Re: 1.1 FPWD: error in timeOnTimeline function

On Thu, 2004-07-29 at 09:14, Dave Peterson wrote:

> The various algorithms would be a lot less susceptible to this kind of
> error (and would be for the most part easier to implement) if the
> property values stored were the local values rather than the normalized-
> to-Z values.  C'est la vie.  (Hint for implementers:  You aren't
> required to actually store Z-normalized property values; you may find
> it easier to store the local values instead.)

This sounds as if you might want to make a proposal to
change them.  The WG did decline to change them when you
brought the topic up at the time we approved this material
in the first place, but speaking for myself my rationale
was not that one was preferable to the other but that
we had a draft for one way and we didn't have a draft
for the other way.  I was unwilling to approve a draft
we didn't have in hand, but speaking for myself I'm not 
unwilling to consider a change, if a draft showing how 
it works is available.

Michael

Received on Thursday, 29 July 2004 13:27:12 UTC