Re: WebID-ISSUE-61 (xsd): xsd datatypes [ontologies]

On 16 November 2011 20:00, Henry Story <henry.story@bblfish.net> wrote:
>
>
> On 16 Nov 2011, at 18:35, Melvin Carvalho wrote:
>> [snip]
>> -1 xsd:base64Binary -- henry is this your proposal?
>
> ( It's not so much my proposal, as something to explore, as people have brought it up in the past, and have again recently. We need to work out if it makes sense. So I am not for voting on these issues immediately, but for exploring the topic first. I agree with the position of the subversion people about not overdoing voting http://www.youtube.com/watch?v=-F-3E8pyjFo )
>
> Anyway, xsd:base64Binary and xsd:hexBinary map to the same value space, so if you are for xsd:hexBinary then you might as well be for xsd:base64Binary too. The same relation can have either as object, and if the tools support either they will support the other too.
>
>>
>> +1 xsd:hexBinary
>>
>> +1 xsd:string
>
> xsd:string does not bring anything new. One would then have to define the type of encoding of the number in the string and how each of these encodings map to the other. That is what xsd:int, xsd:integer, etc... do, and it is what xsd:base64Binary do too.
>
> So xsd:string would be like cert:hex without any of the flexibility.
>
>>
>> Basic reasoning, in general I prefer simplicity, I'd rather reuse if
>> it's practical, than create a new term.
>
> So do I, but there is no God to tell us what is simple. It is not something you can see without using the tools or looking in depth at the standards. You need to explore the space of possibilities before you can come to the conclusion that something is simple.
>
>>
>> Happy to go with whatever is decided, including keeping cert:hex.  But
>> if we are changing cert:int, it would be logical to package a change
>> of cert:hex at the same time.
>
> Well I would like some feedback of how different tools work with base64Binary and hexBinary. Specifically, does SPARQL ask work as well as I put forward. If people are willing to add some empirical evidence then we can get a better picture of where things stand.

Thanks Henry for clarifying.  Preference still stands, but happy to go
with cert:hex until there's some new info.

>
> Henry
>
>
>
>>
>>>
>>>
>>>
>>>
>>>
>>
>
> Social Web Architect
> http://bblfish.net/
>
>

Received on Wednesday, 16 November 2011 19:40:29 UTC