Tuple indexing - no more votes!

From: jaakko.jarvi_at_[hidden]
Date: 2001-06-29 01:56:01

Thanks for the comments!

The solution maximizing overall satisfaction seems to be:

1. Tuple elements should be indexed starting from 0.
2. There should be a set of 'natural language' constants for accessing
them 1-based.

e.g.: get<0>(a) == get<_1st>(a) 

3. This choice and its rationale should be clearly expressed in the
documentation.

We'll change to this scheme.

One additional clarification (which has already been said a couple of
times):

Placeholders in bind/lambda expressions are a separate issue,
they can be named like arg1, arg2, arg3, ...
or free1, free2, ... or _1st, _2nd, ... or something else
Here 1-based naming is more consistent with the established
conventions (cf. bind1st).

Jaakko