[DAS2] DAS/2 weekly meeting notes for 9 Jan 2006

Chervitz, Steve Steve_Chervitz at affymetrix.com
Mon Jan 9 19:27:53 UTC 2006

Notes from the weekly DAS/2 teleconference, 9 Jan 2006.

$Id: das2-teleconf-2006-01-09.txt,v 1.1 2006/01/09 19:32:33 sac Exp $

Note taker: Steve Chervitz

  Affy: Steve Chervitz, Ed E., Gregg Helt
  Sanger: Andreas Prlic
  Sweden: Andrew Dalke
  UC Berkeley: Suzi Lewis
Action items are flagged with '[A]'.

These notes are checked into the biodas.org CVS repository at
das/das2/notes/2005. Instructions on how to access this
repository are at http://biodas.org

The note taker aims for completeness and accuracy, but these goals are
not always achievable, given the desire to get the notes out with a
rapid turnaround. So don't consider these notes as complete minutes
from the meeting, but rather abbreviated, summarized versions of what
was discussed. There may be errors of commission and omission.
Participants are welcome to post comments and/or corrections to these
as they see fit. 

Today's topic: Planning for Code Sprint

Agenda: Focus on the upcoming DAS/2 code sprint in early February in
the the UK and whatever we need to do to stabilize the read part of
the spec by then. 

GH: Andrew proposed in the UK early feb, and in the US later March.
Getting people from the US to the UK can't do. There's no funding for
it. Can teleconf into the UK from US. Maybe get together Bay area
people (suzi, nomi, gregg, ed, steve). Would also be good to get Allen
Day involved. Terminology note: 'code sprint' is preferred to 'hackathon'.

AD: The dates for the UK code sprint are 7-11 of Feb. Have already
schedule flights. In the US, march sometime but it has not been

GH: Will UK people be able to get their schedules cleared to dedicate
a solid block of time? I can do so here in the US.

AD: Planning to spend the whole time on the code sprint.

AP: would be able to attend, planning to write an announcement,
registration mechanism, reserve a room, and start coding...

[A] Andreas will post announcement for Feb code sprint, set up registration
form, reserve room

SL: I will notify Nomi about it today. When is the progress report for
the grant review due?

GH: Not sure. Sent one in may/june last year.

SL: Important to send report before review committee sits. Would be
good to report on the code sprint for the review.

GH: Received notice at end of last year that renewal has been
received. Initial peer review will be done by end of march
2006. funding decision after May 2006. So we could for sure get in a
write up on the UK, and possibly the US one to. Assigned group:
"center for scientific review, special emphasis panel" (not the same
group as last time, unfortunately).

[A] Gregg will prepare reports of code sprint results before grant review

GH: What do we want to accomplish when we have one week to do it?

SL: DAS/2 adaptor for apollo.

  - Make IGB code for DAS/2 client, update to reflect changes in spec
    implement unimplemented stuff.
  - DAS/2 server at Affy - bring it up to the spec as well. not a full
    impl but was compliant 3 mos ago.

AD: Three things:
  - make sure spec works for people providing DAS/2 gets.
  - write up the writeback stuff.
                         - work on validator.

  - brainstorm test cases for validator.
  - testing compliance as well as performance
  - more spec work possibly?

  - Would like to work on the registry.
  - Serving real data with DAS/2.
  - Thomas is also planning to be involved.

AD: been updating the spec to include all the things discussed in
the past month. Want to get it updated at least a week before the code
sprint. Slow at text writing since prefers coding. It's really boring
to do this. Easy to get distracted.

GH: Would also like to get Allen Day to sign on for updating the
chado-based DAS/2 server, bringing it on with latest changes. Possibly
get him to come up with the others in the west-coast. Haven't
contacted him yet.

SL: Can't do it (in Philly). Mark can't either, but likely can be
involved in the March sprint.

GH: March sprint, anytime in the middle 3 weeks of march.

AD: will be at a san diego workshop in march. Back in the states
in Mid feb. 

SC: There is a MAGE v2 workshop/jamboree in Benecia, CA on 7-11
March. Should probably aim for later March. Also will be involved in a
netaffx quarterly update.

GH: We could spend time on Netaffx-related stuff during the DAS/2 code
sprint in March. E.g., pipeling the data prep for the netaffx DAS/2

GH: Other two groups to ping for involvement:
  - NCI stuff that Brian Gilman is doing (see his post to
    list - ready to start going).
  - UCSC - DAS/2 annot server. they're not budgetted until the renewal,
    but could get started. Who from UCSC? will have to look at
    notes. someone whose worked on the DAS/1 server.

GH: Any more ideas for the format of the code sprint?

AD: wait until we get closer to the date.

SL: Need clearly defined goals. too many things could be going on.
Put this together closer to the date.
SC: Yes, would be good to have defined deliverables. Some way of
measuring success.

[A] Someone will write up goals for the Feb code sprint.

AP: Want to write an official invitation to the list, invite people to
register and say what they are planning to work on, and would like to
work on.

AD: Planning to put more spec additions. Searching for features. query
now is set up via params on url. wants the props to have key val = key
is the xml element name and the value is the value. But how to get
this to work with how we do queries...How you'd put it into the query url --
element name is the fully qualified thing. How would this interact
with the search for properties in spec now, the name of property is a
short 6-char thing, not a long url thing.

also some props are numbers, is you search for 1 does it match 1.0?

GH: the property id is now a long URI. this is the problem.

AD: emailed an atom spec developer to ask. said: it's a sop (?) to the
rdf semantic web community. People who wanted rdf wanted to use that
field. So maybe just prop name=.. value=... where name is anything.

SL: can we use 'tag' or 'key' instead of name?

AD: was reading an article/news group: names and cache coherency are
the only really hard things in comp programming.

GH: So now for an arb prop in a feat, it would be
  <property key=somestring value=somestring />
AD: query spec now uses colon (attr:)
GH: feat filters att=property:value  - implies that you can't have a
colon in the property

AD: likes idea of colons for namespaces, but dots could work.

Also, some fields are strings, ints, float, etc., and searches mean
different things for different property keys. Better to have partial
solution and not strive for full genericity.

GH: hasn't been impl in client or server yet (property types).

AD: corba/ soap allow clients to ask servers what they support. rarely
see this actually being used. So best not to put too much work into

SC: Would key make use of URI base? I.e., be implicitly a URI?
AD: no just a string e.g., 'bgcolor'.

GH: like what we do now in types where we have a uri as an id and you
can short hand the uri

AD: if we want one we can stick one in the properties table
GH: can use it if we need to later.

GH: So we're set then for the 6-11 feb .

More information about the DAS2 mailing list