[Ontology-editors] freeze for regulates links

Thu Feb 19 03:19:04 PST 2009

In light of Chris' concern for avoiding format problems -- as well as the 
fact that only one line differs between OE1 and OE2 output at present -- 
it seems to me that this instance of breaking the OE1-commit rule, though 
unfortunate, cannot be construed as callous disregard for the spirit of 
the law or the integrity of the ontology files. Furthermore, I think 
there's nothing to be gained by continuing to dwell on the incident; it's 
much more important to get new procedures in place to support more 
automated checking. We can reasonably expect automated checks to avoid 
errors more reliably than the current save-in-OE1 routine.

I would also like to follow up on Tanya's comment, and make sure it's 
clear that the MF-MF and BP-MF regulates relationships have not been 
committed prematurely. As the ontology development reports indicate, these 
new relationships are the product of several months' work that included a 
hefty dose of quality control. Chris' clearly articulated reasons for 
using the OE2 core at once do not mention, and are not directly connected 
with, the addition of these relationships.

m

On Wed, 18 Feb 2009, Karen Christie wrote:

> My comment was only made in the context of Chris's argument that we -need- to 
> change the rule to allow use of OE2 for committing. If it can all be done in 
> OE1, then perhaps we don't NEED to change the rule.
>
> -Karen
>
> On Wed, 18 Feb 2009, Tanya Berardini wrote:
>
>> Only commenting on this part:
>> 
>>> If these relationships can only be done in OE2, which is not yet fully
>> tested to demonstrate stability, >reliability, etc., then perhaps we 
>> weren't
>> really ready to proceed.
>> 
>> The 'regulates' relationships, both inter- and intra-ontology, can be and
>> have been added, displayed and edited in OE1.
>> 
>> Tanya
>> 
>> On Wed, Feb 18, 2009 at 2:48 PM, Karen Christie
>> <kchris at genome.stanford.edu>wrote:
>> 
>>> Hi Chris,
>>> 
>>> If rules need to be changed, and sometimes they do, it seems that this
>>> should have been brought up before you chose to disregard the rule without
>>> bringing it up for discussion.
>>> 
>>> If these relationships can only be done in OE2, which is not yet fully
>>> tested to demonstrate stability, reliability, etc., then perhaps we 
>>> weren't
>>> really ready to proceed.
>>> 
>>> -Karen
>>> 
>>> 
>>> 
>>> On Wed, 18 Feb 2009, Chris Mungall wrote:
>>> 
>>> 
>>>> Thanks Midori!
>>>> 
>>>> Sorry about ignoring the rule. It seemed arbitrary in this case as the
>>>> diff was identical apart from the saved-by tag.
>>>> 
>>>> We need to change this rule anyway:
>>>> 
>>>> - OBO-Edit is constructed from two software components: OBO-Core and the
>>>> UI
>>>> - OBO-Core in OE2 is production, not beta
>>>> - There are precisely zero people in the GO Consortium who are familiar
>>>> with OBO-Core in OE1
>>>> - We have been having mysterious problems with obo2obo in OE1 on the
>>>> production side
>>>> - OBO-Core in OE1 had no automated test suite
>>>> - OBO-Core in OE2 has a comprehensive automated test suite
>>>> 
>>>> Unfortunately we can't switch to the more stable OBO-Core2 until the OE2
>>>> UI is acceptable for editors. However, in the interim I propose that we
>>>> remove the ban on OE2 committing. Doing a roundtrip will serve no useful
>>>> purpose anyway, as even if problems were to be introduced in OE1, the
>>>> chances of them being propagated in a roundtrip are extremely high. I am 
>>>> in
>>>> fact extremely paranoid and vigilant about errors in either format syntax 
>>>> or
>>>> content leaking into the public version of GO, but I think this rule
>>>> provides only a false sense of security.
>>>> 
>>>> I do thing we should be far more rigorous in automated checking of
>>>> gene_ontology_write.obo. We are now in a position to do this as we have a
>>>> separate filtering step prior to propagation to the public. I seem to 
>>>> recall
>>>> we started collecting possible checks on the wiki, but I can't find that
>>>> right now.
>>>> 
>>>> If the editors are in agreement I will propose to Mike that the rule be
>>>> modified such that it applies to the core rather than the UI.
>>>> 
>>>> 
>>>> On Feb 18, 2009, at 9:47 AM, Midori Harris wrote:
>>>>
>>>>  I have just issued a command to commit my changes from today, saved in
>>>>> 1.101. So this time there's no need for Chris to do a separate commit.
>>>>> 
>>>>> m
>>>>> 
>>>>> On Wed, 18 Feb 2009, Karen Christie wrote:
>>>>>
>>>>>  It looks like you committed the file using beta55. Since Mike Cherry 
>>>>> has
>>>>>> explicitly stated that the GO file should always be produced using the
>>>>>> official release version of the software, and NEVER with a beta 
>>>>>> version,
>>>>>> could you please resave with OE-1.101
>>>>>> 
>>>>>> format-version: 1.2
>>>>>> date: 17:02:2009 18:07
>>>>>> auto-generated-by: OBO-Edit 2.000-beta55
>>>>>> 
>>>>>> -Karen
>>>>>> 
>>>>>> 
>>>>>> On Tue, 17 Feb 2009, Chris Mungall wrote:
>>>>>>
>>>>>>  unfreeze.
>>>>>>> as if by magic... we have inter-ontology and intra-MF regulates links!
>>>>>>> Enjoy.
>>>>>>> Remember, these will be invisible to most users for now, unless they
>>>>>>> explicitly choose to load from gene_ontology_ext.obo. See 
>>>>>>> ontology/README
>>>>>>> for details
>>>>>>> We should post announcements once everyone has had a chance to look
>>>>>>> On Feb 17, 2009, at 4:19 PM, Chris Mungall wrote:
>>>>>>> 
>>>>>>>> I'm freezing the ontology whilst I add the BP->MF and MF->MF 
>>>>>>>> regulates
>>>>>>>> links. Should be done in a couple of hours, I'll announce the thaw 
>>>>>>>> then.
>>>>>>>> _______________________________________________
>>>>>>>> Ontology-editors mailing list
>>>>>>>> Ontology-editors at geneontology.org
>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> Ontology-editors mailing list
>>>>>>> Ontology-editors at geneontology.org
>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>>>>>>> 
>>>>>> _______________________________________________
>>>>>> Ontology-editors mailing list
>>>>>> Ontology-editors at geneontology.org
>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>>>>>> 
>>>>>>
>>>>  _______________________________________________
>>> Ontology-editors mailing list
>>> Ontology-editors at geneontology.org
>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>>> 
>> 
>> 
>> 
>> -- 
>> Tanya Berardini
>> TAIR Curator
>> www.arabidopsis.org
>> 
> _______________________________________________
> Ontology-editors mailing list
> Ontology-editors at geneontology.org
> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>