[Grok-dev] Re: grok.layer branch
kevin at mcweekly.com
Tue Apr 17 14:31:23 EDT 2007
Philipp von Weitershausen wrote:
> On 17 Apr 2007, at 19:59 , Kevin Smith wrote:
>> I'd like to hear some more grokkers weigh in on grok.layer vs.
>> grok.request, I'd like to get this out the door today if possible. :)
> Just chatted with Martijn. We came to a conclusion that grok.layer()
> would probably yield the best code readability which weights higher
> than the actual symmetry. So, let's go with grok.layer then.
> Thanks for doing this. Let me know when you think your branch is
> ready, I'll be happyt to review it.
Great thanks, I'll get right on it.. I appreciate your help and guidance.
>> I prefer grok.layer because IMO it has more narrowly implied
>> definition whereas grok.request has a wider implied definition ( I
>> expect it to do more than just assign layers, if even that. )
> Well, it's not about *assigning* layers or anything. It's about
> expressing a constraint. So, we should at least document it clearly
> that grok.layer() ends up being a constraint on the 'request' of
I'll make a note in the interfaces and doctests.
More information about the Grok-dev