Re: the necessity of describing responses in-band

On 11 October 2015 at 12:16, Robert Sanderson <azaroth42@gmail.com> wrote:

>
> Frankly, I'm -1 on anything that *requires* named graphs to work
> properly.  There are too many implementations out there that would not be
> able to participate, and the added overhead is not insignificant.  One need
> only read the RDF WG archives in discussions for 1.1 to see the
> controversies here.
>
> Secondly, I find the argument that following HTTP and putting metadata
> about the response in the headers, at least disingenuous and at worst
> outright incorrect.  The headers are precisely where metadata about the
> body is intended to live.  This has made interfaces somewhat complex to do
> correctly since day 1 of the web, but that's what we have.  It's not that
> we're making exceptions for the data web, the exceptions are for the
> document web because browsers do not make response headers easily
> accessible for processing, nor process them natively per their semantics.
>

In general, agree with all.

Where to put meta data (headers vs document) is a tricky design decision.
It would seem to me that HTTP related meta data belongs in the header, and
document related meta data belongs in the document.  But there is a kind of
overlap between the two which is not always easy to establish.


>
> Rob
>
>
> On Thu, Oct 8, 2015 at 1:21 AM, elf Pavlik <perpetual-tripper@wwelves.org>
> wrote:
>
>> very relevant to building understanding among LDP NEXT CG & Hydra CG
>> participants, this article also presents opinions about some choices
>> made in current LDP specs!
>>
>>
>> -------- Forwarded Message --------
>> Subject: the necessity of describing responses in-band
>> Resent-Date: Tue, 06 Oct 2015 12:17:57 +0000
>> Resent-From: public-hydra@w3.org
>> Date: Tue, 6 Oct 2015 14:17:24 +0200
>> From: Ruben Verborgh <ruben.verborgh@ugent.be>
>> To: Hydra <public-hydra@w3.org>
>>
>> Dear all,
>>
>> I've written a blog post that describes the necessity
>> of describing responses in-band:
>>     http://ruben.verborgh.org/blog/2015/10/06/turtles-all-the-way-down/
>>
>> More than an argument for REST/hypermedia,
>> it's an explanation of _how_ we should realize that
>> with RDF-enabled representations.
>>
>> In this context, the Hydra Core Vocabulary is a major enabler,
>> because it lets us describe hypermedia controls in RDF.
>>
>> Best,
>>
>> Ruben
>>
>>
>>
>>
>>
>
>
> --
> Rob Sanderson
> Information Standards Advocate
> Digital Library Systems and Services
> Stanford, CA 94305
>