[
Lists Home |
Date Index |
Thread Index
]
Interesting. I noted several issues that should be addressed, but I
don't think any of them are insurmountable.
1. The draft uses this BNF production:
uri = <any valid URI>
In practice, there are namespace names in the world that are not valid
URIs. For instance, they use non-ASCII characters or in other ways
violate the URI syntax. Can these be encoded into your scheme? Either
way this point should be spelled out and addressed.
2. related point: Namespaces 1.1 is defined in terms of IRIs, not URIs.
How is this handled?
3. Should this be any valid URI reference instead of any valid URI?
4. What if the namespace URI uses percent escapes? Do the percent signs
now need to be double escaped? (I expect they do, but this should be
made explicit.)
5. Editorial: the examples you give of these URNs all end with #. This
confused me at first because I thought it was part of the grammar.
However, I think it's just a function of the particular namespace URIs
you chose for examples. You should include some examples that don't end
with a # such as the XHTML namespace URI to make this more apparent.
--
Elliotte Rusty Harold elharo@metalab.unc.edu
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim
|
