Re: Section 4: LDPR/non-LDPR formal definitions

On 3/25/13 4:24 PM, Richard Cyganiak wrote:
> On 25 Mar 2013, at 20:07, Erik Wilde <dret@berkeley.edu> wrote:
>>> I don't see how anybody wins in this scenario.
>> LDP wins, because now LDP clients don't have to (pretend to be) generic RDF clients (and the same is true for servers).
> Boy. This is like introducing a new media type for HTML when forms were invented, and justifying it by saying “now form-capable browsers don't have to pretend to be generic browsers”.
>
> A read-only LDP client *is the same* as a generic RDF client.
>
> A generic RDF server *is the same* as a non-writable LDP server.
>
> The HTTP GET side of LDP is exactly what is already deployed on hundreds of Linked Data sites around the web. All that LDP is doing is adding a POST/PUT/DELETE story to that.
>
> Breaking free backward compatibility by introducing a new media type, and breaking all existing RDF clients that could provide read-only access to LDP server for free, strikes me as an exceptionally poor design choice.

If it breaks then of course that's true. But does it? Will developers of 
RDF clients and servers really balk at a new media type that can be 
given identical treatment to what they already handle?

>
>> they speak LDP and nothing else.
> They also speak read-only linked data, as it is already deployed on hundreds of sites on the web.
>
>> they can make this explicit in HTTP conversations. how is that not a win for a protocol that explicitly intends to bridge RDF and non-RDF worlds?
> LDP doesn't intend to bridge RDF and non-RDF worlds. If it ends up doing so, that's a happy side effect.

"Happy side effect" is really what I believe is possible here.

> LDP intends to provide HTTP-based application integration based on read/write Linked Data.

Yes, but (as you know) the Linked Data tools space is still relatively 
small. It's the published Linked Data that's massive.

To conclude, it's a cost vs benefits thing where the goal here is 
removing a hurdle that could become yet another long term distraction. 
Many of us (you included) have been there and done that. My instincts 
are to encourage the aversion RDF based Linked Data odysseys at first 
whiff :-)

Kingsley
>
> Best,
> Richard
>
>
>
>> cheers,
>>
>> dret.
>
>
>


-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen