Typo in uBLAS documentation

[jseakle@…]

Section 2.2 of the documentation is entitled "2.2 computed assignements", and should be "2.2 computed assignments".

[jseakle@…]

Looking at the documentation some more, it appears that it is in fact rife with grammatical errors. Here are a few more fixes:

WRONG: Set the init to true is equivalent to call w.clear() before axpy_prod.

RIGHT: Setting init to true is equivalent to calling w.clear() before axpy_prod.

The next sentence is pretty incomprehensible: Up to now there are some specialisation for compressed matrices that give a large speed up compared to prod.

Does this have anything to do with the init argument? What does 'Up to now there are' mean? I can't fix this because I do not understand the intent.

WRONG: However, the total speed depends very strong on the combination of blocksize, CPU and compiler.

RIGHT: However, the actual speed depends very significantly on the combination of blocksize, CPU, and compiler.

[jseakle@…]

and what looks like an actual bug:

C = block_prod<matrix_type, 64> (A, B); w = A * B

That w should be a C, right?

[jseakle@…]

Oh, my apologies - all of this is relevant to section 2 of the Overview of Matrix and Vector Operations, itself a part of section 2 of the Documentation as a whole.

[Gunter]

(In [61880]) see #4202: fixed some typos and grammar mistakes

[Gunter]

Most of the documentation was written by non-native speakers. So feel free to send more typos and grammar improvements.

[Gunter]

has been fixed in 1.44