Yotam Gingold
Yotam Gingold
I just mean that we may wish to change the semantics of our summation syntax entirely. Currently, it iterates based on the dimensions of where the summation index is used....
At first glance, I think that expression should be illegal because I couldn't tell you which `i` values would be summed. My guess is that it counts the number of...
I traced the history of our summation conditional expressions. It comes from a discussion around how to handle `L_ii = ∑_( i≠j ) -L_ij`. We started to really get into...
The wiki is inconsistent. I fixed it. I wanted to add the following code, but it doesn't compile: ``` a: ℤ index : an index value b: ℤ^4 index :...
Re-opening this issue until we add support.
I'm not sure about this. It sounds good in this case, but it requires the compiler to determine how the sequence is used. That is, it adds contextual rules to...
Implementing real sequences with vectors seems like an implementation decision. Is this what you mean? Or do you mean that the type should be equivalent? Should you be able to...
Right. The compiler can't statically check that sequences are the right size based on their type definitions. We previously discussed this and decided to allow it with a dynamic check...
Do you want a conditional vector constructor or a true sparse vector type? I could see a sparse vector constructor being useful. Is a sparse vector type? SciPy doesn't support...
A sparse column or row matrix would actually be dense if implemented as compressed row or column (CSR or CSC) or list of lists (LIL). You need something like triplet...