[
Lists Home |
Date Index |
Thread Index
]
At 01:12 PM 5/7/2002 -0700, Dare Obasanjo wrote:
>I'll let you in one an Open Secret, Jonathan; XQuery
>is not going to be interoperable. The XQuery WG set
>out to create the SQL of the XML world and they have
>succeeded. Interoperability between RDBMSs breaks down
>beyond the simple cases with proprietary extensions
>holding sway while lip service is paid to standards.
I agree that this is the case for SQL, and this is a real problem for
users. If vendors care about users, they should not want this to happen
with XQuery.
>No one fully implements SQL92, SQL99 may as well be a
>myth for all RDBMS vendors care about.
Well, our spec tries hard to be unambiguous. There are many things in the
realm of error handling that need to be tightened up, but I believe they
will be. Anybody who does not conform to the spec does not conform to the
spec. There will be areas of ambiguity that need to be tightened up.
> From conversations with XQuery implementers both
>within and outside MS, I am confident that this
>situation is about to be replicated with XQuery
If Microsoft or other vendors do not intend to conform to the spec, that is
bad news indeed. Are you speaking on behalf of Microsoft when you say that?
Perhaps other participants on XML-Dev can say whether they care about
interoperability or not.
My impression is that a lot of people really feel the need to have
implementations out there as soon as possible, and the fact that XQuery is
not finished means that we will see initial pre-releases that do not
conform to the final spec. This is a matter of timing. When we have a REC,
I do expect that implementations will support it.
>especially since a number of its requirements simply
>cannot be efficiently implemented over certain data
>repositories.
Do you have some specific requirements in mind that we can discuss
concretely? I encourage you to raise them on the official comments list so
that they are addressed:
public-qt-comments@w3.org
If there are real problems that keep you from conforming to our spec, I
certainly hope you will bring them up as soon as you can, because we do
want our specs to be unambiguous and implementable.
Jonathan
|
