Scope of instantiated class

Aahz aahz at
Fri Aug 2 21:34:41 CEST 2002

In article <lku1mpl69n.fsf at>,
Michael Hudson  <mwh at> wrote:
>Peter Hansen <peter at> writes:
>> Aahz wrote:
>>> In article <3D3B4F80.B0D6BE07 at>,
>>> Peter Hansen  <peter at> wrote:
>>>>I'm trying to follow Aahz' advice and avoid "reference" in favour of
>>>>"binding" etc, but I guess I carried it too far! :)
>>> <grin>  Your problem wasn't in your use of "binding", but in your use of
>>> "name; I use "target" to refer to the general case of thingy that can
>>> contain a binding, of which names are a subset.  "Names" are words
>>> attached to a namespace (i.e. function local, module global, builtin, or
>>> object attribute).  Another target is an index/key for container
>>> objects.
>> Hmmm... "target" sounds like the wrong direction.  Shouldn't it
>> be a "source" for the binding?  Or maybe another term which doesn't
>> have what seems to me the potential for confusion would be better.

Well, I'm all for alternatives, but target works for me because it
focuses attention on the end of the binding that causes the most
confusion.  That is, people seem to have little trouble with the fact
that bindings are essentially references to objects; it's the from side
that causes confusion.

>Common Lisp calls them "places", which is a reasonable name.  It's not
>that much of an issue in Python as there aren't that many of them
>(names, attributes, subscripts, slices).  In CL, you can define your
>own (tho' it's a bit tricky).

Hmmm...  That doesn't sound quite right to my ear.  <sigh>

All right, let's go to the thesaurus for some alternatives:

    Naw, Python doesn't have gotos for confusion, but people used to
    other languages won't like it.

sigil, totem
    Hmmm...  I like these.  Both synonyms for "symbol", sort of, but
    without any CS-centric connotations already.
Aahz (aahz at           <*>

Project Vote Smart:

More information about the Python-list mailing list