On 2011 Mar 28, at 07:57 EDT, Ketil Malde wrote: > Felipe Almeida Lessa <felipe.le...@gmail.com> writes: > >> I have a small comment: why not use Int64 for Offset as the bio >> library uses? Int is only guaranteed to have 30 bits, which means >> that we can only use Offset on the first 0.5 Gbp, which seems rather >> limiting :). > > Another reason to use Int64 is that this is what (lazy) ByteString > uses, which was perhaps the main reason for using it in biolib. > > I guess a lot of people are now using 64 bit computers, where > GHC's Int is also 64 bits, and this might hide some bugs - and > potentially rather nasty ones, as overflowing numbers silenty wrap > instead of causing exceptions.
Both good points--I tend to use strict bytestrings, but the conversion shouldn't be too costly. I'll make sure the change doesn't have a big impact on performance for my uses and probably release a new version of seqloc along with the BED & GTF parsers. I do think that (Integral a) or Integer could cause a bigger performance hit--most of my uses of seqloc are large sets of CPU-limited "location arithmatic" on deep sequencing alignments. Best, --Nick Nick Ingolia n...@ingolia.org _______________________________________________ Biohaskell mailing list Biohaskell@biohaskell.org http://malde.org/cgi-bin/mailman/listinfo/biohaskell
