Re: Question: localname partition and xsi:type.

Hi Taki,

Thank you very much. I failed to catch the paragraph. Now everything is clear for me.

The reason why we use xsi:type instead of substitution group is needs for ad-hoc extension. For version 1.0 schema we have plenty of type definitions (250 or so). We cannot predict which type to be extended in v1.1, and I don't think it's better to make everything to have abstract type.

Thanks for XML extensibility, without breaking schema validity :-)

Best Regards,

Yusuke


(2012-04-10 03:14), Takuki Kamiya wrote:
> Hi Yosuke,
>
> EXI depends on XML Schema 1.0 specification in expecting how
> attribute's target namespaces are assigned values. In your specific
> case, it ought to be in no namespace namespace (i.e. "").
>
> The eight bit zeros you are seeing most likely is part of the
> representation of localName "PoemType2". Please see the second
> paragraph in section 7.3.3 for how a localName that's already indexed
> gets encoded. It is marked with eight-bit zeros, which is consistent
> with what you saw in the stream.
>
> In general, substitution groups work better with EXI than what can
> achieve with xsi:type in terms of the compactness. This is something
> you might want to be aware when you design your schema.
>
> Regards,
>
> -taki
>
>
>
> -----Original Message-----
> From: Yusuke DOI [mailto:yusuke.doi@toshiba.co.jp]
> Sent: Monday, April 09, 2012 4:33 AM
> To: public-exi@w3.org
> Subject: Question: localname partition and xsi:type.
>
> Dear EXI gurus,
>
> I have two questions on spec.
>
> 1) attribute names and localname partition.
>
> Appendix D.3 tells
>
>> Additionally, when a schema is provided, the string table is also
>> pre-populated with the local-name of each attribute, element and type
>> explicitly declared in the schema, partitioned by namespace URI.
>
> In our schema, attributes are unqualified. Are they in '' partition? or it should be in the partition that the attribute defined in?
>
> 2) encoding of xsi:type
>
> On encoding of xsi:type, we're not able to find how to decode the following pattern of 8-bit zeros. Could you tell me what it is and where I can find it in the spec?
>
> 10000000 header
> 01110100 SE(poems) SE(poem) AT(xsi:type) NS(example.com/example) "PoemType2"
> 00000000<- what's this?
> 10 AT("at2str")
> 00000110 length=4
>
> 01100001 a
> 01100010 b
> 01100011 c
> 01100100 d
>
> (snip)
>
> I have confirmed it with both EXIficient and OpenEXI. However, I cannot find corresponding spec for the 8-bit zero.
>
> The XML and schema we used are attached.
>
> I appreciate your help.
>
> // Yusuke DOI<yusuke.doi@toshiba.co.jp>  Corporate R&D Center, TOSHIBA Corp.
>

Received on Tuesday, 10 April 2012 01:25:05 UTC