binary block designs

[Peter Dobcsanyi]

On Mon, Jul 21, 2003 at 03:53:54PM +0100, Leonard Soicher wrote:
> We need to make clear in the ext-rep and in all its documentation that,
> for now, we are only dealing with *binary* block designs. 

I agree. However, I suggest to do that in the least complicated way for
now, that is simply spelling out this semantical fact in the ext-rep
documentation. It is slightly against explicitness but ... see below.

> Can this be specified in a compulsory indicator?

The whole indicators subtree is optional and all of the indicators are
optional.

> Or in a compulsory attribute of <block_design>?

In whatever way we would go that would indicate that we have (or at
least could have) a choice here: binary or not binary. The problem is
that when we allow not binary some of the current semantical details
and/or structures of the the ext-rep need to be modified, maybe
significantly. So in fact, with the current form of ext-rep there is no
choice, this is good for binary designs only. Which is fine since that
covers a lot.

No to mention that considering our Aug 10 deadline, we are not really in
the position to analyze and foresee all what could come with non binary
designs.

--             ,
    Peter Dobcsanyi