Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!ucsd!ucsdhub!hp-sdd!hplabs!hplabsz!mayer
From: mayer@hplabsz.HPL.HP.COM (Niels Mayer)
Newsgroups: comp.windows.x
Subject: Re: To UIL or not to UIL?
Message-ID: <4855@hplabsz.HPL.HP.COM>
Date: 21 Feb 90 20:38:08 GMT
References: <11658.632634712@hplnpm> <1240@zip.eecs.umich.edu> <1990Feb18.175510.10634@alphalpha.com>
Reply-To: mayer@hplabs.hp.com (Niels Mayer)
Organization: Hewlett-Packard Labs, Software Technology Lab, Palo Alto, CA.
Lines: 40
Summary:
Expires:
Sender:
Followup-To:

In article <1990Feb18.175510.10634@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes:
>I tend to like your suggestion of starting all over again with a new compiler
>for the UID files, although I'm afraid it may be too much work.  Putting a front-end
>on things at least gives some compability, although it does leave you with
>the back-end hacks.

Why do you want a compiler that spits out a file that needs to be
"interpreted" by the client-resident UIL code?

Why not just have an interpreter?

The Motif toolkit is more of an intepreter anyways in that it interactively
responds to widget-object creations and messages sent to those objects.
Until we have a toolkit that compiles into straight X11 code, I think that
compiling into a UID file and then interpreting a UID file to create the
widgets is a circuitous route to take when you can just specify the toolkit
calls for widget creation and messaging in the first place.

I said something to this effect in a previous note in this string, but
nobody commented on the above assertion.

WINTERP takes the interpretive approach hinted at above. If you prefer to
have your description of widgets be hierarchical, or organized in a fashion
that doesn't depend on the creation order and widget parentage, you can
always write a parser (in XLISP) to read in the appropriate description and
spit out "low level" winterp-lisp. The end result will still be readable
and alterable, unlike a UID file. And of course, since the interpreter is
always there, you can make interactive changes to the results of the
"compiled" description too.

[Note to DEC -- this is not a flame against UIL; it's an attempt at
starting a technical discussion... so please don't forward this note on to
my management in an attempt to "get me in trouble".]

-------------------------------------------------------------------------------
	    Niels Mayer -- hplabs!mayer -- mayer@hplabs.hp.com
		  Human-Computer Interaction Department
		       Hewlett-Packard Laboratories
			      Palo Alto, CA.
				   *