John Marshall

This was discussed at yesterday's meeting, and there was agreement that there is room for improvement here. In particular, §1.6.1/7 says “if [circumstances] then FILTER must be `.`” but does...

There is such a building guide, and now there is a link to it in the README on the main repository page. Circle CI artifacts [expire after a very reasonable...

The topic of hard clipping came up during the original PR discussion, in https://github.com/samtools/hts-specs/pull/418#issuecomment-726143844 and the following comments. I believe we concluded that using `MM:Z:N+m,4` to represent this modification of...

You're right that — either way — this needs more details in the spec. The N notation enables non-MM-aware mappers to hard clip reads and leave MM as is, such...

#593 makes a case that `[END_OF_VECTOR, END_OF_VECTOR]` would also be a reasonable encoding of this situation. What is the reason for requiring that “[e]mpty vectors (i.e. vectors with no data...

Thanks, that's looking pretty nice, and seems like a useful convention to establish. Some thoughts from a fairly quick reading: * You don't actually chop up the BAM file on...

CSIv2 was never implemented (or the format design considered workable) because no conclusion was ever reached as to how to correctly include counts from the higher bins (see mailing list...

> I agree about having a more general specification for the format. I'll create a new document for it. Don't worry about a separate document for this, just organised as...

The previous text had > The index must contain the virtual file offset for the first alignment (in the case of a non-empty BAM), and a virtual file offset for...

What the original text about the sentinel last entry in the list is trying to describe is: the virtual file offset of the byte after the end of the last...