Subject:Schema Validation Author:(Deleted User) Date:23 Feb 2004 03:13 PM
I have a schema that uses xsd restriction mechanism and I get such different result from the various validators.
Your default one is absolutely ruthless. Can you please tell me if you are confident with the behaviour of your tool over say the .NET and XSV versions.
Subject:Re: Schema Validation Author:(Deleted User) Date:23 Feb 2004 05:06 PM
At 15.39 23/02/2004 -0500, stylus-studio-tech Listmanager wrote:
>From: "Martin Roberts"
>
>I have a schema that uses xsd restriction mechanism and I get such
>different result from the various validators.
>
>Your default one is absolutely ruthless. Can you please tell me if you
>are confident with the behaviour of your tool over say the .NET and XSV
>versions.
The default schema validator is Xerces-C++; as for saying what is the best
one, I know no official benchmark that can give you an absolute answer. Any
validator has bugs/incomplete implementation, especially given the
complexity of the specs.
However, the feedback I had from people working with schema validation for
a living is that Xerces-C++ is better than the other two; while Xerces-J
and IBM Schema Quality Checker seem to be better than Xerces-C++.
In any case you should test what is the validator that works better with
the schemas you are using and use that one.
Subject:Re: Schema Validation Author:(Deleted User) Date:25 Feb 2004 11:05 AM
At 10.27 24/02/2004 -0500, stylus-studio-tech Listmanager wrote:
>From: "Martin Roberts"
>
>Alberto,
> The problem is 'What is the best validator for me' one that passes
> things I want to do or one that implements the spec best.
>
> As it is always the restriction that gives me problems is there any way
> of knowing if the xercesC++ implements the spec accurately?
The only bug in the 'restriction' area that I am aware of is filed as
bug#14000 (see http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14000 )
and refers to Xerces triggering as invalid a complex type that has a simple
content obtained by restricting a simple type. In a sense, Xerces it's too
restrictive...
There are also a few bugs in the detection of duplicates (i.e. in some case
they are not detected); and others involving the validation of datatypes
like dates and base64.
Subject:Re: Schema Validation Author:(Deleted User) Date:01 Mar 2004 08:06 AM
Alberto,
In your opinion if the schema validates in Xerces J that is supplied with your product as an extension validator, would you be happy?
I am having other problems where the inbuilt validator seems to be having problems with validating contents that have been declared as any##other and skip. Can you help?
Subject:Re: Schema Validation Author:(Deleted User) Date:01 Mar 2004 08:22 AM
At 08.33 01/03/2004 -0500, stylus-studio-tech Listmanager wrote:
>From: "Martin Roberts"
>
>Alberto,
> In your opinion if the schema validates in Xerces J that is supplied
> with your product as an extension validator, would you be happy?
If I were a schema writer and my schema was validated by Xerces-J, I would
be happy.
> I am having other problems where the inbuilt validator seems to be
> having problems with validating contents that have been declared as
> any##other and skip. Can you help?
Subject:Re: Schema Validation Author:(Deleted User) Date:01 Mar 2004 10:18 AM
Given the Unique Particle rule they refer to could we have an option to put the -f flag on to disable this check. This might over come my other problem.