Rob,
> It is possible that a particular element is used only once per
document,
> but is used in every ODF document. If so it is important, since
everyone
> will need to understand it.
True: dc:date for instance...
> In other cases, a particular element will be
> repeated a thousand times when it is used, but is used only in rare
> documents. So there are two different kinds of "popularity" here.
Yes, probably some kind of statistical analysis is required
> Also, does anyone have a large collection of ODF documents that could
be
> tested in this way? I suppose, even if you had such documents
> internally,you could use the python script to do the study and just
upload
> the results. This is easier than uploading the documents themselves.
Mm, a few dozens perhaps, so not that big of a collection, although I
could collect more of them...
Technically, one could just crawl the Net using a search engine. Google
now shows 120.000 hits for filetype:odt, for example, but I'm not sure
how this works out legally (some websites don't even want you to link to
them...)
It should be OK I guess, but I'm not a lawyer...
But it would indeed be easier than uploading the documents: that way you
won't have to worry way less about uploading sensitive documents.
Best regards,
Bart
