This is the archive of the goodrelations dicussion list

Hi Vasiliy,
Apologies for the late reply. See inline comments.

On 17.06.2010, at 17:39, Vasiliy Faronov wrote:

> Martin Hepp (UniBW) wrote:
>> You are right, the use of tel as a plain datatype property is in  
>> partial
>> violation of the latest 2010 vcard ontology spec. However, we  
>> currently
>> recommend to stick with the Yahoo vcard representation pattern, since
>> that is necessary for augmented rendering of phone numbers in Yahoo.
>
> Thanks Martin, that explains it.
>
> But the previous version[1] of the spec, 9 years old, has the same  
> usage
> of vcard:tel as the current one. So basically Yahoo is at fault for
> incorrectly interpreting the spec?
>
> Anyway, this discrepancy should at least be mentioned somewhere in the
> Quickstart guide, as a footnote perhaps.
Added.
See
http://www.ebusiness-unibw.org/wiki/GoodRelationsQuickstart#Remarks

>
>
>> PS: IMO, SPARQL queries must tolerate a bit of deviation from  
>> standards.
>
> When a deviant pattern is widespread, yes, clients practically have to
> do some workarounds. Similarly, when it's necessary to cope with  
> widely
> used but misbehaving clients (as is apparently the case here with
> vcard:tel), workarounds on the server side are of course acceptable,
> but, if at all possible, they should be used *together with* correct
> markup, not in its stead.
>
> For example, if one were to do it like this:
>
> 	[] vcard:tel
> 		"+1 234 56 78" ,
> 		[ rdf:value "+1 234 56 78" ] .
>
> would both Yahoo and spec-compliant consumers be happy?
Yes, but there are two arguments that stop me from recommending that  
at this point:

1. It blows up the size of markup. Implementing this in RDFa adds a  
whole lot of complexity for the casual coder. Keeping markup to the  
minimum is really important for widespread adoption. In fact, the  
impact of RDFa on page loading times has been a frequent concern by  
SEO experts. Fortunately, we were able to show that our trimmed  
recipes cause less than 1 % increase in typical page loading times.

2. The markup would be an obvious contradiction, since the vcard:tel  
property would be used as an object property and a datatype property  
in parallel.

>
>
> [1] http://www.w3.org/TR/2001/NOTE-vcard-rdf-20010222/#5
>
> -- 
> Vasiliy Faronov
>
>