re: defaults and indirect references

At 11:20 21/01/99 -0500, Reynolds Franklin NRC/Boston wrote:
>
>Logically separating delegation (or indirect references) from defaults
>(and the ability to override defaults) is reasonable. But I think
>from a practical point of view, they are much more useful if both features
>are available. I think that indirect references , without value overriding,
>can only be applied to platforms or resources that don't change. Invariant
>resources may be a common case but failing to take into account
>resources or platforms that can change seems like a mistake.

Oh yes!

I believe that orthogonal design helps to promote the kind of synergistic
application of elements that you are proposing.

>
>Default processing without delegation also seems to have marginal
>value to me. If there is no delegation then there is only one source for the 
>profile.

That term "delegation" again:  I think we have different understandings.
My previous message has failed in this respect.

I think we are assessing the same goals here, but are describing them in
different ways.  I meant "delegation" to mean roughly what you mean by
"default processing".

(I mean "delegation" between feature sets, you mean "delegation" between
systems, I think.  IMO, because we are not dealing with transfer protocols
here, I don't think delegation between systems is a useful concept.  Hence
I have latched onto Bill's "indirect reference", which says simply that
here is a reference to some feature set that is defined in some other place
-- which I think is as close as we can usefully get here to delegation
between systems.)

#g

------------
Graham Klyne
([email protected])