Xml Mailing List: Re: Re: comments on draft-frascone-aaa-xml-dictionary-00.txt

[David Frascone (dave]

Ok, let's clean this up a bit, and talk about the issues one at a time.  

1) p-bit in command.  I agree completely with this change, and have
incorporated it into the DTD, and will add it to the draft.

2) e-bit.  I think this is not necessary, since the e-bit is used to
signify an error, and could potentially occur in any command
code.  (Correct me if I'm wrong . . .drafts have been changing rapidly
lately).

3) Ordering of items in base + external entities.  I think forward references
are allowed in libxml2 (my code is working, and I use a dtd/dictionary with
them).  Given that, when one looks at a dictionary, I think the command 
codes are the most important part, and therefore should come first.  I 
ESPICALLY think that the external entities should come at the top.  Otherwise
files would be included anywhere and it could get really confusing.  So,
I don't think the reordering is necessary, or useful.

4) Required AVP syntax.  I like it the way it is, but have no strong objection
to the new format.  I think we need to get some rough concensus from other 
people/authors here.

5) Full ABNF syntax specification . . . This one is where I have the most
problems.  Mainly, because the tighter we tie down the ordering/etc, the 
more flexability we lose.  I know we had originally decided to just provide
rudamentary syntax checking.  But, if it is now possible to do full syntactic
checking (in other words, if the ABNF has become that well defined), perhaps
we should.  Again.  I need input from more people/authors on this one.


Please let me know if I have missed anything,


-Dave

-- 
David Frascone

  Honey, PLEASE don't pick up the PH$ [at] #*&$^(#@&$^%(*NO CARRIER