[Pkg-isocodes-devel] Handling ISO 639-3?

[Alastair McKinstry]

Christian Perrier wrote:
> Hello fellow co-maintainers,
>
>   
Sorry about the delay in replying; christmas holidays and recovering at
work.
> Today, I was looing at ISO 639-3 as of
> http://www.sil.org/iso639-3/codes.asp
>
> As far as I see, our list is correct and complete....however, the ISO
> 639-3 MA site uses constructions like "Moroccan Arabic" while we have
> "Arabic, Moroccan spoken" in our iso_639_3.xml file.
>
>
> There are numerous such examples, however some parts of the ISO-639-3
> MA show that "Arabic, Moroccan spoken" is another possible form. This is
> for instance the one used in Ethnologue....
>
> One is called the "print form", aka "Moroccan Arabic". Another is the
> "inverted form" which makes easier to sort language families, namely
> "Arabic, Moroccan spoken".
>
>
> I think we should indeed provide both forms in our XML file:
>
> 	<iso_639_3_entry
> 		id="ary"
> 		scope="I"
> 		type="L"
> 		name="Moroccan Arabic"
> 		inverted_name="Arabic, Moroccan Spoken" />
>
>   
Good idea; I will add this to a new branch, as it will have to go in
experimental, to stop the inevitable updates from blocking entry to etch.
> This could be done by combining both files:
>
> http://www.sil.org/iso639-3/iso-fdis-639-3_20061114.tab
> http://www.sil.org/iso639-3/iso-fdis-639-3_Name_Index_20061114.tab
>
> Do you guys think that this could be automated ?
>
> I think that Alastair got the file we have directly from the first of
> these two files.....
>
>   
yes :-) I should credit this in the notes.
> PS: if there is some script that is able to update our XML file from
> the TAB file available on the ISO-639-3 MA site, we should probably
> provide it along with our package.
>
>   
It was a once-off at the time; I agree that any future such scripts
should be
included in the repo and the tarball.
Regards
Alastair