Re: Visibility of the data (was Re: Formats and icing)

On 10/4/14 7:07 AM, Sarven Capadisli wrote:
> On 2014-10-02 00:48, Sarven Capadisli wrote:
>> On 2014-10-01 21:51, Kingsley Idehen wrote:
>>> On 10/1/14 2:42 PM, Sarven Capadisli wrote:
>>>> can't use them along with schema.org.
>>>>
>>>> I favour plain HTML+CSS+RDFa to get things going e.g.:
>>>>
>>>> https://github.com/csarven/linked-research
>>>
>>> What about:
>>>
>>> HTML+CSS+(RDFa | Microdata | JSON-LD | TURTLE) ?
>>>
>>> Basically, we have to get to:
>>>
>>> HTML+CSS+(Any RDF Notation) .
>>
>> Sure, why not!
>
> Actually, I'd like to make a brief comment on this. While I agree with 
> (and enjoy) your eloquent explanations on RDF, Languages, and 
> Notations, and that "any RDF Notation" is entirely reasonable (because 
> we can go from one to another at relative ease), we shouldn't overlook 
> one important dimension:
>
> *Visibility* of the data.
>
> Perhaps this is left better as a "best practice" than anything else, 
> but in my opinion:
>
> RDFa is ideal when dealing with HTML for research knowledge because if 
> applied correctly, it will declare all of the "visible" portions of 
> the research process and knowledge. 

The trouble with being notation specific is that it always inadvertently 
opens up a distracting war. It doesn't work, and will never work. You 
have to apply the wisdom of Solomon in the realm of RDF  --- something 
we (as a community) have failed to do, repeatedly, over the years.

RDF is a notation agnostic language, due to its abstract nature. We 
should really take more advantage this RDF virtue.

> It is to make the information available as first-class data as opposed 
> to metadata.

Metadata isn't the issue a hand here. Raw data is the issue, it 
shouldn't ever be confined to any kind of silo, in regards to the Web.

> It is less likely to be left behind or go stale because it is visible 
> to the human at all times.

RDF (various notations) based structured data islands in HTML all share 
this quality. It isn't unique to RDFa, at all. I simply see RDFa as more 
convenient in certain scenarios e.g., that you want to markup structured 
data inline (you allude to this usage scenario further down).

>
> This is in contrast to JSON-LD or Turtle where they will be treated as 
> "dark" metadata, or at least create duplicate information subject to 
> desynchronize. 

No, I think you aren't fully reperesenting the nature of HTML+JSON-LD 
and HTML+Turtle, in your comment above. You can effectively use them as 
raw data islands in HTML documents too, just as you can Microdata and RDFa.

If everyone chooses to use RDFa then fine, my point is that imposing 
(overtly or covertly) any RDF notation never works. In short, that's 
been RDF's problem since the days of RDF/XML.

If RDFa is the most productive notation, in a given scenario, its 
virtues will be obvious to those seeking to make their research data 
more accessible via HTML documents.

> While JSON-LD and Turtle have their strengths, they are unnecessary 
> when concerning the most relevant parts of the document which is 
> already visible, e.g., concepts, hypothesis, methodological steps, 
> variables, figures, tables, evaluation, conclusions.

I don't know how you are arriving at that conclusion when everything you 
mentioned above is an entity that's ultimately describable using RDF 
statements, in any notation.
>
> Again, this is not meant to force anyone to use a particular RDF 
> notation. 

It's better to provide guidelines for different approaches and then let 
folks choose what works for them. An alternative to that is a form of 
imposition, no matter how much its sugar-coated, unfortunately :)

> Getting HTML+CSS in the picture is a huge win itself as far as I'm 
> concerned :) 

That's clear, from your vantage point, but we have to think about 
everyone else, too.

> Then applying RDF notation is a nice reasonable step forward.
>
> * I am conveniently leaving out Microdata from this discussion because 
> I don't feel it is still relevant.

As I've already stated, we shouldn't really be talking about any 
specific RDF notation, when the goal is to set data free from these data 
silos using RDF.


>
> -Sarven
> http://csarven.ca/#i
>


-- 
Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: 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
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this