On Fri, Mar 30, 2012 at 3:07 AM, Sergiu Ivanov
<[email protected]> wrote:
> On Fri, Mar 30, 2012 at 4:59 AM, Ronan Lamy <[email protected]> wrote:
>>
>> This looks like it'll be a solid proposal. In particular, you do a good
>> job of summarising the feasibility of your project, its evaluation
>> criteria and the benefits it would bring. However, some points
>> (particularly the section "Infrastructure setup") could be made more
>> concrete if you wrote some doctest-ish pseudocode. Bonus points if the
>> code actually works.
>
> Thank you for your feedback!
>
> I will throw down some pseudo-code in a couple of days.
>
>>> As a way to better structure the code, I propose a (partial)
>> transition to object-oriented design.
>>
>> Yay! +10.
>
> Cool :-)
>
>>> I suggest defining a very basic class Operation which will initially
>> contain one method with two arguments, which will perform the operation
>> itself.
>>
>> I suspect that having to use the object in order to perform the
>> operation would be cumbersome, so it might be more fruitful to consider
>> Operation as a description of the operation, instead of as the operation
>> itself.
>
> This is exactly the point of Operation: it will eventually store the
> information about the operation *and* will be able to actually perform
> the operation on two elements.  The problem is that implementing a
> full fledged Operation is beyond the scope of my project, so I am only
> going to set up a very basic form of it.  Eventually, when someone
> (maybe myself even, but later) gets to work on base classes for
> abstract algebraic structures, it will be possible to easily migrate
> the existing Groebner code base to the proper structures.
>
>>> I will start by defining the function nextPoint()
>>
>> "nextPoint" isn't PEP8-compliant. It should be "next_point".
>
> Oh, sorry, habits :-( Fixed.
>
>>> July 9 - July 15: start merging the changes upstream; start working on
>> the documentation;
>>
>> That's rather late. Any pull request with more than ~10 commits is
>> painful to review. A pull request at that stage is likely to contain a
>> few dozen commits, so you'd be lucky if it got merged before the end of
>> August. You should endeavour to send pull requests as soon as you have a
>> consistent and well-tested piece of functionality.
>
> I actually had this thought while writing the proposal, but wasn't
> sure about it.  For me, pushing things upstream in small chunks feels
> much more comfortable; it's good news that the developers are also
> comfortable with strategy.
>
> Fixed the proposal by removing explicit statements about merging in
> the timeline and adding a proposition about pushing things in small
> chunks after the timeline.
>
> Sergiu


I've written another mail to this list about this.  We need to really
focus on merging things early and often this summer.

Aaron Meurer

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/sympy?hl=en.

Reply via email to