Minutes of the Bioinformatics Standards Committee Working Group
Nov. 12, 2004, 12:30 PM

Attending:

Vicky Markstein (temp chair)

Betty Cheng (recorder)

Robert Stevens (Manchester UK)

Dan-Wen Tsai (biology background – enrolled Foothill College)

Joanna Luciano (Harvard Medical School, Biopax)

Alex Morgan (Mitre)

Sean Zhou

Bob Davis (chair microprocessor Standards Committee)

Dan Zuros (Advisor, IEEE)

Peter Markstein

Next meeting:

Nov. 23rd at 5:30 PM PST or 8:30 PM EST

Summary of Discussion and Action Items

Schedule these meetings at times that are better for European participants. Develop further community contacts at the organizational level (see details in discussion).

The BSC’s approach to Standards is critical; must be inclusive. The BSC should use Open Source as for all standards developed and released by the BSC. Criteria for standards development should include:

A) Tool development, e.g., validation suites
B) Attempt to incorporate existing standards, de facto and legacy standards (backwards compatibility)
C) Conversion tools
D) Needs to address needs of business (businesses dislike open standards)
E) Involve shareholders to assure standards meet the needs of the users
F) Support for released standards

Set up the mailing lists include rational for the existence of this group.
Send out the minutes and place the minutes on the webpages
Contact Cherry Tom about developing the IEEE’s role to include help for Standards
proposals preparation.

Discussion

The BSC needs more community involvement – we’ve been contacting individuals. The BSC should approach organizations.

A press release is due out shortly; this should improve our community contacts.

Ask Doug Brutlag (Prof. Biochem Stanford) about contacts at NCBI (action item – Betty)
Ask Peter Covitz about contacts at the National Cancer Institute (action item Betty)
Ask Graham Cameron, Director of Services is the appropriate contact at the EBI (suggestion by Robert action item – Betty?)

[Bob Davis] The IEEE has three (3) levels or types of Standards

A) Basic Standards – using an example from floating point this is the “specification”
B) Guides – how to implement the standard
C) Recommended Practices

The IEEE review process takes the standards development group to each of these three levels.

Should we go further and support Standards with workshops to teach users? This is beyond the current effort of the IEEE but could be a mechanism to raise funds to support standards development.

Some bioinformatics groups have already gone through a similar process. For example the Object Management Group (OMG) has similar requirements.

These requirements may be too restrictive. We want to avoid a mechanism that smacks of authoritarianism. The community will simply avoid us.

We must choose project carefully then to first gain community trust.

Is Open Standards really a working solution? As an example, Deviant (a style of Linux kernel) is the best but doesn’t really adhere to the standard. Standards make it too difficult to come up with the best solution – too difficult – for example Genbank… [at this point Robert Stevens clears up a misunderstanding].We need to standardize the data in Genbank. Many perl scripts access the information via the webpages for example and when the format changes, the scripts crash. There is a need to create more XML markup for bioinformatics data. Let’s upgrade this to include Ontology Web Language (OWL) & Resource Description Framework (RDF). Agreement: this is a do-able project. Earlier objection based on misunderstanding.

The NCBI has a service provision – beyond the scope of their mission to do all this…

Agreement:
Existing, defacto, legacy standards must be supported

There should be political/social committment to Open Source or Open Standards

The validation process must be free, e.g. the validation software– open source or open standards is meaningless w/o free validation. Open source validation software. Software that supports standards should be open source.

Those who make modifications to the validation software (and hence the standard) must give those modifications back to the community (effectively those who make modifications must join the standards development community – comment added by BC)

We should be careful about the distinction between an IT standard, e.g., XML, vs. a Bioinformatics standard, e.g., the Gene Ontology (GO).

We need a standard for microarray data.

Action Item: the reflector for the bioinformatics standards committee study group is set up

Stds-1953@ieee.org (the name of the mailing list – use this to post announcements)

Use listserv@ieee.org and “subscribe stds-1953” to register for the mailing list

There is also a website which can be used to sign up for the mailing list

The website for the BSC is http://www.csbcon.org, follow the link to the BSC

Comment : commercial entities, corporations have resources and their own motivations to push standards – could be a source of resistance

We need proposals to write Standards. (agreement)

IEEE should lead this effort – perhaps a role for the BSC – Action item – consult Cherry Tom

Action item: Recorder to send out these notes
Time changes: best times to schedule teleconferences for Europe is 1-5PM

Metaphor for evolution is particularly apt – death and change
Standards advice – this is a thankless job, the most reward you will ever get… no one will notice.

Action item – set up the reflector to indicate rational for the 1953 BSC

Action item – set up a new time for these meetings.