Re: [httpapi] linkset draft ready for next steps

[Phil Archer <[e-mail]>]

Can I clarify your comment please Julian.

Are you suggesting that Linkset could drop support for multilingual titles on a given link? And does that apply to both the native HTTP Link header format and the JSON format, or just one? Such a move would remove complexity that may be little used. For better or worse, it seems that in the real world, a vanishingly small number of websites use content negotiation by language. The new W3C website design is not using it, for example, and is instead going for a {domain}/{lang}/{page} structure. So, yes, having a single link to a resource that is made available with language-based conneg is probably an edge case hanging on by its fingernails.

Even so, it is part of HTTP and so leaving it out might be problematic, process-wise if nothing else? Apologies if I've misunderstood your comment.

Phil

Phil Archer
Director, Web Solutions, GS1
https://www.gs1.org

Meet GS1 Digital Link Developers at
https://groups.google.com/forum/#!forum/gs1-digital-link-developers

https://philarcher.org
[phone]
@philarcher1
Skype: philarcher

On 08 July 2021 15:16, httpapi wrote:
Am 04.07.2021 um 14:10 schrieb Erik Wilde:
> hello rich.
>
> On 2021-06-08 16:17, Salz, Rich wrote:
>> So I think that this is the plan we should follow.  Julian will post
>> his comments (ideally this week, perhaps next) and the authors will
>> respond and perhaps post a new draft. If there's disagreement or
>> controversy, we'll repeat that cycle until issues are settled.  Once
>> that's done, we'll issue a working group last call (WGLC) and proceed.
>
> we just published a new version after resolving all open issues
> without any remaining disagreements. we think that we now are ready
> for the next step, and we're curious to see the result of the WGLC.
>
> thanks a lot to julian for all his valuable and constructive feedback,
> it helped to improve the draft.
> ...

Here's some more...

The one-two-one mapping between link header field syntax and JSON makes the JSON really complex. I understand that this is done for round-tripping.

There is however one case where I would argue that complexity could be removed without major loss. See <https://www.ietf.org/archive/id/draft-ietf-httpapi-linkset-03.html#section-4.2.4.2>
which is about mapping RFC 8187 syntax, which allows inclusion of language tags. I'm not aware of this being used in practice (and I'm author of that RFC). Dropping the language information from JSON would simpliy the datamodel somewhat.

Best regards, Julian

--
httpapi mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/httpapi

CONFIDENTIALITY / DISCLAIMER: The contents of this e-mail are  confidential and are not to be regarded as a contractual offer or acceptance from GS1 (registered in Belgium). 
If you are not the addressee, or if this has been copied or sent to you in error, you must not use data herein for any purpose, you must delete it, and should inform the sender. 
GS1 disclaims liability for accuracy or completeness, and opinions expressed are those of the author alone. 
GS1 may monitor communications. 
Third party rights acknowledged. 
(c) 2020.