[Biopython-dev] SeqRecord per-letter-annotation : avoid lists?

[Peter]

Hi all,

I was thinking that about the SeqRecord object's letter_annotations,
and that perhaps we should only allow strings and tuples (which are
immutable), but not lists.  Because lists are mutable, the user can
(accidentaly) alter the list such that its length doesn't match that
of the associated sequence (which would be bad). Currently we do use
lists in the SeqRecord's letter_annotations, e.g. for qualities. I
don't recall having any particular reason for using a list rather than
a tuple.

Any thoughts on this?

Peter