Normal form vs Normalized

[dsuess]

I am slowly getting angry with my past self for calling the cannoncial form as normal form (and, hence, introducing functions such as MPArray.normalize(...). This will be a big change in the API, but I think we should think about a new naming scheme.

And while we are making ourselves unhappy, I would also think about renaming bdims to ranks and pdims to dims. I think the distinction would help readability a little

[milan-hl]

I think replacing our homegrown names with more commonly used names is a good thing. I think we should keep the old names as aliases to the new names for some time. The aliases could generate warnings if used.

I am not so sure what names to choose, however.

• bdims -> ranks: There seems no doubt about that one. One the other hand, should we make it ttranks?

• pdims -> dims: Some authors [1] say an array with d indices has dimension d, so this might be ambigous. [2] writes "tensor of size n_1 times n_2 ... n_d". The same authors refer to one index of the tensor as a mode and [1] also calls the n_k "mode sizes". We could change pdims to mdims or msizes or mshapes (numpy-style). Strictly speaking, our "modes" maybe composed of multiple sub-modes but this works quite nicely because e.g. msizes[0] = (2, 2) would give a size tuple for the zeroth mode.

• The number of indices of a tensor is called "dimension" or "order". Numpy calls it ndim ("Number of array dimensions.")

• plegs -> submodes or mndim?

• At the moment, we also have dims and legs: dims is "Tuple of shapes for the local tensors" and legs is the number of entries in these tuples. We could call them coreshapes and corendim. (Or maybe use something more consistent with the other decisions we make.)

• Is the "dimension" or "order" of our tensor given by len(mpa) or by np.prod(mpa.plegs)?

• Do you also want to rename other methods such as bleg2pleg, pad_bdim, paxis_iter?

[1] http://dx.doi.org/10.1137/090752286
[2] https://arxiv.org/abs/1302.7121

[dsuess]

Currently at it on the refactor branch. Let me know if you have any suggestions.

[dsuess]

There is one more thing that I find confusing: outer vs louter. How about:

• outer -> kron (because it acts similarly to from_kron) or chain
• louter -> outer or tensor?

[milan-hl]

What about the wording for the documentation? bond leg becomes virtual leg? Do we keep the name physical leg?

[dsuess]

Yes, I think that's OK unless you have a better suggestion ⁣Sent from TypeApp ​

…

On Oct 20, 2017, 19:32, at 19:32, Milan Holzaepfel ***@***.***> wrote: What about the wording for the documentation? `bond leg` becomes `virtual leg`? Do we keep the name `physical leg`? -- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: #31 (comment)

[milan-hl]

Fixed in the refactor branch with fe3e95a and all the preceding commits.