At 17:24 14/12/2004 +0000, Ian Turton wrote:
> Behalf Of David Zwiers
>
> Made a Jira for this, not sure it we actually should make the
> parser case insensitive ... but started some discussions ...
My view is that it is always best to be forgiving in reading and correct
in writing.
Ian
Though it should be remembered that Internet Explorer and Netscape were very forgiving in reading and as a result the HTML specs, to a large extent, fell apart.
It can be very confusing for a user to have a query that works against one implementation and not against another. I can well imagine that parsing a filter that should, according to the spec, fail would not pass the conformance tests.
As annoying as it is for the user I feel that, in the interests of true interoprability we should not actually be forgiving in parsing. However, what we SHOULD do is fail with an exception that provides the user with enough detail to fix their broken expression.
James
At 17:24 14/12/2004 +0000, Ian Turton wrote:
> > Behalf Of David Zwiers
> >
> > Made a Jira for this, not sure it we actually should make the
> > parser case insensitive ... but started some discussions ...
>
>My view is that it is always best to be forgiving in reading and correct
>in writing.
>
>Ian
Though it should be remembered that Internet Explorer and Netscape were
very forgiving in reading and as a result the HTML specs, to a large
extent, fell apart.
It can be very confusing for a user to have a query that works against one
implementation and not against another. I can well imagine that parsing a
filter that should, according to the spec, fail would not pass the
conformance tests.
As annoying as it is for the user I feel that, in the interests of true
interoprability we should not actually be forgiving in parsing. However,
what we SHOULD do is fail with an exception that provides the user with
enough detail to fix their broken expression.
James
I was hoping this conversation would take place in jira 
I essentially said the same thing - we should fail, and we should give
better error messages:
http://jira.codehaus.org/browse/GEOS-260
Chris
--