On Thu, 2003-04-24 at 07:42, Mike Holderness wrote:

In-Reply-To: <E198MfK-0001nq-02@mail.python.org> Quoth Norbert Bollow <nb@thinkcoach.com>:

...

What ever happened to rationality? When did emotional moping take precedence over sanity?

When was there ever a rational definition of "sanity"?

Start with some Foucault (in French, s'il te plait) the questions. Including how the concept "sanity" may

• not that he's right, but that'll get you set on

be used as a coercive instrument of power.

At some point conceptual reality must meet measurable physical reality. The degree of efficiency of that interface defines "sanity". You can only pull that "relativism" card so many times before it gets all worn out. Yes, those who define "sanity" wield a lot of power. Take a close look at the politics of those who currently define "sanity".

The perception of measurable physical reality is a decentralized function and not controllable by any committee. Is it sane/feasible to cater to every emotional whim on the planet no matter how divorced from fact that emotional reality is?

Which, to get back to software, is the whole point of internationali{s|z}ation: to take account of local preferences and expectations *as they exist* so that the product is usable, with the fewest rough edges, by all.

I like this approach.

Think Oxford Dictionary (descriptive of how language *is* used) not Fowler (prescriptive).

So, we go for the ISO codes.

With 'en-co' - Cockney, innit, love - as the default, and list other variants for the weird :-)

LOL. I hadn't noticed that particular code!

I say, if enough people of 'en-co' bent are interested enough, let them write their own module! I'd love to check it out when it's done.

Trying to satisfy the world with committee has nearly always, resoundingly, and utterly, failed.

Give people the elbow room to solve their own problems and magic nearly always resoundingly happens.

So how can we make it easier for people to create pluggable charset and language modules for mailman? That is the question that some are addressing and some aren't. Using a standard (like ISO) to categorize the menu language choices makes a lot of sense, of course, because it is generally accepted. But why not make it flexible enough to handle ad hoc solutions like cockney English since that would not involve a custom charset and would only involve the use of alternate verbiage?

It is an impossible task to paternally cater to all possible morphings of human utterance. But it is far easier to allow for these variations should someone take and interest in some variant and want to do the work.

Marty