On Wed, 6 Sep 2006, Sven Schreiber wrote:
Riccardo Jack Lucchetti schrieb:
> On Wed, September 6, 2006 21:05, Sven Schreiber wrote:
>> Allin Cottrell schrieb:
>>
>> Is that really necessary? For example, I was just told that in a popular
>> German panel dataset (GSOEP) the years-of-schooling variable can also
>> take on half-year (.5) values. But nothing in between, so it's still
>> discrete.
>
> You can always multiply it by 2. Not very clean, but...
Yeah right ;-)
Seriously though, it's an explicit user action, so why restrict it at all?
OK, I can see the case for making it unrestricted. But I'm not
goint to do that for 1.6.0 because it might break some things.
BTW, I've fixed the frequency plot business for discrete
variables in CVS.
Also, IIRC you guys agreed to mark discrete variable in the main
list
(Icon, or 'd', or whatever). Is that still on?
Probably after 1.6.0.
Allin.