[xsl] Re: XSLT 3.0: Question about the meaning of <xsl:modetyped="unspecified"/>

["Dimitre Novatchev dnovatchev@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

> However, I agree that it would be useful to say so explicitly.

Yes, it would be good to have the shape of the decision tree shown
explicitly -- otherwise the 5-valued logic is  rather challenging  :)

Cheers,
Dimitre Novatchev

On Fri, Jan 2, 2015 at 2:20 PM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
> The values yes | no | typed | untyped all impose contraints; they cause errors to be reported if certain conditions apply.
>
> Omitting the attribute, or specifying "unspecified", imposes no contraints. So in a sense, this value has no effect; there are no semantics to be explained. However, I agree that it would be useful to say so explicitly.
>
> Michael Kay
> Saxonica
> mike@xxxxxxxxxxxx
> +44 (0) 118 946 5893
>
>
>
>
> On 30 Dec 2014, at 02:48, Dimitre Novatchev <dnovatchev@xxxxxxxxx> wrote:
>
>> The 2nd Last Call of the W3C XSLT 3.0 specification lists the
>> attributes of the <xsl:mode> declaration and their possible values.
>>
>> In Section "6.6.1 Declaring Modes"
>> (http://www.w3.org/TR/2014/WD-xslt-30-20141002/#declaring-modes), the
>> possible values of the "typed" attribute are enumerated as:
>>
>>   typed? = boolean | "strict" | "lax" | "unspecified"
>>
>> Here it is also said that the default value for this attribute is "unspecified".
>>
>> These values are discussed in section "6.6.3 Declaring the Type of
>> Nodes Processed by a Mode"
>> (http://www.w3.org/TR/2014/WD-xslt-30-20141002/#xsl-mode-typed).
>>
>> It is explained what the two possible Boolean values mean, and that
>> the values  "strict" and "lax" imply the Boolean true() and add some
>> nuances to it.
>>
>> However, there is no mention of the default value "unspecified" and
>> thus it is not known what this value exactly means.
>>
>> My question is:
>>
>>  What is the meaning of the value "unspecified" of the "typed"
>> attribute of <xsl:mode> and how it is different from specifying the
>> Boolean value false() ("no", 0 or "false").
>>
>> Have I missed an existing description of this, or does such
>> description need to be added?
>>
>>
>> --
>> Cheers,
>> Dimitre Novatchev