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: XML Blueberry

At Friday, 22 June 2001, John Cowan <jcowan@reutershealth.com> wrote:

>Those "legacy" systems contain a huge amount of well-maintained
>data, as someone else (Tim Bray?) pointed out.  Anyway, the
>only systems that are not "legacy" are the ones still being
>designed: Fred Brooks told us 25+ years ago that an implemented
>system is an obsolete system.

There's a difference between keeping a system up to date
and wilfully keeping a system using a technology which makes
it incompatible with the rest of the world. Our current world is
ASCII and Unicode. If we can make a small, transparent change
to accommodate those living on planets like EBCDIC, that is a
fine and generous gesture and we should simply do it. Otherwise 
it must surely up to them to change.

>> The time to speak up on this was four years ago.
>A fine attitude to bug-fixing, indeed.

If it's a bug.  The world is full of old systems which have now been
passed by by technology: I'm unconvinced we should be retrofitting all
new systems with bugfixes to the obsolete ones.

>The reason for introducing a new version of XML (or a new
>mark of some sort, anyhow) is to protect old parsers.  Allowing
>NEL and the Unicode 3.1 name characters changes the definition
>of what is a well-formed entity, thus going beyond what an
>erratum can fix.

As I've already said, it may have been a mistake to predicate XML
on the boundaries of another system which was known to be changing,
without making provision for changing with it. But that's just history.
still don't see any reason why the changes shouldn't be made so long
as it's clear *why* they are being made: I'm just opening the box to see
why it ticks.