Re: Examples in the LDP primer

On 12/18/13 5:11 AM, Roger Menday wrote:
> Well, I think there isn't a clear cut answer on this one. I appreciate the answers from Kingsley, but, when appealing LDP to a broader audience, a lot of people will probably appreciate the simplicity that comes from allowing a URI entity to also be directly de-referencable ??
>
> My opinion is that we are 'good' in the Spec, use both ways in the Primer, and (as Steve suggested) outline the pros-and-cons of both in the BP&G.
>
> Roger

Roger,

If you use relative HTTP URIs in the examples, all the glory of Linked 
Data manifests as part of the reading and learning experience. Simple 
rule, stay away from prefixes, use relative URIs, leaving use of 
absolute URIs for properties and classes defined by shared vocabularies.

Users should be able (circa. 2013) to copy and paste examples from specs 
to documents stored in their own personal data spaces e.g., those 
provided by the likes of Dropbox, Google Drive, SkyDrive etc..

I know of  nothing simpler, in the context of RDF based Linked Data 
demonstration and tutorial material, than the pattern I am describing 
here. It just works :-)


Kingsley

>
>
> On 16 Dec 2013, at 20:37, Kingsley Idehen wrote:
>
>> On 12/16/13 1:20 PM, Wilde, Erik wrote:
>>> On 2013-12-16, 10:11 , "Kingsley Idehen" <kidehen@openlinksw.com> wrote:
>>>>     On 12/16/13 12:31 PM, Roger Menday
>>>>       wrote:
>>>>       Maybe we should be good in the Spec, and be naughty in the
>>>>         Primer ... (?)
>>>>     No, once its naughty it just gets naughtier and harder to rectify.
>>>>     History (e.g., RDF and Web) shows, it can even take 13 or so years
>>>>     to fix the ensuing misconceptions.
>>> agreed that it's different for linked data, but on the web, having
>>> identifiers that do not resolve is perfectly acceptable. that's why
>>> hypermedia links are typed: you follow the ones where the type implies
>>> they're dereferencable, and you only use them as identifiers where the
>>> type implies they are identifiers only.
>>>
>>> i am not sure if what roger suggests is to point out that this is what's
>>> natural for the larger web and REST in general. i agree that we should be
>>> careful to promote/show patterns that are not exactly the way how things
>>> should be done in a certain context, but then again, if that demonstrates
>>> how things are done in practice (even though it may not be the ideal way
>>> of doing them), then there might be value in describing those examples as
>>> well.
>>>
>>> cheers,
>>>
>>> dret.
>>>
>>>
>>>
>> As far as I can understand, Roger is seeking clarification about entity (thing) denotation using HTTP URIs. Right now, we have examples in the LDP specs that inaccurately denote entities (that aren't Web accessible Resources) using HTTP URLs.
>>
>> As I said, an HTTP URL is an HTTP URI that denotes a Web Resource.
>> A WebID is an HTTP URI that denotes an Agent.
>> HTTP URIs can be used to denote anything i.e., any kind of entity.
>>
>> Another way to look at this is through the difference between "words" and "terms" .
>> A "term" is a specialization of a "word" i.e., a "term" is a "word" that embodies denotation (naming) and reference such that a name resolves to the description of its referent. [1]
>>
>> The World Wide Web is comprised of webby words i.e., words link to documents (using controls like <a/> in HTML , no referent description is implied this particular association.
>>
>> The RDF based Linked Open Data Web is comprised of webby terms i.e., you have denotation and reference at the very core.
>>
>> [1] http://www.wikihow.com/Differentiate-Between-a-Term-and-a-Word
>>
>> -- 
>>
>> Regards,
>>
>> Kingsley Idehen	
>> Founder & CEO
>> OpenLink Software
>> Company Web: http://www.openlinksw.com
>> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
>> Twitter Profile: https://twitter.com/kidehen
>> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
>> LinkedIn Profile: http://www.linkedin.com/in/kidehen
>>
>>
>>
>>
>>


-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen