W3C home > Mailing lists > Public > public-publ-wg@w3.org > July 2017

Re: addressable identifier?

From: <laurent.lemeur@edrlab.org>
Date: Thu, 27 Jul 2017 22:52:11 +0200
Cc: Hadrien Gardeur <hadrien.gardeur@feedbooks.com>, W3C Publishing Working Group <public-publ-wg@w3.org>
Message-Id: <0C6EDCB6-2E68-4FAB-99F4-BCD3520038BA@edrlab.org>
To: Romain <rdeltour@gmail.com>
. As the manifest is the only resource specific to a WP, its address is a logical choice.
> OK, makes sense.
> But then, the URL to a manifest may change, or two different URLs can point to the same manifest:
> https://resilientwebdesign.com/manifest.json
> https://resilientwebdesign.com/books/../manifest.json
> (dumb case, but you see the point).

I didn't want to raise this uniqueness issue before we agree first on an answer to your question but uniqueness is not really required neither imo.  An unambiguous identifier (meaning that two different WP can't have the same I'd) is sufficient. This idea was discussed in the NewsML G2 WG some years ago, I may find some traces of that. 
Or we can define this id as a "canonical" URL (I Imagine there is such a thing in the W3c toolkit) 
> All I'm saying is I understand the manifest _can_ be used as an identifier in a given context (e.g. a UA), but there's nothing to say about that in the spec, right? There's a URL, people can use that as an identifier or not.
> As far as I can tell, there's no such concept in web sites and apps, and I don't see that publications need anything other than a loosely specified "dc:identifier" property in metadata. (Maybe I'm wrong, I'm just not convinced yet :-)

It was my first view on this. A metadata simply defined as an unambiguous Id. But giving a bit more precise indications may help adoption and avoid ambiguities. 


Received on Thursday, 27 July 2017 20:52:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:52:14 UTC