Re: xyW3 xpl memory etc

["..." yesss@xxxxxxxx]

≪ You've skinned what was, in my view, already a pretty
thin cat. ≫ --Robert Holmgren 

I admire your restraint in not pointing out what must be
obvious to you, that I was able to write my alternative
only because of your fertile work on xyW nulls and your
clarifications to me. I owe you bigtime and wouldn't
have made a peep if the macro hadn't accompanied an
unprovoked insult to the considered decision many of us
made to stick with xyW3 despite v4's indisputably superior
xpl tools and memory management. No v3 user that I know of
(well, maybe one dual user) feels compelled every time
s/he mentions v4 to belabor its flaws.

≪ Shall we revisit your !CL2@ routine. [...] More to the
point, one loses in Xy3 in terms of effort and hours to
write routines, the arduous grind to get from point A to
point B. Everything I do in my Xy4 routine is simple and
straightforward--"baroque" my foot; of necessity, your !CL2@
or whatever you call it is full of tricks and contortions
(and not transparently easy to understand either, with it's
complex entwining of subroutines scattered in many files).
My god, look at _XyWise! ≫

I do, often. Perhaps you should too. I snipped some of
your criticism of !cl2@ because I can't defend a long-gone
pgm whose content and purpose I've long since forgotten.
I do remember that !cl2@ was an ldpm macro in !xyWise's
precedessor. !xyWise has no ldpm files (at startup,
!config.lib ldpms on the fly), and is in constant
evolution. If the existing !xyWise bombs, please email
me details of the circumstances so I can fix it. In my
own usage it's 100 per cent stable. It is quite true that
it deliberately ranges the resource spectrum to minimize
memory drain. That's why those resources are there.

While I envy v4 xpl tools, v3 xpl allows a workaround
--admittedly, sometimes crude (cf !.LIB's {va$pr} kludge)--
for every single omission that's important to me. I find
the trade-off worth it, but think anyone's choice of xyW
version deserves respect.

Tim made several irrefutable points about all versions of
xpl. Powerful it is--and at best a mess. If Sprint weren't
exclusively a menu'd word processor and if it had a fully
remapable kbd I'd probably have switched just for the user
programming language, a subset of C. I dislike BASIC,
but given xyWin xpl flakiness, if the SW beta is more
than the vaporware it increasingly seems to be and
offers VB as a user language I'll port in a NY minute.
(More than halfway through October nobody's even bothered
to note that September was the month we were supposed
to see beta SW fersure. Yawn.)

≪ So how does Xy4 do the same thing as your !CL2@?
	 use
tricks and contortions. Programmers do that at times.
In both cases the variables are defined invisibly and
once defined are simply there for the picking.

≪ Now that's simple! Anybody can do that. Those are
accessible tools. ≫

If {pv105} doesn't qualify on all counts, I've wasted a
lot of time. At the end !xyWise pgms usually issue {pv110}:
places the pre-launch command on the CMline, uses proc
{is105} to move the cursor to {va$cx} and restore pre-launch
{va$gm8} and {va$tx}. In xyW4 you call {pv110} Q2 .
That used to be considered a xyW3 impossibility.
I hope that faint aroma I detect isn't sour grapes.

≪ Anyway, your views notwithstanding, OOM (Out-Of-Memory)
is an ever-present consideration in Xy3, whereas I've never
even *seen* the error in Xy4. ≫

I don't doubt for an instant that OOM is an ever-present
danger if you weigh down xyW3 with huge Help (.HL3) and
big printer (.PR3) files. If you burden xyW3 permanent
memory that way you guaranted xpl memory probs you won't
have otherwise. That's why !xyWise balances resources.
Not that .HL3 isn't an indispensable tool--four frames
are crucial to full !xyWise capability. But .DLG came
to the rescue just about the time it became clear that
dumping all your xpl into one resource is a myopic
approach to xyW3. ... Ciao. 		--a

======================================= adpFisher  nyc
http://www.escape.com/%7Eyesss/_xypro.html ===========================