I was hoping that I could use a list/tuple of separators for the split method:

    "abcde".split(["b", "d"])

vs.

    import re
    re.split("b|d", "abcde")

That's one extra LOC, and (worse still) forces me to learn regexp.

Does this suggested 'improvement' make sense?

> Does this suggested 'improvement' make sense?

Sorry, it doesn't make sense for a number of reasons.

* First the str.split() is already too complex.  It took years to get the docs for it to be correct and informative (two different algorithms, etc).

* Second, it would overly complicate the implementation and preclude various kinds of string search optimizations (i.e. that is why fgrep, grep, and egrep are still separate).

* Third, at some point, when you need a regex, you just need to use a regex.  It would be a mistake to reinvent regex capabilities in string objects.