Search Mailing List Archives
[p4-feedback] Protege, Content negotiation and XML DTDs (Very slow document opening)
tredmond at stanford.edu
Thu Mar 17 09:51:14 PDT 2011
I can give a bit more detail on what is going on. You will notice that
if you open the imported file directly it opens instantly. The OWL api
knows what to do.
It is Protege that is doing things wrong here. There is a class called
the MasterOntologyIDExtractor whose job is to find the name of an
ontology. The intent is that it will find the name of the ontology more
quickly than doing a full OWL parse. When you do an import, it is being
called as a last check to make sure that the imported ontology is really
there. If the ontology is not there a nice dialog can come up offering
I suspect that I can fix the content negotiation quickly problem because
I have seen the correct code in the OWL api and I think that I know how
to find it (it covers a whole slew of cases).
On 03/17/2011 08:27 AM, Marcus Cobden wrote:
> I'm working on an ontology which imports another which uses content
> negotiation to serve RDF from its URI.
> Specifically i am importing http://purl.org/net/provenance/ns.
> From what I can tell, protege isn't setting the request headers
> correctly, resulting in it being given the XHTML+RDFA version rather
> then the RDF version.
> It seems this causes protege's XML parser to go and download the XHTML
> These are very slow to retrieve (just over 30s per module) as the W3C
> is trying to encourage application caching .
> Eventually it will load correctly, but protege really should be
> caching things.
> So, in summary:
> * Protege should set Accept headers when importing ontologies.
> * Protege should be caching DTDs
>  http://www.w3.org/blog/systeam/2008/02/08/w3c_s_excessive_dtd_traffic
> p4-feedback mailing list
> p4-feedback at lists.stanford.edu
More information about the p4-feedback