Re: Express Offers/Credentials as JSON-LD in HTML pages

On 1/15/15 8:05 PM, Manu Sporny wrote:
> The sooner we get the Credentials/OBI vocabulary integrated into
> schema.org the better.

Very interesting -- I went to schema.org and viewed their term at 
'Book' and there was a concise set of examples that show HTML, 
Microdata, RDFa, and JSON-LD side-by-side as tabs so we can compare 
the code easily.

http://www.schema.org/Book

This got me excited because I had several aha! moments about what the 
hell JSON-LD is and what it can become (in my world). :-)

But I'd like to check if these things that I believe about of JSON-LD 
are accurate.

I watched Manu's video this morning about JSON-LD, but that didn't 
help, so:

My questions: If I assume we're using 'schema' as @context, is it true 
that:

a) JSON-LD can be used as a markup for ANY web page, just in the same 
way as RDFa or Microdata. That a single web page describing a single 
book still has valid JSON-LD 'data', and the book's name, authors, 
[price, eventually], etc etc are all expressible though JSON-LD. ?

b) That this addition of JSON-LD to any given web page can be done 
INSTEAD of Microdata or RDFa -- that it will fulfill all their 
functions (and more) and so they aren't required; it isn't necessary 
to do both. ?

c) JSON-LD can be added IN A BLOCK, outside the existing HTML code 
(for example in the header), as opposed to Microdata and RDFa which 
(painfully) have to be added within the HTML code, line by line. This 
would be a major advantage in most situations (especially for 
retro-fitting existing web pages, but even for building new ones), 
over Microdata and RDFa.  ?

And finally one possible disadvantage:

d) JSON-LD's block of code must REPEAT the content of the web page in 
any data that it uses to define that content itself, whereas Microdata 
and RDFa are only written around the existing content and so there's 
no repetition. Thus in some cases, where the entire content of, say, a 
story or a review or a blog, is being specified by the JSON-LD (for 
some reason), then it must be doubled. ?

I apologize again if this seems partly off-topic of the credentials 
group, but in overview it does seem relevant to me: 12 years ago, 
after four years building it online, I took down a 100-page (partly 
experimental) web site that had a lot of content, and was being 
regularly accessed by web searchers, for two main reasons:

    1) Finding that content via web searches was chaotic and 
semi-random. (No widely-used, logical, and simple web semantic search 
system existed as a way for me to express who had created the content 
and what it was about).

   2) Paying for the content was difficult. (No universal web payments 
system that handled micropayments effectively existed).

It looks like #1 maybe now has been solved, and moreover the solution 
may be a key building block in solving #2.   :-)

Steven Rowat

Received on Friday, 16 January 2015 17:44:04 UTC