SV_MEETING_TITLE -- 14 Apr 2010

<JeniT> http://www.epimorphics.com/public/presentations/ldapi/egov-presentation.html#(1)

<JeniT> I'm on the same line as John

<TomM> Conference call is full

<sandro> checkiing on it TomM

<TomM> thx

<DaveReynolds> @olyerickson asked about streaming audio ... maybe another time :)

<TomM> now conference is restricted...

<sandro> still working on it! sorry!

<TomM> thx

<sandro> yeah. :-(

<sandro> I'm so sorry everyone. I thought this was set up properly.

<sandro> Hopefully just another minute or two.

<Cory> Ok, I get back in the loop andd the loop is closed!

<edsu> sandro: no worries, it's a good problem to have imho :-)

<juansequeda> mhausenblas, really? I was expecting this :)

<Cory> I can set up a gotowebinar

<edsu> anyone know how many people are on the call? (that aren't in irc)

<sandro> Yeah, 10. People with w3c member access can see the list of all bridge users here: http://www.w3.org/1998/12/bridge/Zakim.html

<johnlsheridan> how will we know when we can start?

<johnlsheridan> ok

<johnlsheridan> I think it might suit JeniT

<JeniT> Suits me. I hate talking on the phone at the best of times ;)

<JeniT> hi libby :)

<libby> hey jenit!

<libby> and all

<Chris> Evening all

<sandro> hey libby!

<libby> hey sandro :-)

<sandro> waiting, waiting. :-)

<sandro> did that do it?

<JeniT> no

<ivan> aaaahhhh:-)

<Chris> meh - Probs with conf call sandro?

<sandro> Yes, Chris :-( :-(

<johnlsheridan> when we start, will people want to introduce themselves on the call, or just skip straight to the presentation?

<Chris> @sandro no stress - how long till fix?

<sandro> (and totally my fault, for setting up the teleconference reservation incorrectly.)

<johnlsheridan> normally 10 people is enough for us :)

<TomM> twitter-effect.... :)

<libby> hey dave!

I guess so, TomM - re Twitter http://search.twitter.com/search?q=egov+linked+data

<Cory> Sandro - I have a gotomeeting webinar running, if needed

<sandro> Thanks, Cory. In 10 years at W3C I've never had a problem, except once when the bridge's second redundant power supply blew out.

<Cory> Its the every 10 year problems that get you!

<sandro> [The admin I'm talking to now isn't answering me.... hopefully that means she's getting up, or something. But the real expert on the system is away helping with local flood relief operations.]

<sandro> er, getting *help* I mean

<edsu> sandro: so would that be ralph who is helping out w/ the flood relief

<johnlsheridan> I'm being organised by JeniT (again)

<Chris> lol

<johnlsheridan> ok

<Cory> On another note - I would like to try and spool up "outreach" again, looking for ideas and input

<johnlsheridan> can we try and dial back now?

<sandro> yes.

<edsu> dang somebody sounds like they're in a laundromat :-)

<edsu> sandro++

<ivan> isn't technology great?

<sandro> OKAY TO CALL NOW!!

<edsu> ivan: :-)

<TomM> :)

<Cory> So now we are goint to tell the world how to be interconnected, right!

<JeniT> http://www.epimorphics.com/public/presentations/ldapi/egov-presentation.html

<edsu> mhausenblas: sounds like you are volunteering :-)

<daniel> .mhausenblas... this would be good

<edsu> mhausenblas++

<johnlsheridan> +1 - sorry, I should have asked for a scribe!

<olyerickson1> for thoe of us IRC only, please tell us the current slide...thanks!

<sandro> still on slide 2 I believe

<sandro> slide: 3

JeniT: goal is a middleware that hides away RDF/SPARQL complexity

<Chris> @sandro --> PM.

<sandro> slide: 4

<daniel> s, middleware transformation to rdf-based systems is out of scope?

JeniT: the middleware takes HTTP request and wraps the SPARQL endpoint
... idea is that people deploy the middleware with API spec and serves multitude of formats
... make customisation easy

slide 5

JeniT: consumer side
... the RDF format we had are 'ugly' and require changed tool chain

<sandro> slide: 6

JeniT: so in order to simplify the job we remove round-tripping and provide simple JSON
... query
... SPARQL was perceived too complex
... people want to list based on URIs

<sandro> (great stuff, JeniT.... makes a lot of sense.)

juansequeda: what are the 'ugly' formats

<daniel> what is the role of REST?/is there one?

JeniT: RDF/XML
... aim is to make it a RESTful protocol
... get list of things
... read-only ATM

<danbri> is the idea for this API to have multiple implementations, or is it mainly an implementation?

<danbri> <- question in irc as my audio is bad

JeniT: idea is to have a generic spec with many impls

<edsu> i guess this is the specification currently? http://code.google.com/p/linked-data-api/wiki/Specification

DaveReynolds wil now explain how it works behind the scenes

<JeniT> Yes, that's right edsu

DaveReynolds: slide 7
... the middleware takes HTTP request and figures how to translate it to SPARQL queries
... the vocab piece is needed to map types, labels, etc.

slide 8

<daniel> does the mapping have a link to e.g. the d2r-server mapping stuff

DaveReynolds: middleware == selector, viewer, formatter

(DaveReynolds explains details re the three components)

slide 9

DaveReynolds: first you need to define the endpoint
... for a SPARQL endpoint you define 1..N API endpoints
... and then a view definition

slide 10

DaveReynolds: using it
... GET on http://epimorph-pubx1.appspot.com/proxy/api/doc/schools

slide 11

DaveReynolds: you get back the top-level resource
... now in RDF/Turtle

slide 12

DaveReynolds: important that we return a list of resources not a set of "bindings"
... conserve the resource-centric nature of linked data

<danbri> (this is a very healthy emphasis... the resource-centric aspect)

DaveReynolds: further, it's paged
... and finally multiple representations (JSON, RDF etc)

slide 13

DaveReynolds: just put a .json on the end and you get a JSON representation
... result is an array which conserves the Linked Data structure

slide 14

DaveReynolds: goal was to keep the JSON simple/natural
... uniform tree structure
... also gives context information (mapping, etc.)

slide 15

DaveReynolds: filtering
... you can narrow down by changing URI

slide 16

DaveReynolds: sublists and patterns
... package useful filters into endpoints
... leverage URI patterns

slide 17

Michael: wondering if the template is based on http://tools.ietf.org/html/draft-gregorio-uritemplate

DaveReynolds: assisting with formulating queries

<JeniT> mhausenblas: no, simpler, because it's for parsing rather than generation

hmmm

<JeniT> mhausenblas: rather, it's *based on*, but not everything from

Michael: thanks JeniT - I see

<juansequeda> what is the link to the spec?

<DaveReynolds> Spec is at: http://code.google.com/p/linked-data-api/wiki/Specification

JeniT wraps up

JeniT: something we found we really needed

<Cory> we did something similar bit were concerned about acceptance: http://www.modeldriven.org/projects/EKB/XMLView.ppt

<sandro> if you have questions, please use "q+" on irc

JeniT: normal consumers (Web develeopers the scribe thinks?) not willing to handle RDF
... it simplifies many tasks
... no need to learn SPARQL

<daniel> sidenote : is there a way to push the json-output to the libs such as jena2?

JeniT: through preserving the linked data aspects the users can still extend/learn more
... other side is helping the publishers of linked data

<DaveReynolds> daniel - json-output is in an open source project and works with Jena

JeniT: simple config - rich API

<daniel> DaveReynolds: ok, thx - do you have a link for that?

<DaveReynolds> daniel - see http://code.google.com/p/linked-data-api/source/checkout

<DaveReynolds> Not yet formally released and supported ! :)

libby: really impressive stuff
... are people now complaining less?

JeniT: have presented this on certain meet-ups and very positive feedback, yes

libby: does this include also non-RDF-people?

<libby> thanks ;-)

JeniT: most are sort of RDF aware,

juansequeda: people often think RDF == XML, what's you experiences with this?

<Cory> Objection 1: RDF is XML. Objection 2: RDF isn't XML

juansequeda: future plans?

<edsu> ok :-)

<juansequeda> edsu, it is me :)

<edsu> mhausenblas++

<JeniT> PHP implementation: http://code.google.com/p/puelia-php/

<libby> heh

<Zakim> danbri, you wanted to ask (if someone can relay this into audio) ... any interest in going further towards full text search, ranking, auto-completion ('give me schools whose

<JeniT> Ruby implementation: http://github.com/ldodds/linked-data-api

<danbri> danbri: q+ to ask (if someone can relay this into audio) ... any interest in going further towards full text search, ranking, auto-completion ('give me schools whose names/labels begin with...')...

<JeniT> Java implementation: http://code.google.com/p/linked-data-api/

<juansequeda> JeniT, any plans of coming to the US anytime soon?

<Zakim> edsu, you wanted to ask slide 13

<danbri> thanks dave!

DaveReynolds: there are some hooks for these things danbri but no active developments ATM

JeniT: we support extension points in the middleware

<danbri> (i have a hunch auto-complete will end up being important...)

<edsu> http://epimorph-pubx1.appspot.com/proxy/api/doc/schools.json

edsu: question re slide 13
... so you have a list of schools
... what about a link back to the API to get the JSON back

<JeniT> juansequeda, I'm over for WWW2010 and then for Semtech 2010 :)

DaveReynolds: at the top level (next, previous page) there is also the alternative formats
... same works for item-level

<juansequeda> JeniT, great Looking forward to finally meeting you at WWW2010 and at Semtech :) Any free time? Would you like to come to Austin?

DaveReynolds: we have not gone so far to offer it for every resource, no

<JeniT> juansequeda, probably won't have time :( But hope we can meet at the conferences

Cory: you mentioned XXX - plans for it?

<Cory> update = update

DaveReynolds: current motivation and reqs stems from data.gov.uk

<edsu> mhausenblas: i asked the question awkwardly, i meant to ask if the spec would allow people to follow their nose in the json data, so individual schools in that response would have a linked-data-api uri to get the item level json for a school

DaveReynolds: re updates - we'd be open to any ideas, currently not supported

JeniT: have seen such a requirement already, recently

Michael: notes http://esw.w3.org/WriteWebOfData and happy to follow up with you chaps

<edsu> does irc.w3.org respond on port 80?

<DaveReynolds> You're welcome :)

<ivan> clap clap clap

<danbri> edsu, it seems to talk irc: on :80, yes

<danbri> i hadn't thought to look before :)

<ivan> ivan drop me

johnlsheridan: thanks for presenting JeniT and DaveReynolds

<laurent_lefort_cs> * laurent claps

<libby> thanks all

sandro: any announcements?

<kate_geyer> Kate from MA, US, staying until the end

sandro: DCAT will start its work tomorrow
... meeting in two weeks will host speakers from New York state legislature

Cory: in outreach want to find out in what people are interested in
... will do some work in Wiki and send out email

johnlsheridan: will talk about in chairs group but would like to update the note with recent activities

BrianG: some uptake of catalogue in Colorado

OwenAmbur: there are two other activities

scribe can't understand OwenAmbur - anyone?

<OwenAmbur> IAC is conducting a social media best practices study.

<OwenAmbur> They will be interviewing a number of .gov officials for the study.

[adjourned]

<OwenAmbur> Under the auspices of IBM's Center for the Business of Government, Professor Pat Franks is also conducting a similar study.

- DRAFT -

SV_MEETING_TITLE

14 Apr 2010

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output