View source

Minutes of the Shared Names Steering Committee meeting on April 29, 2009

Chair: Jonathan Rees <br/>
Scribes: Trish Whetzel and Jonathan Rees

Present:
Carole Goble (CG), Jonathan Rees (JAR), Scott Marshall (SM), Alan
Ruttenberg (AR), Jeffrey Grethe (JG), Vladimir Mironov (VM),
Marc-Alexandre Nolin (MN), Trish Whetzel (TW)

Absent: Mark Wilkinson, David de Roure

Location: Stata Center, MIT

"Consensus" in the below means consensus of those present.

== Original agenda ==

# Are we serious?
# Procedural
# Expanding the steering committee
# Domain name
# Publicity and outreach
# Web site
# Technical
# Next steps; action items

== Are we serious? ==

JAR asked: Is this group really serious about the project - enough so that
we would be willing to legally incorporate, in the event that that
became necessary?  Consensus: yes.

JAR asked whether anyone could think of another organization that
would be more appropriate to take on this job.  After some discussion
no one could think of an obvious candidate.

== Procedural ==

The practice of assuming that silence after one week is assent was
rejected.  The chair was asked to collect definite agreement from all
steering committee members when consensus is required.

== What is the operational burden? ==

(sidetrack from original agenda, but pertinent)

CG: Free curation [of databanks, in this case] doesn't work.  If
you make the service central, then it will get investment.  But if we
don't have universal coverage, we won't get that adoption.

JAR: What exactly does 'universal coverage' mean?  [a propos staffing
need]  We should ask the shared-names mailing list: What is
*sufficient* coverage?

AR: Sufficient coverage is the union of the databanks in dbxrefs occurring in:
* GO
* NCBI
* EBI
* KEGG
* DBX [? can't find in google]

CG: For comparison, Biocatalog (of web services) has 4 staff members.

JAR asks what a requirement is.  Suggests the question: If the system
lacked X, would it still be worth doing?  If not, then X is a requirement.

Categorization of criteria:
* Killer  (without X we should disband)
* Phase 2 (good idea, but can put off until later)
* There are risks around it (if we don't do it, or do)

CG: We should do the minimum, but do it really well.

Alan: Best working similar system is MIRIAM.  Maybe we should poll
MIRIAM's users to find out what their requirements are.

AR: Review some issues: wiki http://sharedname.org/page/Issues
* Versions  e.g. RefSeq (named versions)
* Provider partitions (e.g. multiple species)
* Proof of Concept for tracking IDs - common response for 404

CG: We need a document that is like a grant proposal, outlining
benefits and risks of the project.

We could have some kind of poll? Determine issue ranking somehow?

Collect Use Cases  [???? JAR doesn't remember this]
* Harold Solbrig - Mayo Clinic
* Sergei - Bio2RDF
* NCBO - Open Biomedical Resources (needs to be confirmed)
* NCBO - Ontologies that currently do not have PURLs
* SM - Other HCLS uses cases

Would be nice to have a best practices document on how to develop URIs
[who said this?]

Benchmarks for measures of success [?]

Curation of Shared Names databanks / resources:
* Initial - generate resource URIs.
* Ongoing - updates when record no longer exist.
* Develop relationships with data providers so we find out when their identifier scheme changes.
* Process - are there existing curation processes in place, eg Integr8 (http://tinyurl.com/oe7uyk), that SN could partner with?
* Support - interest to have a major data provider on board

CG: ChemSpider would be good to have was there an issue with it? AR:
Is the data freely available?

== About expanding the steering committee ==

No consensus on criteria for accepting a new member.
Suggested criterion:  Can the individual provide direct help for adoption?

Consensus to invite Eric Prud'hommeaux to join the Steering committee
as non-voting liaison.

== Domain name ownership ==

Consensus: Every steering committee member should have the DNS
password.  Upon member leaving the committee, the password will be
changed.

ACTION on JAR: investigate how to transfer domain name ownership from
AR to the unincorporated association.

== Publicity and outreach ==

Consensus: Should be discreet and directed.

* Action Item - CG will mention Shared Names initiative at ConceptWeb Alliance meeting
* Action Item - SM and TW will contact people for Use Cases
* No other public announcements for the time being
* Action Item - CG will add information on Shared Names meeting to the 1ELIXIR project report regarding identity management
* Action Item - TW will continue dialog with NCBI on their identifier scheme and how identifiers can be tracked, and to obtain list of resources that are linked from NCBI

== Web Site Policy ==
Continue with Shared Names wiki
Protect SNC process pages from changes from non-SNC members

== Technical issues  ==

Consensus to continue using PURL system unless there is a limitation
that cannot be resolved

== Next steps ==

CG: What is the timetable? When will we make hard decisions?

Need responses from Dave Wood and OCLC regarding suffixes and virtual
hosting, ideally one month from 4/29/09 for feedback and resolution.
* Action Item : AR contact Dave Wood to check-in with OCLC about adding the OCLC server into the replication scheme

AR(?): Next step is to stress test the system by adding/deleting a few
million PURLs

* Action Item - JAR to work on management of A records, adding new records, Nagios monitoring
* Action Item - TW and AR to develop a more formal list of action items, see work from AR at http://sharedname.org/page/Work_packages
* Action item - JAR will write up issue of numeric vs. alphabetic databank names
* Action item - TW and AR will review database prefixes, then pass to SNC for review. Attempt to align with existing database prefixes or naming schemes, however this is not an absolute requirement.

Proposed timeline for release of Shared Names system:
* Interim usage/form of URI will contain at demo.sharedname.org as a silent alpha
* All review in two months (end of June) the form of the URIs [[Issue/Form of the URI]] and databank prefixes
* Aim for steering committee signoff on URIs at end of June
* July - Checkpoint meeting on usage of system ("how the sandbox is going")
* September - review, check in with users
* End of September - release of system under non-demo domain name
* Aim to co-ordinate with timing of International Semantic Web Conference (ISWC, http://iswc2009.semanticweb.org/), October 25-29, Washington DC

-----

Scribe's questions:
* what was the point about success benchmarks and who brought it up? would be nice to attribute it.
* would be nice to know who suggested which use cases.  I have no record of this at all.
* no idea what 'dbx' is

Jonathan