I have some doubts about the identifiers. How will we distinguish if monument number 50 is from Nederland or Andorra? Perhaps should it be identified with a prefix (NL-50, AD-50), or perhaps is there any solution in mind?
Moroever, in Spain we can find different formats in different sources: "(R.I.) - 51 - 0010580 - 00000", "R-I-51-0010580", and others. Obviously we need to unify to an unique format and we are discussing it with wm-es, but I would appreciate some inputs. "R.I." or "R-I" stands for Registro de Inmuebles (buildings' register) and "51" is the category for monuments. Probably it is not a good idea to ask participants to type a complicated number format with spaces, zeros, etc. One solution is to reduce it to its numeric form "10580" or perhaps to construct a simplified format like "RI510010580". What are the technical implications about identifiers, and which would be the best approach?
Vicenç
Hi Vicenç. Please consider the concept of "unique identifier"[1]. The only purpose it has is that it must be able to uniquely identify an entity, in this case a momument. A uniquely identified monument can have multiple properties. One of those could be the country it is in (or normalised a country id which is the unique identifier of a country name), or a reference to a unique identifier in a country's system, a municipality's system or an autonomous region's system for monuments.
Unique identifiers are completely trivial and the only requirement is that they are unique and that you do not run out of them before your collection is complete. Unique identifiers are usually optimised for data processing, and databases index integers usually pretty well.
Cheers!
Siebrand
[1] http://en.wikipedia.org/wiki/Unique_identifier
Van: Vicenç Riullop vriullop@hotmail.com Beantwoorden - Aan: Wiki Loves Monuments Photograph Competition wikilovesmonuments@lists.wikimedia.org Datum: Sun, 13 Mar 2011 16:01:02 +0000 Aan: wikilovesmonuments@lists.wikimedia.org Onderwerp: [Wiki Loves Monuments] Unique identifiers
I have some doubts about the identifiers. How will we distinguish if monument number 50 is from Nederland or Andorra? Perhaps should it be identified with a prefix (NL-50, AD-50), or perhaps is there any solution in mind?
Moroever, in Spain we can find different formats in different sources: "(R.I.) - 51 - 0010580 - 00000", "R-I-51-0010580", and others. Obviously we need to unify to an unique format and we are discussing it with wm-es, but I would appreciate some inputs. "R.I." or "R-I" stands for Registro de Inmuebles (buildings' register) and "51" is the category for monuments. Probably it is not a good idea to ask participants to type a complicated number format with spaces, zeros, etc. One solution is to reduce it to its numeric form "10580" or perhaps to construct a simplified format like "RI510010580". What are the technical implications about identifiers, and which would be the best approach?
Vicenç: The "unique id" is only relevant in the context/table of a country, because the aggregated/overall table would be a merge of all "country tables". In that table, the ID would have always to be a compound key like "countrycode,id".
It seems, however, for Spain you shouldn't use integers. I personally don't like reducing "R-I-51-0010580" to the integer "10580", unless you are pretty sure "R-I-51" can be inferred from other data (like "R-I" being a classification, 51 being the "comunidad autonoma" identifier, etc).
And, from what I've seen so far, it won't work either for Portugal, as it seems we will face a similar issue whenever I merge "Madeira" and "Azores" island heritage.
Reference page: http://commons.wikimedia.org/wiki/Commons:Wiki_Loves_Monuments_2011/Monument...
Oh, that is a problem in Germany too. The greatest we can dream of are unique ids in a state. But within a state, they often have ids only within a municipality. Ziko
2011/3/13 Nuno Tavares nuno.tavares@wikimedia.pt:
Vicenç: The "unique id" is only relevant in the context/table of a country, because the aggregated/overall table would be a merge of all "country tables". In that table, the ID would have always to be a compound key like "countrycode,id".
It seems, however, for Spain you shouldn't use integers. I personally don't like reducing "R-I-51-0010580" to the integer "10580", unless you are pretty sure "R-I-51" can be inferred from other data (like "R-I" being a classification, 51 being the "comunidad autonoma" identifier, etc).
And, from what I've seen so far, it won't work either for Portugal, as it seems we will face a similar issue whenever I merge "Madeira" and "Azores" island heritage.
Reference page: http://commons.wikimedia.org/wiki/Commons:Wiki_Loves_Monuments_2011/Monument...
-- Nuno Tavares Wikimedia Portugal http://www.wikimedia.pt
Imagine um mundo onde é dada a qualquer pessoa a possibilidade de ter livre acesso ao somatório de todo o conhecimento humano. É isso o que estamos a fazer.
Participe também: http://www.wikimedia.pt
Dom, 2011-03-13 às 20:13 +0100, Siebrand Mazeland escreveu:
Hi Vicenç. Please consider the concept of "unique identifier"[1]. The only purpose it has is that it must be able to uniquely identify an entity, in this case a momument. A uniquely identified monument can have multiple properties. One of those could be the country it is in (or normalised a country id which is the unique identifier of a country name), or a reference to a unique identifier in a country's system, a municipality's system or an autonomous region's system for monuments.
Unique identifiers are completely trivial and the only requirement is that they are unique and that you do not run out of them before your collection is complete. Unique identifiers are usually optimised for data processing, and databases index integers usually pretty well.
Cheers!
Siebrand
[1] http://en.wikipedia.org/wiki/Unique_identifier
Van: Vicenç Riullop vriullop@hotmail.com Beantwoorden - Aan: Wiki Loves Monuments Photograph Competition wikilovesmonuments@lists.wikimedia.org Datum: Sun, 13 Mar 2011 16:01:02 +0000 Aan: wikilovesmonuments@lists.wikimedia.org Onderwerp: [Wiki Loves Monuments] Unique identifiers
I have some doubts about the identifiers. How will we distinguish if monument number 50 is from Nederland or Andorra? Perhaps should it be identified with a prefix (NL-50, AD-50), or perhaps is there any solution in mind?
Moroever, in Spain we can find different formats in different sources: "(R.I.) - 51 - 0010580 - 00000", "R-I-51-0010580", and others. Obviously we need to unify to an unique format and we are discussing it with wm-es, but I would appreciate some inputs. "R.I." or "R-I" stands for Registro de Inmuebles (buildings' register) and "51" is the category for monuments. Probably it is not a good idea to ask participants to type acomplicated number format with spaces, zeros, etc. One solution is to reduce it to its numeric form "10580" or perhaps to construct a simplified format like "RI510010580". What are the technical implications about identifiers, and which would be the best approach?
Wiki Loves Monuments mailing list WikiLovesMonuments@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikilovesmonuments http://www.wikilovesmonuments.eu
Wiki Loves Monuments mailing list WikiLovesMonuments@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikilovesmonuments http://www.wikilovesmonuments.eu
Hi all, I don't know if it can be of any help but in France, every monument has an id (or some times more than one if different parts of a monument are classified), but every monument can be identified by at least one id.
These ids aren't integer, they look like PA00083476.
Sylvain
2011/3/13 Ziko van Dijk zvandijk@googlemail.com
Oh, that is a problem in Germany too. The greatest we can dream of are unique ids in a state. But within a state, they often have ids only within a municipality. Ziko
I think I have got it. The special upload form may not be as simple as last year [1] with a field for Rijksmonumentnummer. For Germany, i.e., the user should first select country, state, municipality, and then type the number, unless the identifier includes some codification. For Spain, if all monuments have the same prefix R-I-51, then this part is not relevant as an identifier and it will be easier to type only the integer part.
[1] http://commons.wikimedia.org/w/index.php?title=Special:Upload&uselang=nl...
Vicenç Riullop
There would probably be a seperate form for each country - sometimes in different languages. For that specific country, the form can be as simple as last year, depending on the specific situation (the Germans seem to like more complicated forms ;-) )
In any case, yes - an identifier should be unique within a country - then we can always make it easily unique world wide too by combining it with the country code later. There is no need to do this in advance.
Best,
Lodewijk
2011/3/13 Vicenç Riullop vriullop@hotmail.com:
I think I have got it. The special upload form may not be as simple as last year [1] with a field for Rijksmonumentnummer. For Germany, i.e., the user should first select country, state, municipality, and then type the number, unless the identifier includes some codification. For Spain, if all monuments have the same prefix R-I-51, then this part is not relevant as an identifier and it will be easier to type only the integer part.
[1] http://commons.wikimedia.org/w/index.php?title=Special:Upload&uselang=nl...
Vicenç Riullop
Wiki Loves Monuments mailing list WikiLovesMonuments@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikilovesmonuments http://www.wikilovesmonuments.eu
wikilovesmonuments@lists.wikimedia.org