BioRDF -- 7 Jan 2008

<Susie> Speaking to Josh to try to get it fixed...

oops, is this hcls time?

<Don> I'm having trouble calling in...

<mscottm> was getting "this passcode is not valid".. trying again

<Don> it says that the code 246733 is not valid.

<BillBug> same here

Susie, i'm next door -- can i help somehow?

<mscottm> Something to do with the new year? reschedule conferences?

Don, mscottm, BillBug, our top scientists are working on it as we speak

<Don> Thank you Eric. Happy New Year everyone!

<BillBug> Very comforting :-)

<BillBug> Thanks, Eric.

<mscottm> "I'm doin the best I can Captain!"

<Susie> Things are being set up. It should just be another minute or so now.

"damnit jim, i'm a doctor, not an operator"

thanks should go to Susie who's actually dealing with the problem

i'm just buying time by entertaining y'all

<kei> I have problem dialing in

<alanr> what's the call in code?

kei, yes, we're working on that now

alanr, zakim is ill. josh is working on it now

<alanr> k

<jar> as alan is on the call i think i'll pass today. enjoy -Jonathan

ok all

<matthiassamwald> +??P16 is me.

sribe: ericP

<alanr> scribe: ericP

Susie: focus on documentation

KB documentation

mscottm: had goal of draft by end of year
... didn't work out, but a lot has happened on the doc
... up 'till the 22nd, several of us met on the HCLS telecon slot and made decisions about the doc
... most of that stuff has been edited in
... i've been on vacation and just returned
... Eric documented an issue

<mscottm> http://www.w3.org/2001/sw/hcls/notes/kb/

<BillBug> scott: everyone review current doc?

<mscottm> lost connection

-> http://www.w3.org/2001/sw/hcls/notes/kb/#empty-protein-set Empty Protein Set issue

<BillBug> eric: forward closure - extra triples that we'd normally use OWL to infer

<mscottm> forward chaining or transitive closure, I believe

<alanr> unbound query?

<BillBug> eric: AlanR has put them in named graphs - when including name graphs, you get the "inferred" classes

<BillBug> eric: I can write up a description of why the query would be unbound - Alan thinks there are higher priority tasks.

<BillBug> AlanR: There is a way to be more precise - the problem could be better addressed in a triple store supporting inferencing. Not a trivial task.

<BillBug> AlanR: Some of the bnodes are essentially implementing OWL syntax as opposed to being real "things".

<BillBug> susie: you said problem is unbounded query, but if inferencing done outside of KB, then inferred classes added to KB, then it becomes bounded?

<alanr> it wouldn't really be a problem, even if the query was in owl. But there are different ways of asking questions in OWL.

<BillBug> ericP: yes - wondering if we could tweak query a bit so the same query WOULD work in an OWL triple store. Think can be done by adding one more triple to the query

<BillBug> AlanR: That's where I had issues

<BillBug> ericP: Alan, you seemed to be looking to use a query language other than SPARQL (Alan: yes). I'm trying to work out a way to do this in SPARQL

<BillBug> AlanR: Since SPARQL doesn't directly support inferencing, any proposal would be purely speculative. Similar problem in OWL, since there is not now a specified OWL query language?

<BillBug> AlanR: There are OWL query options BUT none are a part of an excepted W3C spec.

<BillBug> EricP: So there are essentially design patterns that have arisen to help make such inference-dependent queries bounded in SPARQL

<alanr> http://clarkparsia.com/files/pdf/sparqldl.pdf

<BillBug> EricP: Why would this not work in Pellet?

<BillBug> AlanR: Not clear it won't work, just more complicated than Pellet can contend with now. May work in the Jena model (AQL) where there might be support for this sort of query. The general problem is NOT ALL QUERIES ARE DECIDABLE - i.e., there are queries you can express in SPARQL that are not decideable.

<BillBug> EricP: Agreed. However, there are ways of defining a virtual graph that can make unbounded queries work.

<BillBug> AlanR: True but the range of virtual graphs expressible in OWL would certainly allow you to express undecideable queries

<BillBug> EricP: True - but you can still define a virtual graph in SPARQL that would make this particular query bounded. I was just proposing a way of creating extra triples that essentially exist as a Rule that sits outside the query and helps to make it decideable.

<BillBug> EricP: Still having a hard time documenting the current query that explains how we've created a intersection that could be THEORETICALLY EMPTY, but is not so because of the way we've populated our KB triple store.

<mscottm> my connection has been dropped again (!)

<BillBug> AlanR: This is where I don't quite get it.

<alanr> Current example returns pairs of classes and instances

<mscottm> and again..

<alanr> SPARQL effectively returns only instances (or classes viewed as instances)

<BillBug> Kei: Current example is you are using a GO process. If we try another example, would we still be able to find intersection sets?

<alanr> OWL considers this kind of query to be complicated

<alanr> because of the mix of return values

<BillBug> Kei: Also - are there other ways of connecting SenseLab data to other data sets?

<BillBug> Susie: Make sense for AlanR & EricP to have another call to resolve this?

<BillBug> EricP: Yes - we have a call coming up later today

<alanr> I will be in at 1, not 12, as discussed previously

<BillBug> Susie: I beleive we decided we'd write up things up as they are, and also separately descirbe enhancements

<BillBug> EricP: I just felt there was an outstanding issue in describing the current state of the KB - and wanted to make certain we truly understood the issue.

<BillBug> AlanR: A key aspect to clearly specifying which BNODES represent OWL syntax. (EricP: agreed)

Susie: would like to get to WD and solicit feedback

<alanr> In fact, these bnodes could be replaces by constants without changing the OWL-DL meaning (but the OWL-Full meaning would change)

ericP: happy to publish as it is

<alanr> I don't have a lot of time until the end of the week

<BillBug> Susie: Eric - do you have a date in mind for soliciting feedback one more time?

<BillBug> EricP: I'm happy to publish at any point, since the IG has agreed to do so. I'm happy to go as is

<BillBug> Susie: What about asking for additional review by end of this week.

<BillBug> EricP: Rather not wait too much - if we push it off 'til next Friday, that's fine - or get others to review in next few days for publishing by this Friday?

<alanr> need to get names

<alanr> and assign actions

<BillBug> Susie: Think it would be good to aim for the 18th with folks passing their comments to Scott

reviewers: Susie, BillBug

<alanr> 24 hours from now

<BillBug> EricP: what would your time frame be, Scott

<BillBug> Scott: 24 hrs

<BillBug> All: OK

Susie: how goes the SenseLab note?

<BillBug> Eric - can I pass the minutes baton back to you?

matthiassamwald: quite close to completion IMO
... need to reformat citations and change pictures

<matthiassamwald> http://www.w3.org/2001/sw/hcls/notes/senselab/

Susie: can be done by when?

<BillBug> For those of us who have been out of the loop, could all document URLs be listed on the IRC please

matthiassamwald: during the next few days

Susie: ready for review by next Mon's BioRDF call?

matthiassamwald: reviewable now

ericP: would like them reviewed and published simultaneously

Susie: do you need a pub req?

<scribe> ACTION: ericP to send 1st WD transition req to webreq [recorded in http://www.w3.org/2008/01/07-BioRDF-minutes.html#action01]

Topic DSE Note

Susie: haven't read the DSE note yet

<BillBug> Can a drug be efficacious but not safe?

Susie: got write-lock from EricN on Fri
... alanr, did you want write lock?

alanr: yes, but not today

Susie: is there a word version?

alanr: Rachel had one, but don't know the normative one

<mscottm> Here's one that I know of: http://www.w3.org/2001/sw/hcls/Note_DSE_20071108.html

alanr: this is working in GoogleDocs

Susie: works for me

<alanr> Your document is publicly viewable at: http://docs.google.com/Doc?id=dthvctw_5fwwwprc3

-> http://docs.google.com/Doc?id=dthvctw_5fwwwprc3 DSE Note, editor's draft

Susie: were Rachel's edits obvious to you?

<scribe> ACTION: alanr to ask Rachel to edit the google doc, and barring that, will integrate Rachel's comments [recorded in http://www.w3.org/2008/01/07-BioRDF-minutes.html#action02]

URI Note

alanr: gave the note a rest for this month
... discussing issues informally
... will take up editing 1 Feb

Demo (Alan and Kavitha and Holger)

holger: not much update due to xmass
... would like to demo Semantic Web Search Engine and Semantic Web Pipes
... matthiassamwald has been working on these

matthiassamwald: new release in the next two weeks

holger: will jump back into demo work tomorrow
... expect proper update next week

Susie: proposal submission deadline?

alanr: due 11Jan

<alanr> http://www.eswc2008.org/calldemos.html

Susie: you were going to pull out some old abstracts (Society for Neuroscience) and re'use

holger: are they publicly available?

<BillBug> Matthias had the ISMB abstract from last summer, I believe

<matthiassamwald> Holger, here is the link to the ISMB poster: http://neuroscientific.net/res/ismb2007/poster.svg

<scribe> ACTION: Don to send the abstract to alanr and matthiassamwald [recorded in http://www.w3.org/2008/01/07-BioRDF-minutes.html#action03]

Misc

<BillBug> Alan - there is also the PPT for the biocurator presentation we gave - which has at least a few slides that are probably relevant. I believe you have a copy of this, though it is available here on line - http://genome-www.stanford.edu/biocurator/IBCM2007/abshtml/abstract-table.html

Susie: Scott will be presenting at infotech forum in March

<alanr> bye

<alanr> take care...

<BillBug> bye

- DRAFT -

BioRDF

7 Jan 2008

Attendees

Contents

KB documentation

URI Note

Demo (Alan and Kavitha and Holger)

Misc

Summary of Action Items

Scribe.perl diagnostic output