OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: participating communities (was XML Blueberry)

From: Tom Bradford [mailto:bradford@dbxmlgroup.com]

>>"Meanwhile, the existing XML code keeps working."

>What would prohibit it from continuing to work in either case?  Seems to
>me that a superset of allowable characters is backward compatible with
>the existing specification.  And besides, that's what version numbers
>are for.

What indeed?  Perhaps the cost of rebuilding the parsers because 
of a bit of paranoia about where they will be used, feature-envy, 
whatever.  It is like flood-mitigation:  many residents don't like 
to pay for the mistakes of real estate developers twenty years after 
they take the profits because the original land surveys lied about the 
history of the property and the likelihood of flooding.

>The last thing anyone should want to do is fragment XML, because it
>defeats the overwhelming benefit of XML, that being the ability to
>interchange data in a brain-dead easy fashion.  

But it has failed.  Thus Blueberry.

>The last thing I want to
>hear anyone say is that it's perfectly ok to have an incompatible custom
>implementation of XML.   I think we've seen enough history of splintered
>standards turning our lives into living hells.

You are living in Hell?  Do you have flooding there too?

The most I am daring is that this is a problem best solved 
in SGML where the solution is both legal and standard.   It seems that 
you are mistaking XML with a standard.  It isn't.  It is a product of 
the W3C Consortium that depends on a legitimate relationship with the
owner of the intellectual property, ISO, for it's provenance.  If, on the 
other hand, the W3C decides to change XML, it is their property, so theirs 
to do so and to pay for.  The question of cost is basic but undetermined. 
Cowan suggests it is not as high as other think.   If it is, there are 
alternatives.  Clark suggested a hack by relaxing WF name constraints.  A 
version number change will open up other opportunities (here come 

Do you think it possible that these may also be "splinters"?