Article View: comp.sys.mac
Article #23665Perfect operation under Multifinder -- the sequel...
From: Clive Steward
Date: Sat, 16 Jan 1988 01:18
Date: Sat, 16 Jan 1988 01:18
103 lines
4568 bytes
4568 bytes
Well, I have further good news, it seems. Be sure to read through to the
caveats, though.
With the following, I have perfect operation* with large bunches of
complicated applications, for instance, Hypercard, Superpaint 1.0sp,
Word 3.01, MacProject 1.2, and Macterminal 2.2. I cut and paste big
chunks, run several viewable windows to visually trade information, etc.
Here's how it's set up to work:
1) Clean Desktop as before.
2) Applications re-set to normal out-of-the-box memory allocations
(MacProject doesn't have; it's happy at 150k, probably less)
3) Boot blocks set for no (0) additional System Heap space
4) Finder set for 220k allocation instead of the 160k default.
* see caveats below...
The Finder setting seems consistent with the fact that, using the default
allocation instead, rebuilding the Desktop always fails to complete, with
an alert message. With extra memory, it completes fine. I was running 180k,
which was better than 160, but 220 is what gives complete freedom from errors.
Somewhere between might work, don't know.
The reduction of System Heap allocation to the nominal seems to bring
further improvement, as well as give back 80k of memory. Now, da's in the
DA Runner layer seem to self-allocate as they're supposed to. Not enough
experience to know if it's perfect, but so far da's are just fine.
I'd originally increased the initial heap space because with a small
extra (16k), many da's didn't seem to want to work, unless they were opened
in an application's space. It might be a 'near the boundary, but not over,
so don't do anything' situation, and then the da used a little more, and
crashed. Don't know. Curious, but neat when it works.
Letting the system do what it's stated to want, it seems much happier.
Now the caveats.
1) All of this may have something to do, of course, with the
specific disk device driver for my MacBottom 45 meg disk, even
the version I'm on, and about to change. Programmer types
will understand why.
2) I do get the perfect operation, for hours of really
complicated work, and it's a joy. Have you ever tried to put
together a large project plan, without being able to look at
word processor screens when constructing a Pert chart in
Project? I can't say enough for this, it's what I bought the
original 128k Mac, and upgraded this far, to be able to use.
Apple has brought something you can't touch, anywhere else.
But.
As soon as the Mac heats up, 1-2 hours in air-conditioned office,
half hour in overheated winter room, it's that close, something
starts letting go, and I get bombs within seconds or minutes
of startup, often on simple mouse movement (thus I suspect SCC
over-interrupting, overfilling a queue). The funny thing is that
it still works perfectly without Multifinder, but again, other
programming/hardware types will see ways this could happen.
So I have to get the Mac fixed, before it's assured that results
so far are really valid. I think they are, and that the heat
problem is simply additional grief.
Have to see if I can arrange a trial swap. This is what
(established friendly) dealers are for.
Cheers,
Clive
Plus one more. I have an even better running system now, by increasing
the Finder allocation to 220k, and setting the boot blocks so that the
System Heap is 0 over 'normal'.
It's absolutely error free, running 6 or seven complicated apps.
For an hour or two, until the Mac heats up. When cool, it's good again.
Rats.
I will publish this, I guess, but with the caveat that the behaviour
just might be caused by whatever (I have reason to believe SCC) breaks
when hot.
I do believe, though, that the increased allocation is valid, since it
prevents the 'Can't finish rebuilding Desktop' message. Might be because
I have a full 45 meg disk, or due to peculiarities of the disk device driver.
Well, now I have it essentially written, so guess I'll post.
Thanks, and good luck.
Clive
Technical Details
Message-ID:
<6473@drutx.ATT.COM>Path:
arc1-ca.pugleaf.net!utzoo!utgpu!water!watmath!clyde!burl!codas!mtune!mtgzz!drutx!cliveShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!utzoo!utgpu!water!watmath!clyde!burl!codas!mtune!mtgzz!drutx!clive From: clive@drutx.ATT.COM (Clive Steward) Newsgroups: comp.sys.mac Subject: Perfect operation under Multifinder -- the sequel... Keywords: it's neat all the time, when it works Message-ID: <6473@drutx.ATT.COM> Date: 16 Jan 88 01:18:22 GMT Distribution: comp Organization: resident visitor Lines: 103