[Insight-developers] itkResampleImageFilter.h & itkWarpImageFilter.h and Software Archeology

[Bill Lorensen]

I think we just need to be more diligent. Each morning, I usually look
at the files changed from the previous day. If at least one other
person did this, we could catch problems. Also, we should discuss
changes on the mailing list as we are doing now regarding
WarpImageFilter changes.

And, more tests are always helpful. Although the last time I broght
this up, we got sidetracked about a better testing harness.


Bill

On Wed, Mar 25, 2009 at 1:38 PM, Luis Ibanez <luis.ibanez at kitware.com> wrote:
> Gaėtan Lehmann wrote:
>>>
>>>  Being positive and looking to the Future:
>>>  ==========================================
>>>
>>>
>>>  It looks like we *MUST* implement some sort of
>>>  *systematic* code-review practice.
>>
>>
>> Yes, it would be great, but we may have a similar problem than with IJ
>>  contributions...
>>
> ------------------------
>
> Indeed,
> that's what I'm afraid of.
>
>
> For code reviews to be practical,
> they have to be set up in a very easy-to-use framework.
>
> I'm taking a look at review-board, but this is not the only
> system available out there...
>
> An alternative option is to start using "git" for code control
> and exercising the difference between committing code and
> publishing code.
>
> The burden for ITK is that we are trying to do Agile Programming
> under a backward-compatibility constraint.  It really looks like
> we need an intermediate stage were new commits are cured before
> being integrated in a release. We do this for new classes with
> the Review directory, but we don't have a mechanism form simple
> day-to-day patches.
>
>
>
>    Luis
>
>
>