Re: [Wikidata-l] Auto-transliteration of names of humans
Am 02.05.2015 um 11:32 schrieb Thomas Douillard: Yes, that's what they are for, but wikidata is far from beeing complete, and if there is no label in your language, showing next to the original name a transliteration in your language could prove useful. I can very well imagine situations where it is unclear on how to say a Chinese name, for example. Putting a label in french can somewhat be a hard task, for example. Automatic transliteration is already implemented and used. You will notice if you set your user language to sr-el for instance (Serbian using latin alphabet). However: * Transliteration is currently only supported between a handful of language variants, like the different scripts for Serbian, and various variants of Chinese. * Transliteration (and language fallback in general) is only applied inside statements, not for editable labels, descriptions and aliases at the top of the page. It's unclear hoe language fallback should interact with editing. * Transliteration (and language fallback in general) may not work correctly in qualifiers and references at the moment. So, the general mechanism is already there, the question is just how to improve an apply it. It seems to me that we could use transliteration support for more languages, and that we should figure out a way to apply it to the main label and description shown at the top of the page. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Auto-transliteration of names of humans
Yes, that's what they are for, but wikidata is far from beeing complete, and if there is no label in your language, showing next to the original name a transliteration in your language could prove useful. I can very well imagine situations where it is unclear on how to say a Chinese name, for example. Putting a label in french can somewhat be a hard task, for example. 2015-05-01 11:04 GMT+02:00 Bene* benestar.wikime...@gmail.com: Hi, this is what the monolingual text datatype is for. The labels however are multilingual and should provide users in all languages an idea how the name is said. Best regards Bene Am 01.05.2015 um 07:14 schrieb Gerard Meijssen: Hoi, It is still a bad idea. An official name exists only in one language. Thanks, GerardM On 30 April 2015 at 18:50, Thomas Douillard thomas.douill...@gmail.com wrote: I meant add automatically the transliteration, not replace the name. This is a good candidate : we know for sure the source and the target language (the one of the user) so a good choice for transliteration method is always possible, and we don't pretend it should be the way to say orally the name in the target language. It's just a transliteration of the official name. 2015-04-30 15:14 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, It does not quality anything. It is plain wrong. Thanks, GerardM On 30 April 2015 at 15:06, Joe Filceolaire filceola...@gmail.com wrote: Exactly. The official name property always has the name in the original script. But we can and should have the transliteration in a qualifier. Joe On 30 Apr 2015 06:13, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, We transliterate every name from one script to the other. Transliteration the official name is exactly the one you should not transliterate.. What is left after transliteration is not official. Thanks, GerardM On 29 April 2015 at 18:54, Thomas Douillard thomas.douill...@gmail.com wrote: It's always possible to transliterate the official name https://www.wikidata.org/wiki/Property:P1448property. Of course this should be done by a gadget, or we may have to find a special treatment for the ''name'' properties. 2015-04-28 23:06 GMT+02:00 Joe Filceolaire filceola...@gmail.com: I agree up to a point. Transliteration is not appropriate for labels for all items. There are however a few categories of items for which transliterated labels are appropriate. For example : * English labels for villages and towns * English labels for people *English labels for bands and albums I'm sure there are others that could use this too. Joe On 27 Apr 2015 18:09, Leon Liesener leon.liese...@wikipedia.de wrote: The problem with ISO is that it's a standard for language-independent transliteration to Latin script. Since labels on Wikidata are language-dependent, making use of ISO does not make sense really. If you use ISO for Russian names in Cyrillic script, the label you get is not in English. It's still in Russian but transliterated to Latin script. ISO thus would only fit as an alias for the Russian interface language, if at all. 2015-04-26 22:39 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, grin ISO is a reliable source; it is THE standard /grin Wikipedia is definitely not a standard by its own admission. Thanks, GerardM On 26 April 2015 at 22:37, Yaroslav M. Blanter pute...@mccme.ru wrote: On 2015-04-26 22:33, Gerard Meijssen wrote: Hoi My point is that it is not a given that we should follow any WIkipedia for anything. Also the point of romanisation of Russian is not for the benefit of Russian speakers, it is for the speakers of English. Thanks, GerardM On one hand, yes. On the other hand, no reliable source uses ISO. When NYT writes about a Russian person, they do not use ISO, they use what the English Wikipedia uses or smth similar. In my passport, they do not use ISO (fortunately), why should then ISO be used on Wikidata in an entry about me? Cheers Yaroslav ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Auto-transliteration of names of humans
Hoi, When the point is to express how an official name is to be pronounced, IPA is in order not a text in another script. Thanks, GerardM On 1 May 2015 at 11:04, Bene* benestar.wikime...@gmail.com wrote: Hi, this is what the monolingual text datatype is for. The labels however are multilingual and should provide users in all languages an idea how the name is said. Best regards Bene Am 01.05.2015 um 07:14 schrieb Gerard Meijssen: Hoi, It is still a bad idea. An official name exists only in one language. Thanks, GerardM On 30 April 2015 at 18:50, Thomas Douillard thomas.douill...@gmail.com wrote: I meant add automatically the transliteration, not replace the name. This is a good candidate : we know for sure the source and the target language (the one of the user) so a good choice for transliteration method is always possible, and we don't pretend it should be the way to say orally the name in the target language. It's just a transliteration of the official name. 2015-04-30 15:14 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, It does not quality anything. It is plain wrong. Thanks, GerardM On 30 April 2015 at 15:06, Joe Filceolaire filceola...@gmail.com wrote: Exactly. The official name property always has the name in the original script. But we can and should have the transliteration in a qualifier. Joe On 30 Apr 2015 06:13, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, We transliterate every name from one script to the other. Transliteration the official name is exactly the one you should not transliterate.. What is left after transliteration is not official. Thanks, GerardM On 29 April 2015 at 18:54, Thomas Douillard thomas.douill...@gmail.com wrote: It's always possible to transliterate the official name https://www.wikidata.org/wiki/Property:P1448property. Of course this should be done by a gadget, or we may have to find a special treatment for the ''name'' properties. 2015-04-28 23:06 GMT+02:00 Joe Filceolaire filceola...@gmail.com: I agree up to a point. Transliteration is not appropriate for labels for all items. There are however a few categories of items for which transliterated labels are appropriate. For example : * English labels for villages and towns * English labels for people *English labels for bands and albums I'm sure there are others that could use this too. Joe On 27 Apr 2015 18:09, Leon Liesener leon.liese...@wikipedia.de wrote: The problem with ISO is that it's a standard for language-independent transliteration to Latin script. Since labels on Wikidata are language-dependent, making use of ISO does not make sense really. If you use ISO for Russian names in Cyrillic script, the label you get is not in English. It's still in Russian but transliterated to Latin script. ISO thus would only fit as an alias for the Russian interface language, if at all. 2015-04-26 22:39 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, grin ISO is a reliable source; it is THE standard /grin Wikipedia is definitely not a standard by its own admission. Thanks, GerardM On 26 April 2015 at 22:37, Yaroslav M. Blanter pute...@mccme.ru wrote: On 2015-04-26 22:33, Gerard Meijssen wrote: Hoi My point is that it is not a given that we should follow any WIkipedia for anything. Also the point of romanisation of Russian is not for the benefit of Russian speakers, it is for the speakers of English. Thanks, GerardM On one hand, yes. On the other hand, no reliable source uses ISO. When NYT writes about a Russian person, they do not use ISO, they use what the English Wikipedia uses or smth similar. In my passport, they do not use ISO (fortunately), why should then ISO be used on Wikidata in an entry about me? Cheers Yaroslav ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list
Re: [Wikidata-l] Auto-transliteration of names of humans
I'll all with you, on this, unfortunately only a few people reads IPA currently :(. I don't, for example. 2015-05-02 15:19 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, When the point is to express how an official name is to be pronounced, IPA is in order not a text in another script. Thanks, GerardM On 1 May 2015 at 11:04, Bene* benestar.wikime...@gmail.com wrote: Hi, this is what the monolingual text datatype is for. The labels however are multilingual and should provide users in all languages an idea how the name is said. Best regards Bene Am 01.05.2015 um 07:14 schrieb Gerard Meijssen: Hoi, It is still a bad idea. An official name exists only in one language. Thanks, GerardM On 30 April 2015 at 18:50, Thomas Douillard thomas.douill...@gmail.com wrote: I meant add automatically the transliteration, not replace the name. This is a good candidate : we know for sure the source and the target language (the one of the user) so a good choice for transliteration method is always possible, and we don't pretend it should be the way to say orally the name in the target language. It's just a transliteration of the official name. 2015-04-30 15:14 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, It does not quality anything. It is plain wrong. Thanks, GerardM On 30 April 2015 at 15:06, Joe Filceolaire filceola...@gmail.com wrote: Exactly. The official name property always has the name in the original script. But we can and should have the transliteration in a qualifier. Joe On 30 Apr 2015 06:13, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, We transliterate every name from one script to the other. Transliteration the official name is exactly the one you should not transliterate.. What is left after transliteration is not official. Thanks, GerardM On 29 April 2015 at 18:54, Thomas Douillard thomas.douill...@gmail.com wrote: It's always possible to transliterate the official name https://www.wikidata.org/wiki/Property:P1448property. Of course this should be done by a gadget, or we may have to find a special treatment for the ''name'' properties. 2015-04-28 23:06 GMT+02:00 Joe Filceolaire filceola...@gmail.com: I agree up to a point. Transliteration is not appropriate for labels for all items. There are however a few categories of items for which transliterated labels are appropriate. For example : * English labels for villages and towns * English labels for people *English labels for bands and albums I'm sure there are others that could use this too. Joe On 27 Apr 2015 18:09, Leon Liesener leon.liese...@wikipedia.de wrote: The problem with ISO is that it's a standard for language-independent transliteration to Latin script. Since labels on Wikidata are language-dependent, making use of ISO does not make sense really. If you use ISO for Russian names in Cyrillic script, the label you get is not in English. It's still in Russian but transliterated to Latin script. ISO thus would only fit as an alias for the Russian interface language, if at all. 2015-04-26 22:39 GMT+02:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, grin ISO is a reliable source; it is THE standard /grin Wikipedia is definitely not a standard by its own admission. Thanks, GerardM On 26 April 2015 at 22:37, Yaroslav M. Blanter pute...@mccme.ru wrote: On 2015-04-26 22:33, Gerard Meijssen wrote: Hoi My point is that it is not a given that we should follow any WIkipedia for anything. Also the point of romanisation of Russian is not for the benefit of Russian speakers, it is for the speakers of English. Thanks, GerardM On one hand, yes. On the other hand, no reliable source uses ISO. When NYT writes about a Russian person, they do not use ISO, they use what the English Wikipedia uses or smth similar. In my passport, they do not use ISO (fortunately), why should then ISO be used on Wikidata in an entry about me? Cheers Yaroslav ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___
Re: [Wikidata-l] Auto-transliteration of names of humans
I think it would be nice if the *official name* property could have a special treatment in the UI. Something like the way you plan to sort out ids out of the rest of the statements :) But I have no idea. It's an important things, the name the local people gives to a place or a thing, it may be on road signs for example. It's kind of the Main Name. 2015-05-02 11:54 GMT+02:00 Daniel Kinzler daniel.kinz...@wikimedia.de: Am 02.05.2015 um 11:32 schrieb Thomas Douillard: Yes, that's what they are for, but wikidata is far from beeing complete, and if there is no label in your language, showing next to the original name a transliteration in your language could prove useful. I can very well imagine situations where it is unclear on how to say a Chinese name, for example. Putting a label in french can somewhat be a hard task, for example. Automatic transliteration is already implemented and used. You will notice if you set your user language to sr-el for instance (Serbian using latin alphabet). However: * Transliteration is currently only supported between a handful of language variants, like the different scripts for Serbian, and various variants of Chinese. * Transliteration (and language fallback in general) is only applied inside statements, not for editable labels, descriptions and aliases at the top of the page. It's unclear hoe language fallback should interact with editing. * Transliteration (and language fallback in general) may not work correctly in qualifiers and references at the moment. So, the general mechanism is already there, the question is just how to improve an apply it. It seems to me that we could use transliteration support for more languages, and that we should figure out a way to apply it to the main label and description shown at the top of the page. -- Daniel Kinzler Senior Software Developer Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V. ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l
Re: [Wikidata-l] Number of planets in the solar system
Hoi, It strikes me as another example of a search for perfection where we do not even cater for what is good. Our priorities should be with what is common and present it well not with a game of trivia that upset showing what is good and common. Thanks, GerardM On 30 April 2015 at 18:33, Paul Houle ontolo...@gmail.com wrote: @Thomas is close to the right answer. Nothing about Pluto changed, it was the definition of Planet that is changed so you need two different definitions of Planets, but note that the definitions of themselves are somewhat timeless, so you are really pointing to some specific definition of a planet in either case. There is no reason why this is not practical. It is just a matter of putting in another type, and maintenance is not a tough problem since there are fewer than 10 of them. There could be some need for vocabulary to describe the attributes of the definitions, but simply a link to a defining document is good enough from the viewpoint of grounding. On Thu, Apr 30, 2015 at 12:20 PM, Thomas Douillard thomas.douill...@gmail.com wrote: It may not be practical, but it is still possible ;) classes like ''astronomic corp that was thought to be a planet in 1850'' are an option :) 2015-04-30 13:51 GMT+02:00 Andrew Gray andrew.g...@dunelm.org.uk: On 30 April 2015 at 12:37, Thomas Douillard thomas.douill...@gmail.com wrote: Infovarius even complicated the problem, he put the number of known planets at some time with a qualifier for validity :) Just to throw a real spanner in the works: for a lot of the nineteenth century the number varied widely. The eighth planet was discovered in 1801, and is what we'd now think of as the asteroid or dwarf planet Ceres; the real eighth planet, Neptune, wasn't discovered until 1851. Newly discovered asteroids were thought of as 'planets' for some time (I have an 1843 schoolbook somewhere that confidently tells children there were eleven planets...) until by about 1850, it became clear that having twenty or so very small planets with more discovered every year was confusing, and the meaning of the word shifted. There was no formal agreement (as was the case in 2006) so no specific end date. The moral of this story is probably that trying to express complex things in Wikidata is not always practical :-) -- - Andrew Gray andrew.g...@dunelm.org.uk ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l -- Paul Houle *Applying Schemas for Natural Language Processing, Distributed Systems, Classification and Text Mining and Data Lakes* (607) 539 6254paul.houle on Skype ontolo...@gmail.com https://legalentityidentifier.info/lei/lookup http://legalentityidentifier.info/lei/lookup ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l ___ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l