- From: Michael Brunnbauer <brunni@netestate.de>
- Date: Mon, 26 Mar 2012 21:42:40 +0200
- To: Tim Berners-Lee <timbl@w3.org>
- Cc: Leigh Dodds <leigh@ldodds.com>, David Wood <david@3roundstones.com>, David Booth <david@dbooth.org>, Jeni Tennison <jeni@jenitennison.com>, public-lod community <public-lod@w3.org>
hi all the fact that many HTTP clients do not support 303 properly is new to me and I retire my statement that we should leave everything as it is. I have some questions about what would break to the supporters of the change proprosals. Missing bookmarkability of NIRs can be a feature also. Imagine a FOAF app asking the user about the address of his homepage, weblog or pages of interest. The corresponding FOAF properties have a range of foaf:Document. foaf:Document seems to be not a subclass of the IR class (it intentionally includes physical documents) and I am not sure if Dan would see the IR class as a subclass of foaf:Document but surely most IRs are foaf:Documents. So with the current state, the probability that the user enters something that is not a foaf:Document is reduced to a minimum. What will become of this app in the future ? Will it have to check for hint's that the URI is a "IR" (which should be enough to assume a foaf:Document) ? Will it have to warn the user if it finds no hints ? Or lets assume a crawler that crawls HTML pages, does a text analysis and then generates triples like these: <x> a foaf:Document <x> dcterms:subject ex:skosconcept1 <x> dcterms:subject ex:skosconcept2 ... Will people in this community actually refuse to write such a crawler after one of the proposals has gone through ? Regards, Michael Brunnbauer On Mon, Mar 26, 2012 at 01:27:27PM -0400, Tim Berners-Lee wrote: > > On 2012-03 -26, at 06:18, Leigh Dodds wrote: > > > > > I may be misreading you here, but I'm not against unambiguous > > definition. My "show what is actually broken" comment (on twitter) was > > essentially the same question as I've asked here before, and as Hugh > > asked again recently: what applications currently rely on httprange-14 > > as it is written today. That useful so we can get a sense of what > > would break with a change. So far there's been 2 examples I think. > > For me, the fact that you can use the URI of a document on the web > to refer to that document is so built into the semantic web architecture for the last 12 > years that it has been implicit in everything I have coded or designed, > haven't been aware of where I have used it and where not. > It it is really difficult for me to measure which bits would > The SWAP project, CWM has that built in -- the URI of the document > something was read from is kept in the quad store as provenance > for every triple read in. The same with the tabulator store. > The tabulator offers different views of objects as a function of the classes > the are in, and it infers things from HTTP 200s and content types. > It would have to be re-engineered of course, something I'm prepared > to do in the cause of progress, but I feel it had better be something which > adds bath water without throwing out the baby. > > Tim > > PS: Missed the tweet -- ++ Michael Brunnbauer ++ netEstate GmbH ++ Geisenhausener Straße 11a ++ 81379 München ++ Tel +49 89 32 19 77 80 ++ Fax +49 89 32 19 77 89 ++ E-Mail brunni@netestate.de ++ http://www.netestate.de/ ++ ++ Sitz: München, HRB Nr.142452 (Handelsregister B München) ++ USt-IdNr. DE221033342 ++ Geschäftsführer: Michael Brunnbauer, Franz Brunnbauer ++ Prokurist: Dipl. Kfm. (Univ.) Markus Hendel
Received on Monday, 26 March 2012 19:43:05 UTC