Re: HTML5 Test Development Estimation

On 2013-08-06 18:46, Michael Dyck wrote:
> On 13-08-06 09:15 AM, James Graham wrote:
>> 
>> test writing is creative, fulfilling, work, but adding metadata is 
>> not,
> 
> Surely what's fulfilling depends on the person -- there may be people
> who would prefer adding metadata over writing tests. Will the testing
> effort encourage & allow such people to add value to the test suite?

Well certainly I don't want to discourage people looking to contribute 
in that way; as I previously said the metadata can be very valuable. Of 
course that value comes with a cost e.g. with a poorly designed scheme a 
small change to the spec may require a disproportionately large number 
of updates to the metadata in the testsuite (I believe that the system 
being used for XHR tests suffers from this, for example, since XPaths 
are generally fragile in the face of DOM changes). Also it might be 
quite hard to do post-hoc annotations of tests that someone else wrote. 
Nevertheless it would be quite reasonable to experiment with this kind 
of division of labour.

Received on Wednesday, 7 August 2013 04:14:11 UTC