None of this answers why a new language is needed and you can't just do whatever is in this language in a C++ iterator.
The paper is mostly about data structures, this whole thing could be a class in C++, then you could use the entire C++ ecosystem like optimizing compilers, IDEs, other libraries, profiling, static analysis, etc.
When someone invents another language, the entire programming ecosystem is wiped out and everyone has to start over.
You're still having trouble reading so let's try socratic method: if I have a vector with 10 elements and a vector with 100 elements, which iterator from stl should I use to take their dot product?
Lol now I know you're just googling for keyword matches - a parfor has absolutely nothing to do with this.
Overall you seem to be very upset about some perceived cognitive burden/imposition at having to learn a few new ideas. Let me free you from that burden: you don't have to understand co-iteration. But it is possible that one day your code will be slower than someone else's that does understand co-iteration.
Overall you seem to be very upset about some perceived cognitive burden/imposition at having to learn a few new ideas.
The ridiculous part is that you think there are new ideas here.
you don't have to understand co-iteration
I understand it because it is already called iteration.
But it is possible that one day your code will be slower than someone else's that does understand co-iteration.
Feel free to explain in detail the part that is new or faster. You haven't done that at all, your entire post is just saying 'nu uh' with no actual evidence to back it up. Give each thread their own ranges that don't overlap. If you want to make up terms and sell pretend results don't be surprised when people aren't convinced.
Repeating RTFM (especially in such tone!) won't convince anybody to look into this. Quite the contrary…
You, my friend, are dismissive to the fact that people are only willing to put in their valuable time into something if it seems at least on first glimpse interesting enough.
You're doing an extremely poor job showing people why this would be the case here.
Math may be fun, but in the end your success in life will almost exclusively depend on your social skills. Skills which you're obviously lacking still.
>You, my friend, are dismissive to the fact that people are only willing to put in their valuable time into something if it seems at least on first glimpse interesting enough.
simple solution: don't pontificate/comment if you don't have any valuable time - no one asked this guy his opinion on co-iteration.
also the comment you directly responded to doesn't say RTFM - it literally poses a puzzle that illustrates the point.
Anyone can make an iterator, what does this have to do with anything, let alone creating an entire new language for iteration?
there are no iterators in the stl that iterate over a pair of ragged/jagged collections in parallel
Parallel Iteration: https://en.cppreference.com/w/cpp/algorithm#Execution_polici...
Sparse: https://www.boost.org/doc/libs/1_65_1/libs/numeric/ublas/doc...
Parallel Iteration: https://www.openmp.org/
None of this answers why a new language is needed and you can't just do whatever is in this language in a C++ iterator.
The paper is mostly about data structures, this whole thing could be a class in C++, then you could use the entire C++ ecosystem like optimizing compilers, IDEs, other libraries, profiling, static analysis, etc.
When someone invents another language, the entire programming ecosystem is wiped out and everyone has to start over.