- From: Graham Klyne <GK-lists@ninebynine.org>
- Date: Sat, 10 Jul 2010 15:41:14 +0100
- To: Jonathan Rees <jar@creativecommons.org>
- CC: www-tag@w3.org, "Henry S. Thompson" <ht@inf.ed.ac.uk>
I looked at this briefly as part of reviewing the Accept-datetime header field.
At the time, it struck me as being formulated quite cleanly in terms of Web
architecture as I understand it, but I couldn't guarantee there aren't any
lurking dragons.
#g
--
Jonathan Rees wrote:
> Herbert Van de Sompel gave me the following useful pointers regarding
> Memento ("time travel for the Web"):
>
> For those interested in reviewing the proposal at this moment
> in time, and willing to invest 20 minutes in doing so, I think a
> combination of http://www.mementoweb.org/guide/quick-intro/ and
> http://www.mementoweb.org/guide/http/ should provide enough material.
>
> The idea is that for a given resource R there can be a "timegate"
> resource Q having the property that a representation Z is a current
> representation of Q if and only if it at some time t (past or present)
> Z was/is a representation of R. You use Accept-datetime: on a GET Q
> request to ask for a past representation of R.
>
> In the language of HTTPbis, all the representations of Q are the "same
> information" [1] when they come from GET Q even when they're not the
> "same information" when they come from GET R.
>
> (This is only a teaser. Do not critique the protocol based on this
> possibly mangled description. Look at the pages cited above.)
>
> This relates to our ISSUE-53 (generic resources) and ACTION-231.
>
> Jonathan
>
> [1] http://tools.ietf.org/html/draft-ietf-httpbis-p3-payload-09#section-4
>
Received on Saturday, 10 July 2010 14:42:22 UTC