[Python-Dev] Draft proposal: Implicit self in Python 3.0 (original) (raw)

Fredrik Lundh fredrik at pythonware.com
Sun Jan 8 15:25:01 CET 2006

Thomas Wouters wrote:

> Only in the most severe cases does it make sense to create a PEP > specifically to be rejected.

Yet if it isn't recorded, people will keep bringing it up. How about a 'rejected ideas' PEP for ideas that are right out no matter how people argue? A single PEP, with oneliners to describe ideas, one or two lines to explain 'why not', references to the python-list or python-dev discussions, if any, and a sign in big friendly letters saying "if you really must discuss these subjects, do it on python-list, not python-dev". Some of the stuff that could be in there: - implicit rather than explicit self: invalidates too many tricks - omitting () on functioncalls: like implicit self - changing all statements into expressions (e.g. logix): python isn't (going to be) a functional language - methods for tuples: tuples are records, not collections; use lists instead - sigils to indicate 'self.' (e.g. @foo): loses readability, wins too little - '?' and other non-alphanumerical characters in identifiers: like sigils - strict private/public accessing: 'fake' protection; use closures instead etc. No need to come up with them all, I'm sure all the prize ideas will pop back up eventually ;)

the FAQ touches many of these. maybe a couple of clarifications to the relevant FAQ texts (explaining things in terms of design tradeoffs, rather than absolute truths) would be good enough ?

More information about the Python-Dev mailing list