-
Notifications
You must be signed in to change notification settings - Fork 105
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Rename solvers package and move functionals #944
Comments
The rationale is what @adler-j wrote. I have no strong opinion in the question, but moving sounds like a good idea :-) |
Part of the rationale was to not clutter the top-level namespace more than necessary. Let me give you some details about this. The overall structure of ODL currently looks like this:
Everything with a However, one namespace can only hold a certain amount of names before it gets hopelessly cluttered. For that reason, all the subpackages have their own namespaces that are not added to the top level. For example, you need to type Now, in the case of functionals, there are so many that they would lead to a significant amount of clutter if we would simply dump them to top level. Therefore I'm not in favor of making them part of the core in that sense. What we could alternatively think about would be a rename of |
Any new insights here? I just want to mention that there's a related unsolved (although closed) issue #637, namely where to put concrete operators that depend on substantial parts of the library. For example operators that use the Fourier transform (not mentioning names here) which itself depends on a fair amount of stuff. So one way of doing these two things together would be to add a subpackage like What do we think about that? |
I need to think more about this, but I feel that the two most promising options would be
I am personally against having "arbitrary" operators mixed up with the functionals in the same library since this would (at least to me) be somewhat confusing. |
Do we have any further discussion on this? I agree that this naming is perhaps bad and we should think about it before 1.0 |
I'm all for the change to Regarding placement: I think an intermediate solution would be to keep the structure as-is and just do |
I like the answer of @kohr-h. Then the code is still structured well and can be accessed easily in an intuitive manner. |
Why are functionals part of solvers, e.g. odl.solvers.L1Norm? To me functionals should probably have their own category (or at least some non-contradicting category).
The text was updated successfully, but these errors were encountered: