Hi Nathan,
What about putting the code up on github, so that others can see it, and
understand what it does. Then we can start discussing where it belongs. It
should be feasible to extend IRanges, even at the C level. Also, I hope you
don't mind that I've cc'd bioc-devel. This sort of thing belongs on the
public mailing list.
Thanks,
Michael
On Mon, Aug 19, 2013 at 8:57 AM, Nathan Sheffield <nathan.sheffi...@duke.edu
> wrote:
> Hi Michael,
>
> I'm writing because you're listed as an author of IRanges. I remembered
> your name because I had communicated with you earlier on the bioc mailing
> list (but you probably don't remember, and that's OK).
>
> Anyway, first let me say I'm impressed with IRanges and it works very
> well. I have a specific use case that is probably not very common, but I've
> developed some code to supplement IRanges and I thought you might be
> interested...
>
> Briefly, I took your viewSums functions and changed them just a bit to
> something I named viewMuls, which looks across all the views, and
> multiplies each one piecewise by a vector (which is supplied as an
> additional parameter). I am using this to do density estimation from read
> counts, and it works really well and it's really fast.
>
> It works fine for me, and I essentially forked IRanges... but maybe others
> would find this capability useful, and it would be easier for me to
> maintain with upgrades if I could integrate back into the original...
>
> I have a working modified version of IRanges with the added capability.
> It's a bit hacked together, though, because I am only just beginning to
> learn how R/C bindings work, but it functions well for my use case. It
> would probably take a bit of additional work to make sure it is clean and
> everything for many use cases, but the main stuff is already there. Are you
> interested in it? Do you generally accept code contributions like this?
>
> Alternatively, I could make a separate package -- I have been trying to
> just pull out my additions and create a separate R package that would rely
> on IRanges, but the functions are so intertwined with IRanges
> infrastructure that this is proving to be harder than I thought. I don't
> know if it's possible to just link to the IRanges code somehow, through a
> SO library or something, but I can't find much documentation on how to do
> that so I'm not sure it would work.
>
> Let me know what you think.
> Thanks,
> Nathan
>
[[alternative HTML version deleted]]
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
