W3C home > Mailing lists > Public > public-xg-webid@w3.org > November 2011

RE: henry and code

From: Peter Williams <home_pw@msn.com>
Date: Thu, 17 Nov 2011 08:29:14 -0800
Message-ID: <SNT143-W40C1E7D1417A847519ED5592C70@phx.gbl>
CC: "public-xg-webid@w3.org" <public-xg-webid@w3.org>


Henry,  Let me guess that foaf.me (#15) is one of the 15 implementations, as is the test suite (#14). #12 I hereby tie to virtuoso (which is the first commercial-grade product in my list). foaf.me lets me make a foaf card, and I see it still streams it from its endpoint(s). Do take 1m to use the sevice in as minimal way as possible, and take a look at the foaf.me outputs please, and tell me for the expectations of the community, as of today, (a) is the RDF/XML conforming to the spec(b) is the RDFa conforming to the spec while focussing mostly on the only thing I need to really care about: the identity credentials. In my implementation of an Validation Agent, I will be focussing only on IDewntity Agents whose identity credentials are in an RDF/XML formatted document , since the target of conformance requires that the Validation Agent minimally support only one. Ideally, a run of the test suite would give the assurance I ask of you, verbally. Perhaps there is a recent test report to peruse on the foaf.me target, collected in a repository along with other reports?   Looking at http://www.w3.org/2005/Incubator/webid/wiki/Test_Suite, 4 test candidates are identified, and foaf.me is not on the public part of the list. It may have no qualification certificate. It does seems a good step to re-run the test suite against foaf.me and other, having run the self-test of the test suite itself. Once I have an unreprachable stream, ill go find a wordpress or joomla plugin that already streams out RDF/XML today for some purpose (e.g SIOC community), and alter the php so it streams out my card, leaving all other web behaviours unchanged from the default install (with plugin) of the core platform. Ill restrict my implementation to interworking with one browser vendor, operating on Windows. It will be the current release of Mozilla, and will not be a beta. --------- Concerning the test suite, the blog posting at http://fcns.eu/2011/04/fcns-webid-test-suite/ indicates that it too will create a foaf card. It seems reasonable to assume that a download and operation of that software will produce a card which is unimpeachable, current and conforming. This sill be my fallback, used if foaf.me cannot offer assurance that it is a conforming source of identity credentials. I already see that the author of the test suite site also offers https://github.com/WebIDauth/WebIDauth. While is useful to see the implementation of a Validation Agent, I will not be implementing any protocols other than one specified in the Editors Draft. If there is huge deviance between what the spec says and what the test suite rules encode, we will have been able to assess the quality of the spec in terms of consistency with the test suite. We will have a hard engineering metric which asseses the work of the W3C incubator, at that point.From: henry.story@bblfish.net
Date: Thu, 17 Nov 2011 15:30:14 +0100
CC: kidehen@openlinksw.com; public-xg-webid@w3.org
To: home_pw@msn.com
Subject: Re: henry and code




On 17 Nov 2011, at 04:16, Peter Williams wrote:Now, there are 15 other implementations to review, too. so I see what everyone else has done. I can compare and contrast these. Ill write it all up on my blog site, as I find them. I can soon assess the maturity, if they are in source. Ill be as critical as I can, so we can see what its true and what is hype. Ill try to match the third quartile, in demonstrability of what the webid spec actually says. Hows that?
Doing reviews of other people's code before you have the following two pieces seems a bit odd. Furthermore we are going to automate the review process of endpoints, so you would be wasting your time doing that.
There are two things you should have:
1. Your own webid - written by hand even - I'd suggest doing this first. Get yourself a server space, you'll need that if you are serious with 2.2. A server that is running you service that is webid enabled that we can authenticate against 
If you do 2, then the next thing is to have some page  on that server produce EARL output as explained here   http://www.w3.org/2005/Incubator/webid/wiki/Test_Suite
Then we can develop test suites to automatically review your implementations and produce reports on them. You can already use Bergi's code to verify your endpoint now. This needs to be turned into a service.
  Henry


Social Web Architect
http://bblfish.net/


 		 	   		  
Received on Thursday, 17 November 2011 16:29:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 17 November 2011 16:29:49 GMT