[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
- To: <ravi>
- Subject: FromNess, ToNess
- From: Mark S. Miller <mark>
- Date: Thu, 1 Feb 90 00:20:29 PST
- Cc: <bobp>, <xanatech>
- In-reply-to: <Ravi>,53 PST <9001311606.AA27060@xanadu>
Date: Wed, 31 Jan 90 08:06:53 PST
From: ravi (Ravi Pandya)
Summary: I think modest evolution of the current set of Waldos in
Docs&Links will eliminate the need to expose Berts&Orgls.
I've just read all the succeeding correspondence. My overall take is
that you're probably right. If so, that'd be *great* (as an
explanation reduction). However, this clarifies how we should clean
up our terminology to make the issues clear.
There are really three distinct layers:
Docs & Links
Waldos, SubHands, & HandRecords
Orgls & Berts
Assuming it is the case that D&L are built only on WS&H (which would
require removing knowledge of LinkEnds from WS&H on down (which is a
good idea anyway)), the real question is: does (or can) the WS&H layer
make available all the useful power of the O&B layer? I think the
answer may currently be no, but extending WS&H so that the answer is
yes may be a much better thing to do than exposing O&B!
Nevertheless, I think we should also all proceed on the conservative
design principle that the O&B layer interface must be engineered so
that it can be eventually exposed. Even if we don't expose it, this
assumption will be proper discipline to guide its evolution.
And all this time I've been thinking that O&B is the *TRUE XANADU* %)