Re: [entity-resolution] Spec comments to process

[Normand_Montour@lgs.com]

Norm,
   about point no. 2 in the technical area (schema):  My thoughts would be
that
at least the DTD and the W3C Schema should be normative, but not exclusive
of any other one...
Regards
Normand




Norman Walsh <ndw@nwalsh.com> on 20/08/2003 08:52:42

To:    entity-resolution@lists.oasis-open.org
cc:
Subject:    [entity-resolution] Spec comments to process


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

As I went to produce the new version, I found that there were some
comments I had overlooked at our last call. Sorry. We should probably
try to address these. Here's a quick summary, ignoring editorial nits:

Editorial:

- - 4.1. Make schemalocation clearer
- - 4.3. Does the resolver ever see fragment identifiers? It should not.
- - Move answer to "does the resolver resolve itself?" higher up in the
spec.
- - Middle of bulleted list item 1 in 5.1, "ignoring any [SUCH] processing
inst."
- - Two paras before section 6: paragraph is ambiguous. makes it sound like
  there's some subtle requirement to detect all possible circularities.
- - Last paragraph of 6.1, last attribute is unqualified but you can't
  tell if it's local or global.

I'm happy to make these changes since I don't believe they change the spec.

Technical:

- - Public identifier for catalog is not normative
- - Schema should be normative? At least one of them should be normative?
- - Section 6.3, forcing URIs to be encoded removes the benefits of utf-8.

 1. they must be escaped completely to the appropriate ascii form in the
    doc
    either you do the comparison using whatever form is given and you'd
    expect
    the caller to do the escaping
 2. there is no normative required escaping, make sure your catalog
    documents
    are using URIs escaped in the same way as the caller
 3. use the full breadth of utf-8 and escape only the URI reserved symbols.
    do the same thing from the caller.

  note: we don't say anything about the application calling the resolver.
  either the resolver normalizes all of them, or both the catalog editor
  and
  the app layer must do normalization as specified in 6.3.

Thoughts on these?

                                        Be seeing you,
                                          norm

- --
Norman.Walsh@Sun.COM    | A new scientific truth does not triumph by
XML Standards Architect | convincing its opponents and making them see
Web Tech. and Standards | the light, but rather because its opponents
Sun Microsystems, Inc.  | eventually die, and a new generation grows up
                        | that is familiar with it (Planck 1949)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.7 <http://mailcrypt.sourceforge.net/>

iD8DBQE/Q28aOyltUcwYWjsRAgYQAJ0alAbXEbdQX9hhID00gKK9lb51KwCgg/q8
AOvSRDyxDrezw5DWz2F/x8Y==VrXv
-----END PGP SIGNATURE-----

You may leave a Technical Committee at any time by visiting
http://www.oasis-open.org/apps/org/workgroup/entity-resolution/members/leave_workgroup.php