[GitHub] [systemml] Baunsgaard commented on pull request #872: [SYSTEMDS-273] Refactor Compressed Package

[GitBox <gi...@apache.org>]

Baunsgaard commented on pull request #872:
URL: https://github.com/apache/systemml/pull/872#issuecomment-619542266


   > LGTM - overall this is great @Baunsgaard. During the merge, I just fixed a few typos, and minor warnings (generics, static methods).
   
   Thanks @mboehm7  for the fixes :+1:, our editors detect different things, so I think it is a fine trade-off unless it required significant changes :smile:.
   
   > Additionally, you might want to consider moving the `ColGroupSizes` into the class hierarchy of column groups to avoid string-based dispatching (including abstract classes that could never be instantiated) and bring this closer to their implementation. 
   
   Yes I could move it back into the specific ColGroup classes, but the important thing is that they are static methods, not requiring you to construct an object of the ColGroup to calculate the sizes. Therefore I had some difficulties because static methods if can't be overwritten in java. Therefore I decided to have the logic separated into another class, instead of having "Semi Random" names for getting size of objects inside the individual classes. I hope you have a better suggestion, because I two do not like the ColGroupSizes class as it is now.
   
   >Also, please reconsider the unnecessary indirection of `AbstractCompressedMatrixBlock` and `CompressedMatrixBlockFactory`. Is there any intention behind them other than splitting `CompressedMatrixBlock` into smaller files?
   
   The reason was that I wanted to make an LossyMatrixBlock object, that also pointed to the abstract CompressedMatrixBlock, such that we had the option to select compression also based on Object type. I did this before our discussion on keeping everything colocated inside CompressedMatrixBlock,  but while adding the lossy compression logic I will reconsider what could be smartest. That said, i 100% agree that the separation right now is unnecessary.
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org