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

Re: Can a publication change over time?

From: Baldur Bjarnason <baldur@rebus.foundation>
Date: Thu, 27 Jul 2017 13:31:56 -0400
Cc: Hadrien Gardeur <hadrien.gardeur@feedbooks.com>, Laurent Le Meur <laurent.lemeur@edrlab.org>, "public-publ-wg@w3.org" <public-publ-wg@w3.org>
Message-Id: <80541CE8-DCC4-4D05-8D0D-D45B12F53A94@rebus.foundation>
To: Garth Conboy <garth@google.com>
It’s pretty fundamental to the web that its resources be stateless, loosely coupled and not be tied to specific, immutable versions of its subresources.

The nearest analogue (sub resource integrity) almost inevitably leads to broken publications, which is desirable in a security context but isn’t what you’d want from publications, so doesn’t really address your concern.

The web community has been debating versioning and immutability in HTTP and HTML from the moment TBL created them. Probably longer because those concepts predate the web and are an established part of hypertext academic theory.

The closest we have to a workable solution to this is Memento[1] and that is basically a non-starter outside of a tiny minority of the web. Memento’s probably better designed than anything we could come up with and it’s not going anywhere to speak of. There have been quite a few other attempts. "Link Relation Types for Simple Version Navigation between Web Resources”[2] is another but that also didn’t go anywhere. I could dig up more.

All we can realistically accomplish here is to outline a mechanism where publishers can provide metadata of the sort that Hadrien describes. Even then that’s only going to be advisory/informative since the publisher still has the ability to modify those resources and if the UA is going to reject changed resources we’re basically back to using the SRI spec. 

[1]: https://www.w3.org/blog/2016/08/memento-at-the-w3c/
[2]: https://tools.ietf.org/html/rfc5829

- best
- Baldur Bjarnason

> On 27 Jul 2017, at 12:59, Garth Conboy <garth@google.com> wrote:
> Indeed, that may be desirable.
> But, our mission is to bring publishing to the Web.  It's pretty fundamental to (traditional) publishing to write or cobble together a set of "resources" and have said publication not change (or at least know if it changes) over time.  We may end up not being able to find a viable complete equivalent on the Web, but lets not run away too soon, as there are clearly technologies that could possibly be applied.
> Best,
>    Garth
> On Thu, Jul 27, 2017 at 9:17 AM, Hadrien Gardeur <hadrien.gardeur@feedbooks.com> wrote:
> > If you start messing around with immutability, update mechanisms, and resource versioning, you get pretty deep into dark corners of the HTTP ecosystem pretty quickly.
> +1, which is why I'd like to avoid such requirements
Received on Thursday, 27 July 2017 17:32:33 UTC

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