Re: one size fits all vs two-step

> I read both draft-loughney-newtrk-one-size-fits-all-00.txt and
> draft-carpenter-newtrk-twostep-00.txt.
>
> I like twostep much more, because it still includes a notion that 
> obtaining
> interoperability information is important, and if a spec has
> knobs which aren't used or or implemented, at least the fact should 
> be
> documented if those can't be reasonably removed.

This would not be unreasonable, if we made more effort to obtain and 
document interoperability information.

My problem with the current process is that it is engineered for the 
case which is not the common case - count PS, count DS and count IS, 
for a proof of this.

We could change the current process in more than one way - I've 
proposed three versions so far, since 2003. Meanwhile, 2026 still says 
what it says.

I would really like to change it to what we are actually doing, 
whatever that is. If we change it to what we WILL be doing in the 
future, that could be lovely, too.

Spencer

.