[Standards] Namespace versioning

Mark Rejhon markybox at gmail.com
Wed May 25 16:00:57 UTC 2011

To date, I had the opportunity to implement my prototype standard (upcoming
resubmission) in two different programming languages -- C# and Java.   I
learned that different XML libraries handles namespace versioning very
differently.  With one XML library, I was able to monitor all namespaces
that matched a specific wildcard, which is useful for translating that to
the end user warning that a software upgrade is needed.    But for a
different XML library, my software was 100% blissfully ignorant of all
namespaces that did not have an exact match.  So it was not possible (or
easy) to translate that to a warning to the user to upgrade the software.

This is more especially important during the "beta" stage (i.e. year 2011,
or part of 2012), while the specification is during a draft stage.
 Nontheless, if necessary I can still use other techniques of software
version negotiation, if necessary (there's already XMPP specification for
that, from what I recall).

(I am also in collaboration with a emergency 911 center that is considering
implementing my spec, so I am under some pressure, too.)

On Thu, May 19, 2011 at 7:43 AM, Dave Cridland <dave at cridland.net> wrote:

> On Thu May 19 08:24:43 2011, Jacek Konieczny wrote:
>> Yes. It is wrong before <dialback xmlns='urn:xmpp:features:dialback:1'/>
>> is a totally different element than <dialback
>> xmlns='urn:xmpp:features:dialback' />.
>> It is not a new version, it is a new protocol. And a new namespace. Why
>> would we want that?
>> That is the always returning problem with XMPP protocols – treating the
>> namespace like an attribute. It is not an attribute, it is a part of
>> element name.  '<a xmlns="a"/>' is different from '<a xmlns="b"/>' the
>> same way '<a/>' is different from '<b/>'. Namespaces give us
>> availability to use e.g. '<error/>' element in different protocols in a
>> different way, for a different purpose. They should not be used to give
>> extra meaning to an existing element.
> Since I think this'd be the third time this month I'd try to explain
> namespace versioning, please just read the attached. FWIW, the vast majority
> of people working on XMPP (including I might add everyone contributing to
> this thread, as far as I know) do understand XML namespaces. The way we
> allocate namespaces is, however, fairly confusing when you see it and I
> don't blame you for (apparently) misunderstanding it.
> Also, Peter, Oh Mighty XEP Editor, I apologise for its rough form, but
> hereby submit it as an Informational XEP to provide background reading for,
> and expansion on, XEP-0053§4. I beseech thee, as well as those nice chaps on
> the Council, to help me clear it up a bit, too.
> Dave.
> --
> Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net
>  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
>  - http://dave.cridland.net/
> Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/standards/attachments/20110525/065f34d2/attachment.html>

More information about the Standards mailing list