Re: Type Coercion is confusing (ISSUE-34)

On 10/01/2011 05:58 PM, Ivan Herman wrote:
> ... and I understand all that compactness issue. But, remember Manu,
> I hit exactly the same issue when I began to read JSON-LD and the
> current way of spec-ing was very confusing to me. I wonder whether
> the compactness argument does not lead as in direction of a possible
> general user confusion and if we have to balance compactness vs.
> confusing spec, then, well... this reminds me of rdf/xml

Hmm... yes, good point. This is something that we've received a number 
of complaints about.

The data that we're missing to make a real educated decision about this is:

1) In the average JSON-LD document, how many coercion rules are
    specified, on average?
2) Are these rules typically provided via an external JSON-LD Context
    Document or in-line?

A wild guess at #1 says that there wouldn't be that many coercion rules 
in small systems, but large systems may have 20-50?

A wild guess at #2 says that most people are going to use an external 
JSON-LD Context... so, perhaps the "size" argument is negligible since a 
context will only be downloaded once and cached. Processing the document 
would add negligible time to the full document parse... in some cases, 
contexts may be cached by the application in a binary form to speed 
processing. Also, since the internal representation is going to be more 
like what Ivan and Markus wants, it would simplify the implementation.

So, perhaps the line of reasoning we made behind this particular 
mechanism is outdated. We did not have external @context documents at 
the time and therefore were more sensitive to the in-line context 
becoming too big.

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
Founder/CEO - Digital Bazaar, Inc.
blog: Standardizing Payment Links - Why Online Tipping has Failed
http://manu.sporny.org/2011/payment-links/

Received on Sunday, 2 October 2011 21:33:12 UTC