From: Edward C. Zimmermann
> It does not mean the same thing. 1950/1959 is an interval (a set of
> points, linearly ordered containg all the points in-between the start
> and end) and its precision is that of year. 195x is a point and its
> precision is 10 years. 

Are you saying that 195x means some instant in time within the interval
1950/1959? That's certainly not what I thought was intended. 
Now while I don't suggest that that isn't a real requirement, and it
certainly isn't otherwise supported by the spec, if you want to say "some
instant in time with the interval 1950/1959" why wouldn't someone want to
say "some instant in time with the interval 1955/1958"?   The x notation
won't be of any use for that, it's only good on a decade boundary.  

> I think I did a rather long expose of points and intervals some time
> ago..

I think we're all clear on the concepts.

> Or shall the chorus now suggest that 1999, 1999-02, 1999-02-01, 1999-
> 02-01T12:00 are all intervals too?  

No. Nobody is (still) suggesting that. I don't think we need to rehash it, I
think we all understand the model. However, what I don't want us to have to
do is try to articulate it in the spec. 

> We don't need or want to define the concept of decade, century or ..
> But I think we do perhaps want to have a concept of 10 years, 100 years

Accepting that for argument sake, if the 10 year period must start on a year
divisible by 10, and the 100 year period on a year divisible by 100, that
doesn't give you a way to represent an arbitrary 10 year or 100 year period.

Nevertheless, if you're happy with that restriction, and since it seems that
only you are going to use this, I don't have a problem with defining it as
you suggest. 

> This is not the domain of CQL, at most its for CQL to accept our work
Yes that's how I see it.