[Please note: I moved this onto the cellml-discussion list because the 
ABI team-cellml list is supposed to be for Auckland-local correspondence 
such as discussions about meeting times, which are not of interest to 
people outside Auckland].

David Nickerson wrote:
>>> I would also like to reiterate that the primary model repository on the 
>>> live site is not a suitable place for this kind of testing. If the test
>>> server is not suitable, can we at least create a separate model 
>>> repository (people can add one to their own home folders, right?) and 
>>> use that for testing and feedback on these features.
>>>       
>> Is Matt is working on this? There was some talk about Michael's
>> SiteTools at the last meeting being useful in this regard.
>>     
>
> Regardless of the state of a properly set up test server, as I 
> understand it there is nothing stopping anyone adding a model repository 
> to their own home folder and putting modes in there. Links can then be 
> sent out for discussion of the look and feel, etc, of these test models 
> which illustrate some new feature before they are added to the live 
> repository. While this is a less than ideal test situation it would be 
> much more preferable to asking for feedback on models already entered 
> into the live repository.
>   
I think that there are several quite different issues getting confused here:

1) Changes to the actual Plone product (including skin changes) should 
be tested elsewhere before they go up (to check that no existing 
functionality is broken and the change has the effect that it is 
intended to have). There is no option but to test this on a completely 
separate Zope instance because code changes could affect any objects in 
the Zope instance. I believe Tommy currently carries out this kind of 
testing locally in an ad hoc fashion, but site_tools will eventually 
automate the process.

2) It would also be useful to allow people to see the effects of code 
changes to decide if they agree before the live site is updated. This 
probably also requires a completely separate Zope instance. I believe 
that the IT staff have been working on a server that will help with 
items 1 and 2 on my list in conjunction with site_tools.

3) Submitting new models to the repository or making updates to these 
models is intended to be eventually become a fairly decentralised 
process, which certainly wouldn't require any sort of review except 
within the community the model relates to (provided the general curation 
rules about metadata and so on are followed). I don't think it makes 
sense at all to have every submitted model and every change to the 
models go into a 'test' repository, get approved via one of the CellML 
mailing lists, and then go into the live repository. Instead, changes 
should go into the live repository (with proper metadata describing the 
model status), and any changes after this should be tracked by the 
repository. If there are any different or interesting features in the 
model, the author of the model could draw them to the attention of one 
of the CellML lists.

Unfortunately, a lot of the model documentation structure which would 
ideally be generated by the Plone site currently resides in the 
temporary documentation in CellML models or in manually created pages 
(i.e. in the content, not the code), and so there is a blurring of the 
line between these cases. However, unless James is going to be changing 
something across all the models within a month or so, I think that case 
3 above probably still applies to such changes made through the model 
repository.

Best regards,
Andrew Miller

_______________________________________________
cellml-discussion mailing list
cellml-discussion@cellml.org
http://www.cellml.org/mailman/listinfo/cellml-discussion

Reply via email to