Solution proposal for versioning problem

[Jirka Kosek]

Norbert Bollow wrote:
> Jirka Kosek <jirka at kosek.cz> wrote:
> 
>> I think that any new features we are going to introduce in post
>> 29500:2008 will be designed to be backward compatible itself or relying
>> on MCEs so there will be no need to change namespace.
> 
> For those points where ISO/IEC 29500:2008 differs from ECMA-376 ed 1
> or from implementation decisions of existing software on points that
> ECMA-376 ed 1 leaves unspecified, what precisely is the target of this
> "backward" compatiblity?

I suppose ISO/IEC 29500:2008. But note then there are some changes
underway in AMD and COR which are trying to fix some unnecessary
incompatible changes made between ECMA-376 and ISO/IEC 29500:2008.

>> But what's the point of versioning information then?
> 
> See
> https://members.ecma-international.org/get.php?group=SC34-WG4&file=2010_sub_sc34-wg4-2010-0135.zip

Just quick comments:

ad "Use Case 1: Automated validation of documents against a schema"

Usually you are interested in knowing whether some document is valid
against particular version of schema which is supported by your toolchain.

ad "Use Case 2: Using different import filters for different OOXML versions
...One reasonable way to address this added complexity in programs which
are able to import OOXML is to use different import filters for
different versions of OOXML..."

If this will be necessary then we have definitively failed in standards
development and maintenance process. Any widespread markup language like
HTML or DocBook exists in several versions but user agents contain only
one "import filter" for reading such content.

I can imagine that there will be several import filters which will
compensate for bugs in respective OOXML implementations. But I don't
think that such thing should be supported by standard. It is simply
failure of implementations.

ad "Use Case 3: Humans seeking to determine the OOXML version"

If you want to be reasonably sure that document conform to particular
version of OOXML it will not be sufficient to read version information
from document itself. You have to additional validate content of
document, otherwise your software can be mistaken by supplying false
version identifier.

My feeling is still little bit split. I'm not convince that we strongly
need version information inside documents. But I'm not against it. But
if we introduce version info, we have to define processing expectations
for various combinations like -- application receives document which
conforms to older/newer/unknown/... version of OOXML. IMHO this is the
most hard and important issue. So discussing just versioning info
without knowing how such information should be processed doesn't make
big sense to me.

Have a nice day,

				Jirka

-- 
------------------------------------------------------------------
  Jirka Kosek      e-mail: jirka at kosek.cz      http://xmlguru.cz
------------------------------------------------------------------
       Professional XML consulting and training services
  DocBook customization, custom XSLT/XSL-FO document processing
------------------------------------------------------------------
 OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member
------------------------------------------------------------------

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 259 bytes
Desc: OpenPGP digital signature
URL: <http://mailman.vse.cz/pipermail/sc34wg4/attachments/20100712/6385ac61/attachment.pgp>