Summary: for better international support, 1) change validation of code in lookup.state to [A-Z0-9_]* and 2) in lookup.suburb increase the max length of the postcode field to 10.

Discussion:

In the lookup.state the validation for code is '[A-Z]*' - this is different for the normal validation for the code field which is '[A-Z0-9_]*'

I got bitten by this when loading postcodes for South Africa. SA does not use provinces (its equivalents of states) in its postcodes, but rather 'areas' - which are effectively the name of the nearest reasonable sized town.  There are some 1061 of them, and a number of them are two words (eg Wingate Park).

When generating the xml data, I used my standard recipe to convert something to a code - uppercase and remove any ' and replace spaces by underlines.

However, for the two word areas, this ran foul of the [A-Z]* validation.

Can I suggest that we change the validation of code in lookup.state to the same as that in lookup.suburb and elsewhere - ie [A-Z0-9_]*

Note that the UK also uses a similar system or 'locations' and some of these are two word.

The UK situation reveals another problem - the postcode field in lookup.suburb has a maximum lenght of 6 characters. The UK postcodes can be up to 8 characters long (in the format 'AA99 9AA').

From http://en.wikipedia.org/wiki/Postal_code it would appear that we should increase the max lenght of the postcode field to 10 characters.

Regards, Tim G