned off by default and off on all WMF wikis but my implementation feels too liberal so I'm sitting on it until I can get an update on exactly what the requester wants. Currently "main p", "page m" "ma pa", "m p", and "p m" all match "Main Page". I'm not sure if that is the right way to do it but you can see how it does fulfil your request when you ask it in the way that you did. Another way to think of it is that "page", "main p", "mai", "m", and "p" should all match "Main Page" but "page m", "m p", and "p m" should not.
There is also the question of how this should be enabled. Should it be changed per wiki? Should it be a user configuration setting? If it is a user configuration setting should its default be configured per wiki? Should it be accessible via some magic prefix typed into the box like "~" does now? Should the prefix invert the default? For the beta tester I was going to add it as a wiki wide setting because the suited his needs.
I've added Dan Garry to this particular reply because he is the WMF Product Manager who is managing this search replacement who might be able to help sort this out.
Nik
On Thu, Oct 17, 2013 at 6:35 PM, Strainu strainu10@gmail.com wrote:
Hi Nikolas,
One of the grievances I have with the current search system is that it's incapable of offering suggestions with matches in the middle of the title (e.g. when searching for "X", also find "County X (some other words)" if such an article exist). Is CirrusSearch capable of this?
Thanks, Strainu
2013/10/17 Nikolas Everett neverett@wikimedia.org:
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and
I've
been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new
search
features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than
the
current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they
are
also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through
the
old search. You'll be able to get search results from the new search
engine
for comparison by adding a url parameter to the search results page. If
you
and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome,
we'll
switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter
similar
to the one that you used to test CirrusSearch. If anything goes wrong
we'll
switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because
CirrusSearch
understands their grammar and old search doesn't. Many other languages
will
see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
2013/10/18 Nikolas Everett neverett@wikimedia.org:
Short answer: it's going to but doesn't yet. I'd love to know more about _exactly_ what you mean.
Thanks for your answer.
On ro.wp we have articles about both the communes and the villages in those communes. The commune often has the name of the main village and it's possible to have several communes with the same name in different counties, so we'll have articles like "Bucșani, Giurgiu", "Bucșani, Dâmbovița", "Comuna Bucșani, Giurgiu", "Comuna Bucșani, Dâmbovița". If I search for "Bucș" I get the first 2 articles, but not the latter two.
Long answer:
A version of this feature was requested by someone beta testing CirrusSearch on their private wiki. I implemented it as an option turned off by default and off on all WMF wikis but my implementation feels too liberal so I'm sitting on it until I can get an update on exactly what the requester wants. Currently "main p", "page m" "ma pa", "m p", and "p m" all match "Main Page". I'm not sure if that is the right way to do it but you can see how it does fulfil your request when you ask it in the way that you did. Another way to think of it is that "page", "main p", "mai", "m", and "p" should all match "Main Page" but "page m", "m p", and "p m" should not.
The last version sound right to me. While google-like inversions would allow far more flexibility, they wouldn't bring much value in the suggestion box, as the number of results is limited. Perhaps enable the inversions only in the results page and place them at the end?
There is also the question of how this should be enabled. Should it be changed per wiki? Should it be a user configuration setting? If it is a user configuration setting should its default be configured per wiki? Should it be accessible via some magic prefix typed into the box like "~" does now? Should the prefix invert the default? For the beta tester I was going to add it as a wiki wide setting because the suited his needs.
Disabled (by default) user config would probably be best. This is an advanced feature and not all users would need it. I'm not sure about the performance impact if enabled on the wmf cluster though.
I've added Dan Garry to this particular reply because he is the WMF Product Manager who is managing this search replacement who might be able to help sort this out.
Nik
On Thu, Oct 17, 2013 at 6:35 PM, Strainu strainu10@gmail.com wrote:
Hi Nikolas,
One of the grievances I have with the current search system is that it's incapable of offering suggestions with matches in the middle of the title (e.g. when searching for "X", also find "County X (some other words)" if such an article exist). Is CirrusSearch capable of this?
Thanks, Strainu
2013/10/17 Nikolas Everett neverett@wikimedia.org:
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than
the current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the
syntax here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and
I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question - sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages - please point me to the right place if so.
- CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example < https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour%3E, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily) return the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me < https://fr.wikisource.org/wiki/Les_mots_tristes%3E ; but http://ur1.ca/fwoytdoes return me < https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!%3E...
Hello,
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
From: jeanfrederic.wiki@gmail.com Date: Fri, 18 Oct 2013 11:32:30 +0200 To: wikitech-ambassadors@lists.wikimedia.org CC: chorohoe@wikimedia.org Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question − sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages − please point me to the right place if so.
* CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily) return the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me https://fr.wikisource.org/wiki/Les_mots_tristes ; but http://ur1.ca/fwoyt does return me https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!…
Hello,
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
From: jeanfrederic.wiki@gmail.com Date: Fri, 18 Oct 2013 11:32:30 +0200 To: wikitech-ambassadors@lists.wikimedia.org CC: chorohoe@wikimedia.org Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question − sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages − please point me to the right place if so.
* CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily) return the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me https://fr.wikisource.org/wiki/Les_mots_tristes ; but http://ur1.ca/fwoyt does return me https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!…
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentation http://sv.wikipedia.org/wiki/Anv%c3%83%c2%a4ndare:Hannibal @aliasHannibal http://twitter.com/AliasHannibal - på Twitter
"*Tänk dig en värld där varje människa på den här planeten får fri tillgång till **världens samlade kunskap*http://sv.wikipedia.org/wiki/Portal:Huvudsida *. Det är vårt mål.*" Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org. You can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like so: Current: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6... Cirrus: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett neverett@wikimedia.orgwrote:
Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentation http://sv.wikipedia.org/wiki/Anv%c3%83%c2%a4ndare:Hannibal @aliasHannibal http://twitter.com/AliasHannibal - på Twitter
"*Tänk dig en värld där varje människa på den här planeten får fri tillgång till **världens samlade kunskap*http://sv.wikipedia.org/wiki/Portal:Huvudsida *. Det är vårt mål.*" Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Nikolas Everett, 22/10/2013 14:24:
I'm requesting another deployment window to do se.wikimedia.org http://se.wikimedia.org and it.wikisource.org http://it.wikisource.org now.
Any other wikis want in?
I've proposed it.wiki to join too, I hope we're on time https://it.wikipedia.org/wiki/Wikipedia:Bar/Discussioni/Lanciamoci_nel_futuro_motore_di_ricerca_interno
Nemo
Federico Leva (Nemo), 22/10/2013 14:56:
Nikolas Everett, 22/10/2013 14:24:
I'm requesting another deployment window to do se.wikimedia.org http://se.wikimedia.org and it.wikisource.org http://it.wikisource.org now.
Any other wikis want in?
I've proposed it.wiki to join too, I hope we're on time https://it.wikipedia.org/wiki/Wikipedia:Bar/Discussioni/Lanciamoci_nel_futuro_motore_di_ricerca_interno
And there's definitely consensus! I hope to see it enabled as soon as possible. :)
Nemo
Hi Nik, all
Now I have some time to spend, so I might assist you to deploy Cirrus Search in Asturian (ast) wiki. I'm mostly a translator, so don't expect more than basic tech skills from me.
Regards -- Xuacu
On Tue, Oct 22, 2013 at 2:24 PM, Nikolas Everett neverett@wikimedia.org wrote:
Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org. You can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like so: Current: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6... Cirrus: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett neverett@wikimedia.org wrote:
Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål." Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
I'd suggest using new search as default for any newly created project. Would that already happen now? For those users don't even have to get used to change. I'd also add recently added vi.Wikivoyage or all wiki voyages . If there are already docs how to create indices I'll help.
Hi Nik, Thanks for enabling this new feature.
I have tested with some other test cases and it is working fine. 2 search system shows 2 set of results. Between these two results i found that the new search results are more useful and most of the cases it show more results than the previous search system.
I will inform the local community and ask them to test as well. I will let you know the feedback here.
thanks Nasir
*-- **Nasir Khan Saikat* http://profiles.google.com/nasir8891 www.nasirkhn.com
On Tue, Oct 22, 2013 at 7:23 PM, Xuacu xuacusk8@gmail.com wrote:
Hi Nik, all
Now I have some time to spend, so I might assist you to deploy Cirrus Search in Asturian (ast) wiki. I'm mostly a translator, so don't expect more than basic tech skills from me.
Regards
Xuacu
On Tue, Oct 22, 2013 at 2:24 PM, Nikolas Everett neverett@wikimedia.org wrote:
Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org.
You
can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like
so:
Current:
http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
Cirrus:
http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett <
neverett@wikimedia.org>
wrote:
Sorry for being confusing. Accents folding is turned on for English
wikis
but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål." Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try
out
a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with
your
j-words. Åland is followed by Alaska, for example. So, as one user
answered
this question, does this function do this better? If it isn't, why
should we
test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Can you please enable it for gu.wp as well? Apologies for late reply On 22 Oct 2013 13:25, "Nikolas Everett" neverett@wikimedia.org wrote:
Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org. You can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like so: Current: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6... Cirrus: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett neverett@wikimedia.orgwrote:
Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentationhttp://sv.wikipedia.org/wiki/Anv%c3%83%c2%a4ndare:Hannibal @aliasHannibal http://twitter.com/AliasHannibal - på Twitter
"*Tänk dig en värld där varje människa på den här planeten får fri tillgång till **världens samlade kunskap*http://sv.wikipedia.org/wiki/Portal:Huvudsida *. Det är vårt mål.*" Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
It looks like that fr.wikisource is also interested: https://fr.wikisource.org/wiki/Wikisource:Scriptorium/Octobre_2013#Tester_le... Sorry for answering late, Thomas
Le 22 oct. 2013 à 19:39, Dhaval S. Vyas dsvyas@gmail.com a écrit :
Can you please enable it for gu.wp as well? Apologies for late reply
On 22 Oct 2013 13:25, "Nikolas Everett" neverett@wikimedia.org wrote: Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org. You can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like so: Current: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6... Cirrus: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett neverett@wikimedia.org wrote: Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote: Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål." Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson l_guldbrandsson@hotmail.com wrote: One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
el.wikipedia would love to test the new search. (Discussion: http://b.geraki.gr/HfjdJ5)
Konstantinos Stampoulis geraki@geraki.gr http://www.geraki.gr ----------------------------------------------------------------------- Οι παραπάνω απόψεις είναι προσωπικές και δεν εκφράζουν παρά μόνο εμένα. Το μήνυμα θεωρείται εμπιστευτικό μόνο εάν το έχω ζητήσει ρητά, διαφορετικά μπορείτε να το χρησιμοποιήσετε σε οποιαδήποτε δημόσια συζήτηση.
2013/10/22 Nikolas Everett neverett@wikimedia.org
Just to round up: bn.wikipedia.org se.wikimedia.org it.wikisource.org have all expressed interest.
Yesterday we enabled CirrusSearch as a secondary for bn.wikipedia.org. You can get its results by performing a normal search and then added 'srbackend=CirrusSearch' to the url. You can compare the results like so: Current: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6... Cirrus: http://bn.wikipedia.org/w/index.php?search=%E0%A6%8F%E0%A6%95%E0%A6%9F%E0%A6...
I'd love to be able to say that it is working well but I can't read any Bengali so I have no idea.
I'm requesting another deployment window to do se.wikimedia.org and it.wikisource.org now.
Any other wikis want in?
Nik
On Fri, Oct 18, 2013 at 10:06 AM, Nikolas Everett neverett@wikimedia.orgwrote:
Sorry for being confusing. Accents folding is turned on for English wikis but off for all other wikis.
On Fri, Oct 18, 2013 at 10:01 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
Sorry, I didn't understand what you meant by "that" in the sentence "Right now that is on for English but off for all other languages."
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg Presentationhttp://sv.wikipedia.org/wiki/Anv%c3%83%c2%a4ndare:Hannibal @aliasHannibal http://twitter.com/AliasHannibal - på Twitter
"*Tänk dig en värld där varje människa på den här planeten får fri tillgång till **världens samlade kunskap*http://sv.wikipedia.org/wiki/Portal:Huvudsida *. Det är vårt mål.*" Jimmy Wales
From: neverett@wikimedia.org Date: Fri, 18 Oct 2013 09:50:59 -0400 To: wikitech-ambassadors@lists.wikimedia.org
Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
On Fri, Oct 18, 2013 at 9:17 AM, Lennart Guldbrandsson < l_guldbrandsson@hotmail.com> wrote:
One problem on Swedish Wikipedia with the current search engine is that it just doesn't do characters like å, ä and ö (which are three common letters in Swedish) well. They are lumped together with a and o respectively, which is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Right now that is on for English but off for all other languages. Some other languages may want it but Swedish certainly isn't one of them.
Nik
_______________________________________________ Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Lennart,
Conversely, for those of us who don't have ready access/knowledge to add the variant characters to our searches specific to languages (even if we know or loops from our chicken scratchings), it becomes difficult to find things in searches unless they are clumped together. For someone non-indigenous to the language finding Åland when I can only type or know Aland means I probably won't find something, which indicates means that there needs to be some wriggle room here.
When working at Wikidata, especially in your non-native language where they are not standard ascii characters is very difficult if you don't have know the characters.
Regards, Billinghurst
On Fri, 18 Oct 2013 13:14:43 +0000, Lennart Guldbrandsson lennart@elementx.se wrote:
Hello,
One problem on Swedish Wikipedia with the current search engine is that
it
just doesn't do characters like å, ä and ö (which are three common
letters
in Swedish) well. They are lumped together with a and o respectively,
which
is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri
tillgång
till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
From: jeanfrederic.wiki@gmail.com Date: Fri, 18 Oct 2013 11:32:30 +0200 To: wikitech-ambassadors@lists.wikimedia.org CC: chorohoe@wikimedia.org Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try
out
a new search?
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question − sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages − please point me to the right place if so.
- CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily)
return
the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me https://fr.wikisource.org/wiki/Les_mots_tristes ; but
does return me https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!…
Conversely, for those of us who don't have ready access/knowledge to add
the variant characters to our searches specific to languages (even if we know or loops from our chicken scratchings), it becomes difficult to find things in searches unless they are clumped together. For someone non-indigenous to the language finding Åland when I can only type or know Aland means I probably won't find something, which indicates means that there needs to be some wriggle room here.
Wriggle room is fine, but should the search be optimized for the people who speak the language or anyone who might pass by? The statistics would probably be considered here. As a non-Mandarin speaking person, I would rather not edit things in Mandarin, because it makes more sense to leave that to people who can actually speak Mandarin.
In Swedish, ÅÄÖ are as much different letters from A and O as W is from V. To lump results from A and Å together will make a lot of Swedes confused. Now, I can manage (though it takes longer to find it), but there is no non-Wikipedian who will find Åland if the search demands that you replace Å with A. And there are of course plenty of exemples of words that differ on that one first letter, for instance Arla (which is a milk company) and Ärla (which is a type of bird).
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri tillgång till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
To: wikitech-ambassadors@lists.wikimedia.org Date: Wed, 12 Mar 2014 11:27:15 +1100 From: billinghurst@gmail.com Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try out a new search?
Lennart,
Conversely, for those of us who don't have ready access/knowledge to add the variant characters to our searches specific to languages (even if we know or loops from our chicken scratchings), it becomes difficult to find things in searches unless they are clumped together. For someone non-indigenous to the language finding Åland when I can only type or know Aland means I probably won't find something, which indicates means that there needs to be some wriggle room here.
When working at Wikidata, especially in your non-native language where they are not standard ascii characters is very difficult if you don't have know the characters.
Regards, Billinghurst
On Fri, 18 Oct 2013 13:14:43 +0000, Lennart Guldbrandsson lennart@elementx.se wrote:
Hello,
One problem on Swedish Wikipedia with the current search engine is that
it
just doesn't do characters like å, ä and ö (which are three common
letters
in Swedish) well. They are lumped together with a and o respectively,
which
is like getting your i-words lumped together with your j-words. Åland is followed by Alaska, for example. So, as one user answered this question, does this function do this better? If it isn't, why should we test it on Swedish Wikipedia? His words, not mine.
Best wishes,
Lennart Guldbrandsson
070 - 207 80 05 http://www.elementx.se - arbete http://www.mrchapel.wordpress.com - personlig blogg
Presentation @aliasHannibal - på Twitter
"Tänk dig en värld där varje människa på den här planeten får fri
tillgång
till världens samlade kunskap. Det är vårt mål."
Jimmy Wales
From: jeanfrederic.wiki@gmail.com Date: Fri, 18 Oct 2013 11:32:30 +0200 To: wikitech-ambassadors@lists.wikimedia.org CC: chorohoe@wikimedia.org Subject: Re: [Wikitech-ambassadors] Would any other wikis like to try
out
a new search?
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question − sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages − please point me to the right place if so.
- CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily)
return
the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me https://fr.wikisource.org/wiki/Les_mots_tristes ; but
does return me https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!…
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
As JeanFred says, Wikisources would *love* a new search engine :-) At it.source we'd like to try yours.
Aubrey
On Fri, Oct 18, 2013 at 11:32 AM, Jean-Frédéric <jeanfrederic.wiki@gmail.com
wrote:
Hi Nik,
I'm looking for volunteer wikis to try out the new search that Chad and
I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
I would have a question - sorry if it was answered already but I did not find it mentionned in the Mediawiki.org pages - please point me to the right place if so.
- CirrusSearch indexes expanded templates.
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
See for example < https://fr.wikisource.org/wiki/On_ne_badine_pas_avec_l%E2%80%99amour%3E, which is merely the super-transclusion of text from the Page: namespace. Search of whatever string from the play text does not (necessarily) return the page of the play itself (though your mileage may vary [1]).
Thanks!
[1] For example, http://ur1.ca/fwoys does not return me < https://fr.wikisource.org/wiki/Les_mots_tristes%3E ; but http://ur1.ca/fwoytdoes return me < https://fr.wikisource.org/wiki/Malheur_%C3%A0_moi_!%3E...
-- Jean-Frédéric
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
On Fri, Oct 18, 2013 at 5:32 AM, Jean-Frédéric jeanfrederic.wiki@gmail.comwrote:
How does CirrusSearch handle cross-namespace transclusions like the one used on Wikisource?
If Mediawiki puts it on the page it we include it. Except things like the table of contents, the [edit] links, and the innards of <video> tags. Otherwise, it is all searchable and highlighted int he search results. We hook into Mediawiki at a really low level so we get things like transcluded categories as well.
el.wikipedia would love to test the new search. (Discussion: http://b.geraki.gr/HfjdJ5)
(I had sent this some days ago but it bounced...)
Konstantinos Stampoulis geraki@geraki.gr http://www.geraki.gr ----------------------------------------------------------------------- Οι παραπάνω απόψεις είναι προσωπικές και δεν εκφράζουν παρά μόνο εμένα. Το μήνυμα θεωρείται εμπιστευτικό μόνο εάν το έχω ζητήσει ρητά, διαφορετικά μπορείτε να το χρησιμοποιήσετε σε οποιαδήποτε δημόσια συζήτηση.
2013/10/17 Nikolas Everett neverett@wikimedia.org
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than the
current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
I'd been meaning to reply to this earlier but all my time has been sucked up by a conference and then getting the flu (my wife insists its just a cold....) Anyway, here is the list I have of interested wikis:
se.wikimedia.org (4865) ast.wikipedia.org (34957) gu.wikipedia.org (44141) el.wikipedia.org (249208) fr.wikisource.org (1505859) ---------------------------------------- nl.wikipedia.org (3171221) it.wikipedia.org (3484831)
We've deployed CirrusSearch as a secondary on bn.wikipedia.org and all wikivoyages. You should be able to test it with the srbackend=CirrusSearch parameter on the search results page. I think bnwiki's community is already trying out CirrusSearch. It'd be wonderful if you could ask the wikivoyage communities to do the same.
I'll see about getting a deployment slot for the wikis above that line. The ones below it are "bigger" and might want to wait until we get the hardware that we've been promised "real soon now." Once those above the line are in I'll measure our headroom again and grab either nlwiki or itwiki in a later deployment window.
Thanks for your patience,
Nik
On Tue, Oct 29, 2013 at 5:08 PM, geraki geraki@gmail.com wrote:
el.wikipedia would love to test the new search. (Discussion: http://b.geraki.gr/HfjdJ5)
(I had sent this some days ago but it bounced...)
Konstantinos Stampoulis geraki@geraki.gr http://www.geraki.gr
Οι παραπάνω απόψεις είναι προσωπικές και δεν εκφράζουν παρά μόνο εμένα. Το μήνυμα θεωρείται εμπιστευτικό μόνο εάν το έχω ζητήσει ρητά, διαφορετικά μπορείτε να το χρησιμοποιήσετε σε οποιαδήποτε δημόσια συζήτηση.
2013/10/17 Nikolas Everett neverett@wikimedia.org
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than
the current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
On Tue, Oct 29, 2013 at 5:08 PM, geraki geraki@gmail.com wrote:
el.wikipedia would love to test the new search. (Discussion: http://b.geraki.gr/HfjdJ5)
(I had sent this some days ago but it bounced...)
Konstantinos Stampoulis geraki@geraki.gr http://www.geraki.gr
Οι παραπάνω απόψεις είναι προσωπικές και δεν εκφράζουν παρά μόνο εμένα. Το μήνυμα θεωρείται εμπιστευτικό μόνο εάν το έχω ζητήσει ρητά, διαφορετικά μπορείτε να το χρησιμοποιήσετε σε οποιαδήποτε δημόσια συζήτηση.
2013/10/17 Nikolas Everett neverett@wikimedia.org
Dear Ambassadors,
I'm looking for volunteer wikis to try out the new search that Chad and I've been working on called CirrusSearch.
<sales pitch>Be a part of the second wave of wikis and influence new search features!</sales pitch>
Reality:
- We're reasonably sure CirrusSearch's language support is better than
the current search. [1]
- CirrusSearch indexes expanded templates.
- CirrusSearch indexes articles within a few seconds of when they are
changed. Articles that contain a changed template take longer but they are also updated.
- Most of the special search syntax is the same. You can read the syntax
here: https://www.mediawiki.org/wiki/Search/CirrusSearchFeatures
What it means to volunteer: If you volunteer your wiki we'll turn CirrusSearch on in "secondary" mode where it'll keep itself up to date but all queries will still go through the old search. You'll be able to get search results from the new search engine for comparison by adding a url parameter to the search results page. If you and the community that you represent aren't immediately blown away by how much better it works we'll work with you to make it awesome.
At some point, shortly after the new search has been deemed awesome, we'll switch CirrusSearch to "primary" mode and all queries will go through it. You'll be able to get at the old search results with a url parameter similar to the one that you used to test CirrusSearch. If anything goes wrong we'll switch you back to the old search. We'll keep that option open for a few months.
So who is ready to help make search better?
Nik Everett
[1]: Some languages (20ish) will see a huge improvement because CirrusSearch understands their grammar and old search doesn't. Many other languages will see an improvement because CirrusSearch is happy to search all kinds of character sets while the current search isn't. Esperanto is very well supported by the old search so would get worse. eo wikis should probably wait until we've improved support.
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
On Wed, Oct 30, 2013 at 5:29 PM, Nikolas Everett neverett@wikimedia.orgwrote:
se.wikimedia.org (4865) ast.wikipedia.org (34957) gu.wikipedia.org (44141) el.wikipedia.org (249208) fr.wikisource.org (1505859)
These wikis now have Cirrus as a secondary. You can test it by performing a regular search and then adding "&srbackend=CirrusSearch" to the url. If you perform another search you'll have to add it back. Please give it a shot and let me know if you have any problems with it. Here is a search in frwikisource: http://fr.wikisource.org/w/index.php?search=Platon&title=Sp%C3%A9cial%3A...
If you find the results better let me know and I'll switch you to primary. If not, please let me know why and I'll fix it.
nl.wikipedia.org (3171221)
I'll do this one a week from yesterday. I should be sending a similar email about it a week from now.
it.wikipedia.org (3484831)
pl.wiktionary.org (409728)
I'll do these November 14th in the evening UTC. I should be sending an email around this time on November 15th about it.
Around that time I'm going to get hungry for more. I'll have new servers with tons of head room and WMF folks will start asking me, "When can we turn off the old search?" and I like to be able to answer that with "When no one uses it. We're adding 6 wikis to Cirrus this week so we'll get there eventually." What do folks think about adding _all_ wikisources and/or all wiktionaries as secondary? I've been told that they really need template tranclusion and so the old search isn't really working for them any way.
Otherwise, I'm looking for more volunteers.
Thanks so much,
Nik
Nikolas Everett, 06/11/2013 15:13:
Around that time I'm going to get hungry for more. I'll have new servers with tons of head room and WMF folks will start asking me, "When can we turn off the old search?" and I like to be able to answer that with "When no one uses it. We're adding 6 wikis to Cirrus this week so we'll get there eventually." What do folks think about adding _all_ wikisources and/or all wiktionaries as secondary? I've been told that they really need template tranclusion and so the old search isn't really working for them any way.
+1 This is more true for Wiktionary than for Wikisource, because many Wiktionaries produce tables of declension etc. via templates and that's actual if not the main content of entries. Additionally, with a similar "volume" it allows you to test in hundreds languages. So maybe Wiktionary first and Wikisource shortly after?
Nemo
it.source is available. if you want, you can send a message to the Wikisource ml, or I can send it for you.
Aubrey
On Wed, Nov 6, 2013 at 3:41 PM, Federico Leva (Nemo) nemowiki@gmail.comwrote:
Nikolas Everett, 06/11/2013 15:13:
Around that time I'm going to get hungry for more. I'll have new
servers with tons of head room and WMF folks will start asking me, "When can we turn off the old search?" and I like to be able to answer that with "When no one uses it. We're adding 6 wikis to Cirrus this week so we'll get there eventually." What do folks think about adding _all_ wikisources and/or all wiktionaries as secondary? I've been told that they really need template tranclusion and so the old search isn't really working for them any way.
+1 This is more true for Wiktionary than for Wikisource, because many Wiktionaries produce tables of declension etc. via templates and that's actual if not the main content of entries. Additionally, with a similar "volume" it allows you to test in hundreds languages. So maybe Wiktionary first and Wikisource shortly after?
Nemo
Wikitech-ambassadors mailing list Wikitech-ambassadors@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-ambassadors
Hoi, When you are interested in more functionality out of search, try combining sources.. like Wikidata and Wikipedia..
I actually blogged about it. http://ultimategerardm.blogspot.nl/2013/11/divcon-search-beyond-tail.html Thanks, Gerard
On 6 November 2013 15:41, Federico Leva (Nemo) nemowiki@gmail.com wrote:
Nikolas Everett, 06/11/2013 15:13:
Around that time I'm going to get hungry for more. I'll have new servers with tons of head room and WMF folks will start asking me, "When can we turn off the old search?" and I like to be able to answer that with "When no one uses it. We're adding 6 wikis to Cirrus this week so we'll get there eventually." What do folks think about adding _all_ wikisources and/or all wiktionaries as secondary? I've been told that they really need template tranclusion and so the old search isn't really working for them any way.
+1 This is more true for Wiktionary than for Wikisource, because many Wiktionaries produce tables of declension etc. via templates and that's actual if not the main content of entries. Additionally, with a similar "volume" it allows you to test in hundreds languages. So maybe Wiktionary first and Wikisource shortly after?
Nemo
Wiktionary-l mailing list Wiktionary-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiktionary-l
wikitech-ambassadors@lists.wikimedia.org