Jacobi

I think that it can be implemented in constant time for floating point numbers. For integers I know that the exponentiation by squaring takes O(log(n))

There are plenty. The analysis of basic algorithms involve math (e.g. QuickSort, binary search, etc). Programming a user friendly website can involve advanced mathematics, for example you want to recommend to your users a product that might interest them (Collaborative Filtering)

The PDF document format is incredibly complex. It can incorporate JavaScript actions !

I think that MS R&D is valuable for the research community because they do publish in peer reviewed journals.

He works at Netgem a french start-up. It's not really a big corporate.

For me sometimes it switches to Arabic even if I set the settings. And here in north Africa most people are francophone and prefer French or English !

I appreciate the OP's efforts, certainly sharing code is a good thing. And one of the benefits of the sharing is to get a lot of feedback. Criticizing implementation details is important to spot, discuss and hopefully fix potential problems.

I don't know why it uses a garbage collector when simple reference counting would suffice. Also I hate libraries with global states/variables like this one ...

One could exploit the continuity of the image regions. (The similarity of neighbors in intensity and colors). Also the output is an image with the same histogram ...

The migration from MS Office to LibreOffice is not straightforward, especially if you have millions of documents. In my experience, any complex document will almost always fail to render correctly ...