James Bonfield

Is it like this: https://manpages.debian.org/unstable/biobambam2/bamseqchksum.1.en.html I believe we use this locally to offer similar functionality, but there has been discussion about having something native to samtools too. That tool allows...

I've thought about similar things in the past, but never really got my head around the actual need for perfection. However as you say, maybe some people interpret the rules...

This is the "birthday paradox". See https://stackoverflow.com/questions/14210298/probability-of-collision-when-using-a-32-bit-hash for details of a 32-bit hash. Basically a 50/50 chance is of any two hashes colliding for bit-length N, is more or less...

Correct. If you are designing barcodes rather than randomly constructing them, then a set can be designed that avoids collisions. Maybe we should add a tool to take a list...

It depends on the tool. For textual based ones, absolutely automatic stdin/stdout makes a lot of sense and as you say is quite standard in Unix. For binary ones, stdin/stdout...

Problems still to consider. How does this work from htslib, which although having this as a submodule doesn't run the configure script ind instead builds the source direct. 1. Htslib's...

Good thought. Possibly. It'd need validating whether it also solves the problem on the Mac, but given it was a compiler warning about missing prototype as well as a link...

So... pros and cons. The configure.ac is simpler: just 1 line as you suggest. However the code that uses it is now ugly, made more so by the 0/1 nature...

I added those changes as a separate commit anyway, so we can either squash them together which means we get the new mechanism, or drop that commit and keep my...

I did initially try as you suggest, but the if-found action is called for each in turn, so it ends up as an OR instead of an AND operation which...