SBML.org — the global portal for all things SBML

Level 3 Package Proposal Voting Results: Qual, FBC, Annot and Dyn

The call for votes for the following four packages was issued on May 22 and closed on June 6, 2011:

A total of 31 votes were cast. Removing 3 votes from people who are not actually members of the sbml-discuss mailing list (as this is a requirement in the process), and another vote cast by someone who later submitted a revised vote, left a total of 27 valid votes. The results and comments are presented in separate sections below. Note that not all individuals voted on each proposal, which leads to totals of less than 27 when the votes are summed for each proposal.

Qualitative models proposal

The outcome of this vote is accept because more than 50% of the votes cast were cast for 'accept'. The following graph presents the results:

The following are the comments made by people who voted. The votes in particular are indicated in parentheses before the comments.

  • (Voted 'accept') I almost said revise here as I think that this proposal can be simplified e.g. is new "qualitativeSpecies" really required, or can one extend an L3 core species?. However, I think there is a need for such a proposal.
  • (Voted 'accept') Voted accept though not entirely convinced this package is necessary. My understanding though is that it is far along and will continue regardless.
  • (Voted 'reject') I personally think that the encoding of qualitative models should be left to other efforts, eg. PNML (or BioPNML). So, I'd favor to better communicate with the related efforts and figure out how to combine models from these different languages to SBML models instead of having a new extension. However, I am not an expert with the qual extension, so maybe people from other efforts are participating here and this is more or less to move existing efforts under the umbrella of SBML... So, this reject doesn't reflect any strong opinion (mainly due to the fact that I don't know too much about the package extension).

Flux Balance Constraints proposal

The outcome of this vote is accept because more than 50% of the votes cast were cast for 'accept'. The following graph presents the results:

The following are the comments made by people who voted. The votes in particular are indicated in parentheses before the comments.

  • (Voted 'reject') FBA feels more like something that software does with a model than a description of a new kind of model. Is there something fundamentally different between FBA and any other parameter optimization scheme, for instance fitting time course data with a Levenberg-Marquardt algorithm? SBML does not, for instance, support upper and lower bounds on adjustable model parameters or any of the many objective functions that have been devised to support least squares fitting.

Annotations proposal

The outcome of this vote is accept because more than 50% of the votes cast were cast for 'accept'. The following graph presents the results:

The following are the comments made by people who voted. The votes in particular are indicated in parentheses before the comments.

  • (Voted 'accept') I think you guys shouldn't necessarily be wedded to RDF, if it turns out that we're taking it in directions it was not meant to go. If other approaches more naturally match the data you want to encode, feel free to switch--I don't think the effort you've put in so far to RDF will be wasted at all; it has served to clarify many many issues.
  • (Voted 'revise') The package is under development. Contrary to quali for instance, the community did not reach an agreement yet. In particular, a fundamental question remains the use of pure RDF (essentially an extension of the core) or OWL2.

Dynamic Structures proposal

The outcome of this vote is accept because more than 50% of the votes cast were cast for 'accept'. Revised 2011-06-19: the vote on the dynamics proposal has been rescinded; a proper proposal awaits development and will be put to vote in the future.

The following graph presents the results from the original vote:

The following are the comments made by people who voted. The votes in particular are indicated in parentheses before the comments.

  • (Voted 'revise') Can't comment as there is no detail on the approach and lot of whether this should exist as an independent package (or be part of e.g. a multicompartment/multicellular description) will depend on that.
  • (Voted 'revise') I am wondering whether this is not rather part of the experimentation with a model (a simulation step) and therefore might better be covered by SED-ML. I could not find a more detailed package proposal so I am not sure I understood correctly how this extension will look like and if SED-ML could be imagined to cover the needs. If I could I'd vote for "unsure, want to know more" ;-)
  • (Voted 'revise') At this stage it is difficult to say whether the approach is reasonable, I agree that there is a need for such a package.
  • (Voted 'revise') I reconsidered my vote here--I think there's just too little information to go on to give a proper vote to the approach. The need is clear, but the only listed approach is 'maybe we'll use this other method', which begs the question: what exactly is the approach being proposed and voted on?
  • (Voted 'revise') Sounds good, but I did not find much of a detailed proposal apart from the abstract. As this package would depend on the set/array packages, shouldn't there be votes on these first?
  • (Voted 'revise') There is no package. Therefore the option ""accept"" is void of meaning. But we certainly need a mechanism to build dynamic structures. IMHO, it should be part of the core.
  • (Voted 'revise') There is no visible proposal at http://sbml.org/Community/Wiki/SBML_Level_3_Proposals/Dynamic_Structures
  • (Voted 'revise') This is not sufficiently formulated as a "proposal", so revise is the only logical option, IMHO. The idea of having such a package is a good one, however.


Please use our issue tracking system for any questions or suggestions about this website. This page was last modified 16:34, 19 June 2011.