RE: Objection to hexBinary and base64Binary

Michael --

Sorry for taking a few days to reply to this, I'm travelling right now.

I can't say that I'm really satisfied with the response.  The response
that Ashok outlined concerns me more than yours, since that response
essentially says that having two lexical spaces for the binary value
space is merely a prelude to a more complex scheme that would allow an
arbitrary number of lexical encodings for all datatypes.

I think the WG should re-examine its requirements document.  The
requirement for datatypes is to promote interchange of richly structured
XML documents.  I also believe that the requirements document states a
bias towards solutions that are simple and implementable (certainly I
pushed for this -- not going to check from dialup whether it made it
in).
Choosing a single lexical representation for each datatype advances both
of these requirements.  I cannot think of a requirement that is advanced
by allowing multiple lexical representations for a single datatype.

While it is true that supporting hex and base64 isn't technically that
difficult, my objections at this stage is a combination of saying it
sets a bad precedent and the idea is inelegant and inconsistent with the
handling of other datatypes.

I would once again re-iterate that if this is to be revisited for
Schemas 1.1 or 2.0, the conservative choice is to choose a single
lexical representation for binary and add additional representations
later only if experience proves them necessary.

I would suggest the single representation be base64, since that is
clearly more efficient.


-----Original Message-----
From: C. M. Sperberg-McQueen [mailto:cmsmcq@acm.org]
Sent: Friday, April 20, 2001 7:41 AM
To: Charles Frankston
Cc: W3C XML Schema Comments list
Subject: Re: Objection to hexBinary and base64Binary


At 2001-04-05 22:42, you wrote:

>There was a change made to the binary datatype back in January, between

>the Candidate Recommendation and the first Proposed Recommendation:

Charles -

in preparing for the director's decision on whether XML Schema
should go forward or not, it would be helpful to know whether
you are at all persuaded by the WG's response to your note on
the binary types.

Your argument, if I understood it right, was:

   1 this exposes the lexical form of the binary blob to
     downstream processors, which is bad
   2 this forces schema authors to choose a particular notation
     instead of just saying 'binary'
   3 this requires all processors to support both encodings

The WG position (you have already seen Ashok Malhotra's
description; here's another) is

   1 it would be bad, but since hexBinary and base64Binary
     both have the set of bit strings as their value space,
     and since downstream apps should normally have access
     to the value, not just the lexical form, it should be
     possible for downstream apps to ignore the lexical form
   2 schema authors can define a 'binary' type as a union
     of the hex and base64 types, so they can in fact
     just say 'binary' if they wish
   3 true, but neither encoding is really hard to support

Does this persuade you at all?  Let us know.  Thanks.

Michael