10427
Reporter: mdoering
Assignee: bko
Type: Task
Summary: Fix Spanish checklists
Priority: Major
Resolution: Fixed
Status: Closed
Created: 2011-11-17 10:00:50.217
Updated: 2013-12-06 17:55:13.701
Resolved: 2012-10-01 14:16:26.633
Description: The registered checklists of GBIF Spain need to be fixed before they can be indexed:
http://gbrds.gbif.org/browse/agent?uuid=6c4a0bb0-2a4d-11d8-aa2d-b8a03c50a862
for example:
http://www.gbif.es/ficheros/GNA/Inventario-DwCArchive_Pli.zip
]]>
Author: mdoering@gbif.org
Created: 2011-11-17 10:01:32.809
Updated: 2011-11-17 10:01:32.809
15.Nov:
In Vernacular.txt
no. 2466 has language code spaM.
And I am not sure names like "N·car", or "Discoglosso-ibÈrico" (line 2411), which seem to have broken accent marks, need corrections. Please decide if it needs fix.
In SimpleImages.txt
title1 has 2 values: DitribuciÛn and DistribuciÛn. Should be typos and I think the Ü should be ó, please confirm.
Author: mdoering@gbif.org
Created: 2011-11-17 10:02:44.028
Updated: 2011-11-17 10:02:44.028
16.Nov:
We spot some other similar concatenations in reference.txt before we proceed to re-indexing. Sorry if I didn't remind perhaps because of focusing on others.
I.e.:
taxonID 1000 actually contains 5 references which should become 5 separate records:
DÍAZ, M., B. ASENSIO & J.L. TELLERÍA (1996) Aves ibéricas I. No paseriformes. Edit. Reyero. Madrid...
HERRERA, C.M. & J.L. ALCAIDE (1977) Posible nidificación de Jynx torquilla en Sierra Morena occidental. Ardeola, 22: 148-149...
MUNTANER, J., X. FERRER & A. MARTÍNEZ-VILLALTA (eds.) (1983) Atlas dels ocells nidificants de Catalunya i Andorra. Ketres ed. Barcelona...
PARDO, R. (1984) Las aves del naranjal de la provincia de Castellón. Caja de A. y Monte de Piedad de Castellón de la Plana...
TOMIALOJC L. (1994) Jynx torquilla. En: Tucker, G.M. & M.F. Heath (eds.) (1994). Birds in Europe: their conservation status, pp.: 342-343. BirdLife Intern. Cambridge.
Also, in distribution.txt, although it seems reasonable to map to distribution extension, in locality usually a name of a place is expected. So for the long descriptive text you have, we would suggest you to map them to description extension as separate records, and remember to give a proper type in
http://rs.gbif.org/vocabulary/gbif/description_type.xml
Third, the Type.txt only have institutionID mapped, which seems not helpful using this extension. If there is no type specimen to refer to and to associate with a name, perhaps you don't need to use it.
Now the core DarwinCoreA_Pli.txt.
You mapped an acceptednameusage column in which a lot of values look like free text and some are referring to more than one name. To correct this, we suggest:
1. For synonyms, each synonym should have their own records in this file, and each of the record refer to its accepted name(one name). Would be better if you use acceptednameusageID and parentnameusageID to just refer to the ID, which provide better accuracy of reference.
2. For descriptive text you can use taxonRemarks to map it.
Author: bko@gbif.org
Created: 2012-05-02 09:11:35.821
Updated: 2012-05-02 09:11:35.821
In their recent revision, they use vernacular column in the core, without providing language. Since all vernaculars are already mapped in the vernacular extension, I suggest they can omit the vernacular in the core.
The description in the bibliography is more relevant to description extension and in fact containing several types. Also suggested for revision.
Author: mdoering@gbif.org
Comment: Providing a language for vernacular names is important - even if its a static mapping with always the same value, e.g. "es"
Created: 2012-05-02 09:51:20.637
Updated: 2012-05-02 09:51:20.637
Author: bko@gbif.org
Created: 2012-05-02 10:46:51.195
Updated: 2012-05-02 10:46:51.195
I can see they have mapped that in the vernacular extension.
A question is, if the language is provided in the core mapping, does it by default referring to the vernacular attribute? Or could it be considered as the language of the whole record?
Author: mdoering@gbif.org
Created: 2012-05-02 11:15:02.541
Updated: 2012-05-02 11:15:02.541
This is a good question - I don't think we can safely link a core language column to a core vernacular name column.
For a "complex" vernacular name class the extension has to be used.
Author: bko@gbif.org
Created: 2012-05-02 11:23:38.777
Updated: 2012-05-02 11:23:38.777
I think perhaps we should always suggest people use vernacular extension for vernacular names, and the meaning of language in the core mapping should be defined better.
Should the discussion of definition happen here? Or should we put it somewhere else?
Author: bko@gbif.org
Comment: The checklist resource itself had been updated since then. So I think this issue can be closed. The issue of language attribute of a core probably should happen somewhere else.
Created: 2012-10-01 14:15:28.029
Updated: 2012-10-01 14:15:28.029