Lists Home |
Date Index |
- To: firstname.lastname@example.org
- Subject: Re: [xml-dev] Is the XML Schema Test Collection still alive
- From: Michael Champion <email@example.com>
- Date: Fri, 15 Apr 2005 09:12:37 -0700
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=E+RFVIK1ypgVYBmftKmCTeRKJ7GTkm1/l8oj8HcePMJJATNH0aaKB/geGVg3hCtx9k4tj9tdaFbSOTxim/JyKlBiyOuLvjDjPw+su07flm4iZsgY691id65dC8IRSKTiveUi811tEVmodWlKBpnj5zCzKFfBe6ux42wHbRRBrUQ=
- In-reply-to: <425e9db0.65761d1d.06e6.420eSMTPIN_ADDED@mx.gmail.com>
- References: <firstname.lastname@example.org> <425e9db0.65761d1d.06e6.420eSMTPIN_ADDED@mx.gmail.com>
- Reply-to: Michael Champion <email@example.com>
On 4/14/05, Michael Kay <firstname.lastname@example.org> wrote:
> > I generally agree, except that more and more tools are forced to add
> > "buggy tool compatibility modes" to make this happen with minimal
> > customer pain.
> That's a road to ruin. It only prolongs the pain.
Yes, that's a good point.
> Conforming to the schema spec may be hard, but conforming to another
> implementor's undocumented and changing bugs is completely impossible.
I do think a good case can be made for putting in flag to cover a
well known class of conformance issues, such as those related to the
UPA constraint, especially since a widely used tool apparently
deliberately chose not to follow the spec in this area. I have heard
about a fair number of our users who don't have the ability to fix
such schemas, but don't want the validator to choke on them either.
Gresham's Law being what it is, such users will simply use validators
that don't worry about (or give you a switch to ignore) that class of
conformance issues ... and then the "bad validators" will tend to
drive out the "good validators."
It's a dilemma for sure. All I can think of suggesting, wearing the
Day Job hat, is to provide this particular flag but strongly warn
against its use except as a last resort.