PEP 285: Adding a bool type

[Paul Rubin]

Erno Kuusela <erno-news at erno.iki.fi> writes:
> since 1.5, i find the following changes to the language come
> immediately mind as worthy:
> - nested_scopes
> - string methods
> - unicode (maybe, haven't really used or needed it yet)
> - new-style classes (ditto)
> 
> the following language changes come immediately
> to mind as feeling superfluous:
> - list comprehensions (map/filter)
> - f(*args, **kw) (apply)
> - staticmethod/classmethod
> 
> i haven't decided if this is a good ratio or not, yet.

I see it as being a deeper matter than a "ratio".  I think it's wise
to resist making deep changes to the languages without deep benefits.

Nested scopes and new-style classes were deep changes with deep benefits,
so they were justified.

List comprehensions and f(*args) were shallow changes, so they can be
justified even if they don't provide deep benefits.  I think both were
worth adding, though their benefits weren't all that deep.

I don't understand staticmethod/classmethod so I won't remark on it.

I see Unicode as correcting an implementation deficiency-- it may have
been distasteful to fix, but it probably had to be done if the
language is supposed to stay current.  There just wasn't much
discretion available.

Bools seem to me to be an example of a deep change with shallow
benefits, so it's harder to justify than any of the above.