[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xdi] Minutes: XDI TC Telecon Thursday 1:00-2:30PM PT 2012-01-12
Granted I've been out of XDI for the past few months due to work, but I thought the situation when I last checked in was that composition was performed via a predicate (=Bill+home/+phone). This is in contrast to specialization, which was performed via either a plus-narrowing (+homephone instead of +phone) or more preferably delegation (+phone+home). I say preferably because of my own bias to for the view of everything in XDI as sets, so +phone+home represents the elements of +phone that are also elements of +home. This matches other discussions we've had, that $$S/$$P/$$O <==> $$S/$has/$$P, which does not imply anything about $$S$$P. Also $$S$$P ==> $$P/$is/$$S, but not $$S/$is/$$P.
Does the above match current TC consensus? If not please explain how it differs so I can get back up to speed.
From: xdi@lists.oasis-open.org [xdi@lists.oasis-open.org] on behalf of Giovanni Bartolomeo [giovanni.bartolomeo@uniroma2.it]
Sent: Friday, January 13, 2012 5:40 AM To: Markus Sabadello Cc: Drummond Reed; OASIS - XDI TC Subject: Re: [xdi] Minutes: XDI TC Telecon Thursday 1:00-2:30PM PT 2012-01-12 Oh, yes that document! (Thanks Markus!)
The original idea was to merge patterns and constraints (e.g., the FROM and the WHERE clause in SeRQL, http://www.openrdf.org/doc/sesame/users/ch06.html) into one single xdi graph, allowing this latter to contain logical expressions. About cross referencing the subject: http://lists.oasis-open.org/archives/xdi/200803/msg00071.html this was raised from the ambiguity about the order of evaluation of subsegments. In a longstanding old thread ("associativity of $has") I've always supported the idea that the composition operator is left associative (with "left associative" intended as reported here: http://en.wikipedia.org/wiki/Operator_associativity) - and I've still this position (I know Drummond has a different one). In Markus' original example however the problem was that +phone+home is not a composition, rather it is a specialization (should be +homephone, with somewhere else stated that +homephone is a subproperty of +phone). Model it as a composition result in the ambiguity problem Markus had outlined. Best Regards, Giovanni Def. Quota Markus Sabadello <markus.sabadello@xdi.org>: > I think this document is what I used back then as a basis, plus a few > e-mail threads > > http://www.oasis-open.org/committees/download.php/27547/Draft-Contrib-OASIS-XDI-dollar-words-v03.pdf > > > http://lists.oasis-open.org/archives/xdi/200804/msg00000.html > http://lists.oasis-open.org/archives/xdi/200804/msg00016.html > http://lists.oasis-open.org/archives/xdi/200803/msg00071.html > ... > > Markus > > On Thu, Jan 12, 2012 at 11:56 PM, Drummond Reed <drummond.reed@xdi.org>wrote: > >> Following are the minutes of the unofficial telecon of the XDI TC at: >> >> Date: Thursday, 12 January 2012 USA >> Time: 1:00PM - 2:30PM Pacific Time (21:00-22:30 UTC) >> >> ATTENDING >> >> Mike Schwartz >> Giovanni Bartolomeo >> Markus Sabadello >> Drummond Reed >> Joseph Boyle >> >> REGRETS >> >> Bill Barnhill >> >> >> THE IDEARPAD LINK FOR TODAY IS: >> http://piratepad.net/3GsPcga5is >> >> >> 1) IDTRUST BUDGET >> >> Great news - the budget request for $5K for interop testing of an XDI >> client operating against two XDI servers was approved! Dee Schur has sent a >> contract to Markus to review and will have a call with him to go over it. >> >> We discussed the configuration that we would use for it. Markus and Mike >> will coordinate the details. >> >> # MARKUS AND MIKE will work out a plan for executing and keep the TC >> appraised of progress. >> >> >> 2) OPEN SOURCE PROJECT UPDATES >> >> Mike said that the OpenXDI Project released version .2, and is now working >> on the front end needed by MIT. Work now will go into OxAuth. >> >> The next improvements on OxServer will be optimizations on search, plus >> work on the OGM (Object Graph Mapping) tool. >> >> Markus said he is beginning work on changes to the Project Danube code >> base, and will have more to update next week. >> >> >> 3) LINKED DATA UPDATE >> >> Giovanni said that Linked Data is becoming the defacto standard for >> sharing data, and governments are starting to look at it as the standard >> for publishing data. He said the Italian Government (after the US, UK and >> French Governments), for example, is exploring it now. >> >> Giovanni believes that XDI should be 100% compatible with Linked Data. The >> key question that he would like to have researchers explore is defining >> what XDI can add to Linked Data. The first point of addition is the ability >> to address single nodes inside the graph. The second point is control of >> data sharing using link contracts. >> >> Giovanni explained that Linked Data implementers approach the problem of >> identification from a different perspective. For example, two different RDF >> graphs from two different authorities may use different URIs to represent >> the same entity. What is emerging now is the pattern of using hubs to share >> URIs that represent a specific entity (including a concept, such as an >> ontology definition, or a place, or a product, etc.) >> >> If this trend continues, of referencing/sharing a single "very popular >> URI", then Giovanni believes we can map to that URI from an XRI using >> either (or both): >> >> - A cross-reference >> >> >> - A short-form XRI, which would provide a single short-form >> "canonical" identifier for the entity. >> >> >> Giovanni feels that if we take this approach, we can help solve the >> problem of disambiguating attributes assigned to the same entity >> (identified by a popular URI/XRI) in two different contexts from two >> different RDF authoritities. The solution is to make each set of attributes >> relative to each of the two authorities using the contextual capability of >> XDI. >> >> Giovanni is currently working using statistical techniques to analyze part >> of the Linked Data cloud (which is up to 32 billion Linked Data triples) to >> identify what appear to be the same entity in different contexts. Once >> these are identified, then the XRI can be assigned that would become the >> canonical identifier for that resource. >> >> Using this approach -- this XDI dictionary "hub" that establishes the >> equivalence of different URIs -- software that uses the hub can do the >> mapping across RDF graphs, and then the next step is that the data >> publishers can start using longer URIs that incorporate XRIs into their >> URIs - making them longer but also embedding the XRI identifier in such a >> way that the equivalence can be recognized by processing agents. >> >> Mike pointed out that by creating URI "hubs", the Linked Data community is >> already starting to add a data abstraction layer and create canonical >> identifiers. Giovanni agreed and said that because this is happening, it >> sets up the business case for them to start to use what is specified in the >> XRI 3.0 spec, which is to use a concrete http: or https: URI to identify >> the provenance of the RDF graph bound to an XRI to identify the entity. >> That is a very elegant solution that provides a bridge between the RDF >> Linked Data world and the XDI world. >> >> # GIOVANNI will work out a roadmap for how he sees this unfolding, and >> then post an update. >> >> >> 4) AUTO-INCREMENT >> >> Mike sent this message to the list: >> >> http://lists.oasis-open.org/archives/xdi/201201/msg00006.html >> >> He explained that the use case is very common where you need the server to >> add the next available i-number for identifying an node. >> >> Drummond and Markus immediately recognized this pattern as something they >> had to deal with in a Higgins-related project several years ago. The >> solution that we developed was to send the server a $add operation with a >> variable in the XRI for the XDI object. The server would see the variable >> and then replace it and return the value to the client. >> >> Here is what we agreed on the call, to use variables: >> >> <$sender>/$add/(=schwartz/()/($1))<http://piratepad.net/ep/search?query=schwartz/%28%29/%28$1%29%29> >> <$sender>/$add/(=schwartz/()/($2))<http://piratepad.net/ep/search?query=schwartz/%28%29/%28$2%29%29> >> <$sender>/$add/(=schwartz($1)/()/+friends)<http://piratepad.net/ep/search?query=schwartz%28$1%29/%28%29/+friends%29> >> <$sender>/$add/(=schwartz($2)/()/+friends)<http://piratepad.net/ep/search?query=schwartz%28$2%29/%28%29/+friends%29> >> >> The server would assign the next available unordered index and returns it >> in the response to the client. >> >> >> 5) PROPOSALS FOR OPTIMIZING $GET - XDI QUERIES >> >> Continuation of last week's discussion. See: >> >> http://ox.gluu.org/doku.php?id <http://ox.gluu.org/doku.php?id=xdi:get> >> =xdi:get <http://ox.gluu.org/doku.php?id=xdi:get> >> >> Drummond and Markus brought up that a basic query syntax was developed a >> few years ago and actually implemented by Markus. Originally developed by >> Giovanni and with some patterns suggested by Bill, it was very simple yet >> very elegant. >> >> # MARKUS agreed to look up the source documents and help port that syntax >> forward to the XDI TC wiki page(s) that Mike is creating for a query syntax. >> >> Mike suggests we also discuss via email to be ready for a live discussion >> on next week's call. >> >> >> 6) XDI MESSAGE ORIGINATION SOLUTION >> >> Drummond said that he had added the basic solution as the revised first >> line (the "from" line) of the XDI message template: >> >> http://wiki.oasis-open.org/xdi/XdiMessagePatterns >> >> We ran out of time to discuss it on this call, but Drummond will add more >> documentation, and we will discuss it next week. >> >> >> 6) NEXT CALL >> >> Although Drummond will be in Europe, he still hopes to be able to attend >> the call, and in any case it will be next week at the regular time. >> >> >> >> >> >> >> >> >> >> >> >> > ---------------------------------------------------------------- Invito da parte dell'Ateneo: Il tuo futuro e quello della Ricerca Scientifica hanno bisogno del tuo aiuto. Dona il 5 x mille all'Universita' di Roma Tor Vergata codice fiscale: 80213750583 http://5x1000.uniroma2.it --------------------------------------------------------------------- To unsubscribe, e-mail: xdi-unsubscribe@lists.oasis-open.org For additional commands, e-mail: xdi-help@lists.oasis-open.org |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]