Also note that Slug is defined very loosely in terms of syntax and needs additional processing on the server to derive a proper url segment. In my implementation, I typically perform Unicode NFKD normalization, replace common whitespace with dashes (-) and strip remaining non-ascii (A-Za-z0-9) characters. This typically yields acceptable results. YMMV, however. - James On Mon, Nov 5, 2012 at 1:35 PM, Wilde, Erik <Erik.Wilde@emc.com> wrote: > hello richard. > > On 2012-11-05 13:08 , "Richard Cyganiak" <richard@cyganiak.de> wrote: > >On 5 Nov 2012, at 20:40, David Wood wrote: > >> Our resolution to ISSUE-20 [1] left me wondering about compliance with > >>regard to clients providing a (presumably optional) AtomPub-like Slug: > >>header (as Callimachus does). > >> > >> [1] http://www.w3.org/2012/ldp/track/issues/20 > >I think the question of user-supplied ³slugs² sort of fell under the > >table when we discussed ISSUE-20; we focused on the issue of relative > >URIs. > > +1 to look at slugs or slug-like concepts. it is a useful implementation > of the concept of providing clients with some control over the namespace, > but still giving servers final control. it's important to keep in mind, > though, that slugs are meant mainly to control centralized naming (the > header determines the final path component of the created URI), so for > supporting decentralized scenarios (where you might want to supply the > canonical URI that's not under the control of the LDP server), this > probably could not be directly adopted. > > cheers, > > dret. > > >
Received on Monday, 5 November 2012 21:43:28 UTC