Re: participating communities (was XML Blueberry)

[John Cowan <cowan@mercury.ccil.org>]

David Carlisle scripsit:

> But that seriously underestimates the cost. It may only take one
> Microsoft programmer an afternoon to update MSXML (and even if not, MS
> could afford the develpoment cost) but the real cost is the lost time
> and confusion caused when newly created conforming blueberry files get
> distributed to the millions of installed systems and die with fatal non
> well formed errors, without any helpful warning to the end user that an
> update to the system would fix the problem. 

How is this different from changing Microsoft Word format, with the result
that new Word documents can't be read by old versions of Word?  Back
in Word 2.0 days, Word 1.0 would try to accept Word 2.0 documents and
then crash because they made the wrong kind of sense.

At least there will be a positive indication, early on, that the
document is Blueberry.

> (I can't see any
> benefit at all in the NEL proposal, for IBM users or anyone else, the
> end result of making NEL be white space is that it would be mapped to
> #10, in which case that may as well just be in the ebcidic mapping table
> and not affect XML at all)

But the EBCDIC mapping table correctly maps CR to CR, LF to LF, and NEL
to NEL.  They have distinct functions on IBM systems: it's just that
NEL is the plain text line ending character.

To do what you suggest is to require there to be an XML-specific EBCDIC
mapping table.

-- 
John Cowan                                   cowan@ccil.org
One art/there is/no less/no more/All things/to do/with sparks/galore
	--Douglas Hofstadter