[tei-council] Fwd: bug report for Council, if you like

[Martin Holmes]

On 10-10-04 12:43 PM, Lou Burnard wrote:
> as noted before using data.numeric would be consistent with
> svg:points and also with what we currently allow for coordinate
> points on zone etc. so i say go for that.

Right. So the only remaining question is one Sebastian raised, and which 
I couldn't find the answer to in our previous correspondence: do we:

a) Require that the last component of a polygon be identical to the 
first, thus closing the polygon, and deem the attribute value invalid if 
this is not the case; or

b) Deem any attribute value whose final coordinate is not identical to 
its first to be closed anyway, thus assuming the existence of the final 
point where it is not explicit; or

c) Allow that some shapes may not be closed (in other words, they may be 
paths), and therefore assume that any shape not explicitly closed is open.

Thoughts? My instinct is to be permissive and choose c), but then we 
have the issue of possible conflict between the definition of a <zone> 
and the idea of a path. That's somewhat moot anyway, because the current 
definition of a <zone> is inconsistent with @points:

<zone> defines a rectangular area contained within a surface  element.

Cheers,
Martin

>
>
> from the pad
>
>
> On 4 Oct 2010, at 19:03, "Martin Holmes"<mholmes at uvic.ca>  wrote:
>
>> I think we need to decide whether the idea to use data.numeric is a
>> good one or not. Does anyone see any objections?
>>
>> Cheers, Martin
>>
>> On 10-10-04 10:34 AM, Lou Burnard wrote:
>>> thanks for clarifying the history martin... do you think you
>>> could update the ticket syd has just submitted with a proposed
>>> fix?
>>>
>>> from the pad
>>>
>>>
>>> On 4 Oct 2010, at 18:11, "Martin Holmes"<mholmes at uvic.ca>
>>> wrote:
>>>
>>>> I believe I've dredged up the history of this. On the TEI
>>>> Graphics SIG, in April, I posted this:
>>>>
>>>> [quote] HI all,
>>>>
>>>> Yesterday I worked out a simple regular expression to validate
>>>> the content of the proposed @points:
>>>>
>>>> [\d]+,[\d]+([\s]+[\d]+,[\d]+){2,}
>>>>
>>>> meaning:
>>>>
>>>> each point consists of two integers separated by a comma; a
>>>> points value consists of three or more of these points,
>>>> separated by a whitespace.
>>>>
>>>> However, going back to look at the specification for
>>>> @svg:points, which was our original model, I realize that it
>>>> allows negative values as well as positive, and floating-point
>>>> values as well as integers.
>>>>
>>>> Since @ulx and friends are defined as data.numeric, which also
>>>> allows negatives and floating-point numbers, it would make
>>>> sense to me that @points allows this too. This is the
>>>> definition of data.numeric:
>>>>
>>>> data.numeric = xsd:double | token { pattern =
>>>> "(\-?[\d]+/\-?[\d]+)" } | xsd:decimal
>>>>
>>>> <http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-data.numeric.html>
>>>>
>>>>
>>>>
In view of this, I think @points should be defined as three or more
>>>> instances of data.numeric, separated by whitespace. What do you
>>>> think? [/quote]
>>>>
>>>> So it seems that I'm the source of the erroneous regexp, but in
>>>> fact I wasn't proposing that it be adopted; I favoured using
>>>> three+ instances of data.numeric instead. Both Sebastian and
>>>> Conal agreed with this. However, on the SF feature request
>>>> discussion, this doesn't seem to have been considered
>>>> (presumably my fault -- I did include my faulty regexp in the
>>>> discussion, but neglected to bring over the resulting
>>>> discussion from the GRAPHICS SIG.
>>>>
>>>> Cheers, Martin
>>>>
>>>> Cheers, Martin
>>>>
>>>>
>>>>
>>>> On 10-10-03 10:46 PM, Laurent Romary wrote:
>>>>> Just pointing to it to confirm this would be the most
>>>>> sensible way to rpoceed. Laurent
>>>>>
>>>>> Le 3 oct. 10 à 20:14, Martin Holmes a écrit :
>>>>>
>>>>>>> what do you think about closing the path?
>>>>>>
>>>>>> I seem to recall a discussion about this before. Did we not
>>>>>> decide that if a path is not closed, it would be deemed
>>>>>> closed as if the first point were re-iterated as the last?
>>>>>
>>>>> Laurent Romary INRIA&    HUB-IDSL laurent.romary at inria.fr
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>> -- Martin Holmes University of Victoria Humanities Computing
>>>> and Media Centre (mholmes at uvic.ca)
>>>> _______________________________________________ tei-council
>>>> mailing list tei-council at lists.village.Virginia.EDU
>>>> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
>>
>> -- Martin Holmes University of Victoria Humanities Computing and
>> Media Centre (mholmes at uvic.ca)
>> _______________________________________________ tei-council mailing
>> list tei-council at lists.village.Virginia.EDU
>> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council

-- 
Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)