Question about prefixes and N3

[Dan Armbrust]

Could someone explain this to me?

Given this code:

                Model model_ = ModelFactory.createDefaultModel();
                String uri = new URI("http://foo/contentA/version#";).toString();

                model_.setNsPrefix("a", uri);
                model_.setNsPrefix("l", 
com.hp.hpl.jena.vocabulary.RDFS.label.getURI());
                
                Resource source = model_.getResource(uri +
"951d9b63-bc07-3e58-bbb4-8dffd26e57f1");
                Resource target = model_.getResource(uri +
"a51d9b63-bc07-3e58-bbb4-8dffd26e57f1");
                
                Statement s = model_.createStatement(source,
com.hp.hpl.jena.vocabulary.RDFS.label, target);
                model_.add(s);
                
                RDFWriter rdfWriter = model_.getWriter("N3-TRIPLE");

                File file = new File("export.n3");
                OutputStream out = new FileOutputStream(file);
                rdfWriter.write(model_, out, null);
                out.close();

Jena outputs the following:

    @prefix a:       <http://foo/contentA/version#> .
    @prefix l:       <http://www.w3.org/2000/01/rdf-schema#label> .

    <http://foo/contentA/version#951d9b63-bc07-3e58-bbb4-8dffd26e57f1>
<http://www.w3.org/2000/01/rdf-schema#label>
a:a51d9b63-bc07-3e58-bbb4-8dffd26e57f1 .


You will notice that it uses the prefix for one of my values, but not
the other.  The only difference between these values is one of them
starts with a number, and the other doesn't.

Does a fragment really have to start with a letter?  Is this a bug?
Or a feature?

Also, you can see that I tried to make it write a prefix for the label
to shorten the output... but that didn't work either.  Is there a way
to do that?

Thanks,

Dan