Re: ANN: Wikipedia as topic map

Hi Benjamin,

[...]
> I tried requesting the Leonard Cohen example below and found that the
> resulting file was transferred without Content-Encoding even though my
> client accepted it (i.e. had the Accept-Encoding header set). If it

That's true. I already thought about support for compression, but I
haven't implemented it yet.

> The file is also indented very cleanly which makes it easy to read for
> humans, but it also costs about 56163 bytes (uncompressed and

Also true. :) It was a debug setting which I've forgotten to remove.
Removed now.

> Third, I wondered why many subjectIdentifierRef elements are repeated,
> e.g. http://dbpedia.org/resource/Person occurs four times, artist
> occurs three times in the instanceOf list of Leonard Cohen. They get
> merged in the end, but why are there many of them in the first place?

They are, for some reason, duplicates in the RDF source I use. If the
service gets 30 times the predicate

   cohen rdf:type dbpedia:Person

It writes it 30 times. Since there is no Topic Maps engine involved
it's difficult to filter these duplicates. Anyway, a Topic Maps engine
would, as you said, merge the duplicates if it reads the topic map.

> Enough ideas for improvement for now. Thanks for this very good
> addition to the Topic Maps toolbox.

Thank you :)

Best regards,
Lars
-- 
Semagia
<http://www.semagia.com/>