Dmitry Selivanov

Would be great to have them as well.

Fork backend potential issues - external pointers

3

Looking to [fork_driver.R](https://github.com/vertica/ddR/blob/master/ddR/R/fork_driver.R) we can see that `dmapply` is essentially `mcmapply`. So we rely on the fact that every object can be easily lazily copied from master to worker process....

S4 generics for coercion function

1

I think it make sense to extend S4 method `as` instead of using `as.dlist`, `as.darray` and so on. Correct me if I missed something.

worker to worker communication

1

I read the article [dmapply: A functional primitive to express distributed machine learning algorithms in R](http://www.vldb.org/pvldb/vol9/p1293-ma.pdf) and found following diagram: That was surprise to see peer-to-peer communication. I started to...

Hi! thx for you work. I found minor issue: ``` r library(Matrix) library(rfunctions) xx = 1:10 m1 = sparseMatrix(i = xx, j = xx, x = xx, dims = c(10,...

Support for up to 64-bit signatures

1

`data.table` works nicely with `bit64`. For Rcpp integration [check this thread](https://github.com/RcppCore/Rcpp/issues/33).


Current algrorithm is **vectorized** and reasonably fast, but it can be even faster.

- hash function family construction - S-curve visualization

Dear `rmongodb` users, today I received letter from CRAN: > See https://cran.r-project.org/web/checks/check_results_rmongodb.html . > The Fedora results are now from Fedora 24 which has mongodb 3.2.7 and it seems your...