AW: Canonical EXI section 4.2.2

Hi Taki,

I agree with you that "MUST NOT" should be lower cased in a note.

Having said that, I am not sure whether Canonical EXI processors should be allowed to implement a feature that accounts for XML schema validity.

Let me give you one example with the following XSD snippet

    <xs:simpleType name="myRangedUnsignedInt">
        <xs:restriction base="xs:unsignedInt">
            <xs:minInclusive value="500"/>
        </xs:restriction>
    </xs:simpleType>

The type "myRangedUnsignedInt" maps to UnsignedInteger [1]. According to my understanding every Canonical EXI processor should also use the UnsignedInteger representation even if the integer value is below 500.

IF we allow for schema validity in such a case we end up with canonical EXI processors behaving differently.

Do you see my concern?

Thanks,

-- Daniel

[1] https://www.w3.org/TR/exi/#encodingUnsignedInteger





________________________________
Von: Takuki Kamiya [tkamiya@us.fujitsu.com]
Gesendet: Dienstag, 28. Februar 2017 09:26
An: public-exi-comments@w3.org
Betreff: Canonical EXI section 4.2.2

Hi,

Excerpted below is a note in section 4.2.2 [1].

"A Canonical EXI processor MUST NOT account for XML schema validity
 (similar to an EXI processor) in order to maintain high-performance efficiency."

The statement uses "MUST NOT", therefore it should not be placed in a note.

Furthermore, I think canonical EXI processor can optionally choose to implement
a feature to account for XML schema validity. In that case, "MUST NOT" may be
a bit too strong an expression.

[1] http://www.w3.org/TR/2016/CR-exi-c14n-20161103/#exiEventMatching

Thank you,

Takuki Kamiya
Fujitsu Laboratories of America