SHULMAN@RED.RUTGERS.EDU (Jeffrey Shulman) (10/04/86)
Delphi Mac Digest Saturday, 4 October 1986 Volume 2 : Issue 49
Today's Topics:
2 MB + RAMSafe = Heaven (3 messages)
Uses of Memory (8 messages)
Report on TML Pascal 2.0 (4 messages)
Line Conditioners/AC power (7 messages)
MonsterMac/Switcher ? (2 messages)
floating pt boards
another 800K dies
RE: User Interface Problem: Macintosh Alert Boxes
Re: Misc. programming problems
-----------------------------------------------------------------------
From: IVANOVIC (13325)
Subject: 2 MB + RAMSafe = Heaven
Date: 2-OCT-01:47: Programming
2 MB + RAMSafe = Heaven
I just got a Levco MonsterMac clip-on upgrade to my 512E with
RAMSafe*, bringing it to 2MB. RAMSafe is a RAM disk that doesn't go
away when the system crashes. It consists of two PROMs that fit on
the Levco board and an installation/configuration program. The way
it works is the PROMs snatch the signal that resets memory from the
jaws of death, turning the electrons into sweet smelling ether.
Voila! your RAM disk doesn't go away when your system bombs. Turning
the power off though is more karma than the PROMs can deal with,
though, and that will do the RAM disk in. The RAM disk is
configurable from 512K to 1.5MB in 256K increments. It is HFS.
I've had some problems with the upgrade (system won't come back after
a power off) but I think that's due to the upgrade altering the power
supply voltage rather than RAMSafe. The seller, K.B. Company of
Boston has been quite helpful. They extend Levco's warranty to 120
days from 90 days and also warrant the original motherboard for the
same time. The clip-on Levco board (if removed before an Apple dealer
sees it!) won't void the AppleCare warranty.
RAMSafe itself works well. The system reboots in 6 seconds and that
includes installation of a 73KB TMON. Just for fun, I ran DiskBench
on the RAM disk and got 2.42 seconds (145-146 ticks) for 100 32KB
reads or writes, and 0.07 seconds (4 ticks) for 80 1MB seeks. Wow!
Rebooting is so fast it's hard to turn the Mac off before the welcome
message comes on unless you've got your finger on the on-off switch
already. I like that! If the rumors that Apple is going to come out
with a 8MB SCSI RAM disk are correct, they have one hot product. I
can't wait, though I probably can't afford it.
While on the subject of disk performance, I'd like to observe that
according to studies done by Digital Equipment Corp., the
contribution that the transfer rate of a disk makes to overall,
perceived performance is way overshadowed by that of access or seek
times. Since most disk accesses are for a small number of blocks,
the time spent transferring the data is small compared to preparing
to transfer the data, i.e. the seek times. (At 330KB/s a 2048 byte
transfer takes 6 ms. Average seek times are 65-85 ms, depending on
the drive.) Go for the faster access times, unless you've got a
specialized application like real-time data collection. Hence,
therefore, ergo, the XP upgrade to the DataFrame 20 is in all
likelihood a $99 (?) donation to SuperMac Technology. Now they're
deserving people and all that, but it won't do squat for performance
you'll see. And oh, BTW, what's the meaning of DiskBench's "1 MB
seek"? To me, a seek is the drive's process of setting up for a
transfer. There is no data size associated with it.
P.S. I don't keep anything in the RAM disk I'm not prepared to lose.
I have faith, but I also have experience! Save early and save often,
to paraphrase Mayor Daley.
Update on the upgrade/RAMSafe saga. KB Co. replaced my mother board
and adjusted the voltage to exactly 5.00 volts. I haven't had any
crashes and the system reboots fine. I hope the problem has been
licked.
* RAMSafe is a trademark of somebody or other. (They got the first
part right: put the letters "tm" after the trademark; but somewhere
you gotta say who owns the trademark!) RAMSafe is distributed by
CJS Systems, 3051 Adeline, Suite 1, Berkeley, CA 94703.
(415) 849-3730. It was written by Paul Mercer. I assume that Levco
will sell it to you also. I got mine through the local Levco dealer,
K.B. Company in Boston.
-- Vladimir
------------------------------
From: BRECHER (13335)
Subject: RE: 2 MB + RAMSafe = Heaven (Re: Msg 13325)
Date: 2-OCT-08:16: Programming
DEC's studies most likely used DEC's operating systems -- almost certainly
multi-user. No doubt that access time is important, though.
DiskTimer's "1MB seek" is a seek across 1MB of disk capacity. The
test consists of reading 1 sector at/near the beginning of the volume
and then 1 sector at an offset of 1MB from the first read (this
process done 80 times).
------------------------------
From: MACINTOUCH (13342)
Subject: RE: 2 MB + RAMSafe = Heaven (Re: Msg 13335)
Date: 2-OCT-17:18: Programming
Just a note on recoverable RAM disks. The one in the HyperDrive 2000 also
sticks around until you power off. I don't know how they do it, but I don't
think Paul had anything to do with it <grin>.
Ric
------------------------------
From: IVANOVIC (13326)
Subject: Uses of Memory
Date: 2-OCT-01:48: Programming
On the Uses of Memory and Other Important Matters
With a bountiful supply, I started to think about the best way to use
the 2MB I now have. I started experimenting, and the first
configuration I came up with was 512K for the RAM disk and several
Switcher partitions. I run with TMON and Tempo but without the disk
cache. Well, folks, that doesn't work. The menus end up with
non-ASCII junk in them, even though switching around seems to clear
things up, why I don't know. Anyway, why keep a program on RAM disk
AND in Switcher-controlled memory! Talk about double-buffering!
After several changes, the configuration I currently have is a 1MB
RAM disk, a 73MB TMON installed in high memory before Tempo gets
started and no Switcher. The RAM disk contains the System Folder
and an application with its supporting executables (e.g. TML Pascal,
Edit, Linker and RMover, or LightspeedC). All modifiable data files
are on either a 800K internal floppy or an external 400K floppy.
(My hard disk is on order.) I've now reduced the number of system
bombs to what they were after I upgraded to the 128K ROMs, but still
very much higher than with the 64K ROMs. Has anyone else noticed the
increase in frequency of bombs with the 128K ROMs? (I use System 3.2
and Finder 5.3 but no HD20.) I suppose I ought to keep track of what
the bombs are... but there are so many of them!
Should it be possible to have a RAM disk, TMON or Macsbug, Tempo, and
a disk cache on, plus several Switcher partitions all at the same
time? (All these guys do funky things with memory.) It's really not
that farfetched. I'm almost there. With the imminent installation
of a hard disk, I'll soon want a disk cache. With the rumored 8MB
SCSI RAM disk, I'll definitively want many Switcher partitions. And
I want to be able to run a debugger also. Tempo saves time. With
the current state of affairs, I don't think it is possible to expect
that all those programs will work flawlessly together. Didn't the
IBM PC folks just go through a "learning experience" just like the
one we're about to what with all their RAM-resident pop-up
applications? What lessons did they learn?
-- Vladimir
------------------------------
From: PEABO (13337)
Subject: RE: Uses of Memory (Re: Msg 13326)
Date: 2-OCT-12:02: Programming
I don't see why a properly installed RAMdisk should interfere with other high-
memory-resident products like TMON. As a basis of comparison, you probably
should try RAMStart. Version 1.23 is in the database now, and I have a version
1.3 which should be in the database here this weekend.
The SCSI RAMdisk is external and therefore (if it exists) will not have nearly
as good performance as a RAMdisk composed of addressible memory. Your point
about reduced seek time is well taken though.
peter
------------------------------
From: MACINTOUCH (13343)
Subject: RE: Uses of Memory (Re: Msg 13337)
Date: 2-OCT-17:20: Programming
The TurboCharger 2.0 disk cache lets you select which disks to cache, so that
you can avoid "double buffering" situations.
Ric
------------------------------
From: IVANOVIC (13350)
Subject: RE: Uses of Memory (Re: Msg 13343)
Date: 2-OCT-21:21: Programming
I was curious about TurboCharger 2.0. Do you use it? Pros, cons? The slim
amount of information that I have come across seems to indicate that it's the
most sophisticated of the lot. With statistics gathering ability, it ought to
perform the best.
-- Vladimir
------------------------------
From: MACINTOUCH (13356)
Subject: RE: Uses of Memory (Re: Msg 13350)
Date: 2-OCT-21:52: Programming
I have a very unstable, changing environment at the moment, after
suffering a month without a hard disk and having problems getting
upgraded to a Mac Plus, given the HyperDrive and the stuff the dealer
is giving me about month-long delays for upgrade kits. (Is that
bull?) I like TurboCharger, but don't always use it because of
testing of all these different hardware and software configurations.
If I had a standard Mac Plus and one hard drive, I'd seriously
consider it after checking the TOPS compatibility.
Ric
------------------------------
From: IVANOVIC (13358)
Subject: RE: Uses of Memory (Re: Msg 13337)
Date: 2-OCT-21:58: Programming
In addition to "properly installed" one should add "properly written." I hope
none of the authors have resorted to "tricks" which are System version
dependent.
I just remembered that TMON has a "Launch..." menu item. I now know how to use
RAMSafe, Switcher and TMON all at once.
-- Vladimir
------------------------------
From: IVANOVIC (13370)
Subject: RE: Uses of Memory (Re: Msg 13358)
Date: 2-OCT-23:30: Programming
Well, _I_ can't get RAMSafe, Switcher and TMON all to work. I can get as far as
launching TMON in a Switcher partition, butall goes to pot as oon as I try to
launch an application under TMON. Maybe I should try the system heap...
-- Vladimir
------------------------------
From: DWB (13390)
Subject: RE: Uses of Memory (Re: Msg 13370)
Date: 3-OCT-01:28: Programming
You will definitely have to load all memory resident stuff (TMon,
RamDisk Disk Cache, etc.) into memory before running switcher.
Switcher won't cooperate well with applications which change the size
of memory after it gets started. To take message 13387 one step
further, I have actually used a configuration of TMon, RamStart 1.24,
and Switcher at the same time. It worked quite well. RamSafe may not
play by the rules but I would be extremely suprised to find that it
doesn't.
David
------------------------------
From: IVANOVIC (13327)
Subject: Report on TML Pascal 2.0
Date: 2-OCT-01:49: Programming
Report on TML Pascal v2.0
I am posting this message to Delphi rather than privately to Tom
Leonard because last time I did so (version 1.2), it had no effect.
Tom didn't bother to reply ("Yes, I got the bug report, thanks." was
all I was looking for) and he didn't correct the bugs I submitted.
Notes
"ANSI Pascal" refers to the ANSI/IEEE X3.97-1983 definition of
Pascal.
"ISO Pascal" refers to the ISO dp7185 defininition of Pascal.
Level 0 of ISO Pascal is equivalent to ANSI Pascal.
Level 1 of ISO Pascal incorporates conformant array parameters
over Level 0.
The "User's Guide" refers to the TML Pascal Users's Guide v2.0 of
August 1986.
The "Reference Manual" refers to the TML Pascal Reference Manual,
v2.0 of August 1986.
Omissions
There is NO statement in either the Users's Guide or in the
Reference Manual of compatibility with anything. In the User's
Guide, p. 1-2, Jensen and Wirth (3rd ed.), ANSI/IEEE Standard 770
and Henry Legard's annotations are listed as "good references to
the Pascal language itself." What does this mean? Is TML Pascal
intended to be compatible with ANSI or with ISO Pascal? A
superset of either of them? The Macintosh Pascal Reference Manual
states in line 1 of the Preface (!) what these documents are
(descriptions of the language supported by Macintosh Pascal) and
in line 2 that this language is intended to be ANSI Pascal
compatible. What is TML's equivalent statement?
Procedural and functional parameters (noted as an Implementation
Note) are not supported. This is Level 0 stuff! When is it
going to get supported?
No error numbers are listed with the errors reported in Appendix
A, Section 2 of the Reference Manual. Why list them at all? How
about a suggested course of action for each error message?
The non-terminal "sign" is never defined in the Reference Manual.
The file MacIntf.pas is incomplete. In particular, a number of
I/O completion codes are missing, namely -1 to -32. There are
also others.
The compiler's version numbers does not appear in the listing file.
The linker's version number does not appear in the map file.
The file "Boot Paths" with a system icon does not have any
documentation. Does anyone know what it does? Note: it's not
the "Paths File".
Errors
The compiler reports an error without any message. The error
number is 105.
The compiler reports errors when given the text:
PROGRAM subrange_errors;
CONST
a = 15; { (2**4) - 1 }
b = 255; { (2**8) - 1 }
c = 32767; { (2**16) - 1 = maxint }
d = 2147483647; { (2**32) - 1 = maxlongint }
TYPE
a_type = -a..+a;
b_type = -b..+b;
c1_type = -c..+c;
c2_type = -maxint..+maxint;
d1_type = -d..+d;
d2_type = -maxlongint..+maxlongint;
BEGIN
END.
Errors (# 105) are incorrectly reported for a_type, b_type and
both c_types. No errors are reported for either d_types. The
d_types are permitted by TML Pascal but are NOT ANSI Pascal.
Macintosh Pascal v2.0 compiles the above program but flags the
d_types as illegal.
The compiler reports errors when given the text:
PROGRAM goto_errors;
LABEL
1, 10001;
VAR
x : integer;
PROCEDURE a;
BEGIN
GOTO 10001;
END;
PROCEDURE b;
BEGIN
GOTO 1;
END;
BEGIN
a;
10001: b;
1: x := 0;
END.
ANSI Pascal says that 10001 is an illegal label. The Reference
Manual implies that labels can have values up to maxint. Both
the GOTOs should compile. The compiler flags them as being an
implementation restriction. No such restriction is noted in
the Reference Manual.
Macintosh Pascal v2.0 flags 10001 as illegal and compiles the
rest.
When the /Code option of the linker is selected, there appears at
the end of those procedures which define strings, code without the
usual addresses to the left, but, with instead, 5 characters (in
hex: 4E BA FE 80 3C) unrelated to either the address or to the
string. The code for the strings is presumably correct. The
listing of the addresses continues correctly in the next module.
What are these characters?
Typos
"Miscellaneous" is misspelled in the Table of Contents of the
Reference Manual.
Pages A-4, A-3 and A-2 of the Reference Manual are out of order.
Suggestions
Have a standard/no standard option for the compiler. VAX-11
Pascal compiles anything the manual describes, including DEC
extensions, but will flag as non-portable, optionionally, those
extension it finds.
Have an option to turn off optimization. Sophisticated debuggers
get confused by optimization.
Provide a symbolic, source code debugger. Provide a symbolic,
source code debugger. Provide a symbolic, source code debugger.
Get the hint?
Enough information should appear in the listing and map files to
recreate the identical compilation or linking scenario. In other
words, the listing file should also list the options used, the
date and time and the system version number. The linker should
reproduce the .Link input file along with the system version
number, its version number and the date and time.
It should be possible to create a file that contains the following
information:
The source line.
The corresponding assembly code.
The corresponding object code.
I am very close to the point of abandoning TML Systems. I need a
Pascal that will compile standard ANSI (at least, and better yet
ISO Level 1) Pascal. TML Pascal doesn't do it. I assume (and will
check out BEFORE I buy) that Lightspeed Pascal recognizes the same
language as Macintosh Pascal. If so, it's been real TML. LSP has
an entire development environment that is fast. Now, if it only
supported MacApp...
-- Vladimir
------------------------------
From: MACINTOUCH (13344)
Subject: RE: Report on TML Pascal 2.0 (Re: Msg 13327)
Date: 2-OCT-17:23: Programming
Tom has done an astounding amount of work for a one-man company. I'm
not surprised to hear he may have dropped the ball once or twice on
support. He is online here and on CompuServe, and has an MCI account.
It might be worth trying to send a note to him this way.
Ric
------------------------------
From: IVANOVIC (13352)
Subject: RE: Report on TML Pascal 2.0 (Re: Msg 13344)
Date: 2-OCT-21:38: Programming
I realize the amount of work he's done alone. The cynic in me says
that to compete effectively in the marketplace, he might have to learn
to leaverage his knowledge by hiring and working with others. I
bought TML Pascal v1.0 because it was the first Pascal for the
Macintosh. He's now got powerful competition in MPW Pascal and
Lightspeed Pascal. I've added my thoughts on how he can retain me as
a user. I will ultimately use the product that best meets my needs.
The actual support I was looking for was a very simple response (should be
automatic) like, "I've received your bug report," nothing more. I can usually
figure things out myself if the compiler emits assembly language (are you
listening LightspeedC people?)
I'm more adamant on the issue of standards. Pascal is well-established as both
a national and an international standard. The draft ISO version was available
well before the final version. There's really no excuse for not being up to
snuff. I don't want to have to re-write an entire application simply because
TML Pascal can't compile ANSI (not even ISO!) standard Pascal. Portability is
important to me - I don't want to re-invent the wheel. I have enough other
troubles as it is.
-- Vladimir
------------------------------
From: PEABO (13373)
Subject: RE: Report on TML Pascal 2.0 (Re: Msg 13327)
Date: 2-OCT-23:55: Programming
Another reason why it's nice to be able to shut off optimization is
that code generators occasionally make gruesome errors while
optimizing that they wouldn't make when not optimizing. It's really
nice to be able to shut off the optimizer and try running the program
again.
peter
------------------------------
From: MACINTOUCH (13346)
Subject: Line Conditioners/AC power
Date: 2-OCT-17:34: Hardware & Peripherals
You know how Apple is always saying the LaserWriter is their
most powerful computer? Yeah, but it's not because of the 1.5MB
of RAM, it's because it sucks up 800 WATTS! This was dimming
the lights in our new, old office, and giving us fits about
HyperDrive longevity. A UPS (uninterruptible power supply) looked
like a good idea, but a good one seemed to cost about $1K. We
compromised and got a 300Watt "line conditioner." I thought
people might like to know how it worked out. Basically, it
cut back the power drops to the point where you can just barely
notice a change in light intensity when the LaserWriter cycles --
before, it was like flashes of lightning (almost). The one we
got without shopping around was $200. (Rush, rush). One anomaly
is that now the light dims momentarily when the Mac is switched
on! Anyone got an explanation for this? (Any power engineers
in the SIG?)
Ric Ford
"MacInTouch"
------------------------------
From: MACINTOUCH (13360)
Subject: RE: Line Conditioners/AC power (Re: Msg 13346)
Date: 2-OCT-22:10: Hardware & Peripherals
Following up on a couple of things. The device makes a hum that's mildly
annoying. It also get so hot you can't touch it. Anybody know of good UPS's
well under $1K that produce sine-wave output, are well-built, and have fast
enough switching times?
Ric
------------------------------
From: MOUSEKETEER (13362)
Subject: RE: Line Conditioners/AC power (Re: Msg 13360)
Date: 2-OCT-22:49: Hardware & Peripherals
Hi Ric,
Ya, UPS's are hefty in price. There is a smaller company somewhere making
them up at around $750, but if you dig into the specs, you aren't getting
much at that price in terms of switching times, etc. A lot of that $1000
goes into liability insurance...just in case that UPS running the Acme
Oxygen Pump cuts out sometime.
I'd suggest you give Frank Stifter at Electronic Specialists Inc. a call...
he's right up there in Natick, MA near you at 617-655-1532. While they
do market a very complete line of UPS and such gear, Frank is great at talking
to you about just what problems are going on, and the best way to handle them.
It's the heater unit in the Laser sucking all the power, and the way it cuts
in and out at various times doesn't help matters. I'd be just as concerned
about what those wires in the wall are doing while the Laser is hooked up...
it can get pretty toasty. Maybe a wall run of plastic conduit and some 12 ga.
directly to the Laser from the junction box would be a safer (and considerably
less expensive) solution (assuming you can bypass local regulations and do the
work yourself...grin).
I'm more into overkill. I have 8 separate ciruits in a room 12x17, each set
for 30 amps. Some people say that 56 outlets is too many in a single room,
but....
Alf
------------------------------
From: DWB (13389)
Subject: RE: Line Conditioners/AC power (Re: Msg 13346)
Date: 3-OCT-01:24: Hardware & Peripherals
About two months ago I went out and bought I UPS at a local discount
house (Frye's Electronics for those of you that care) I am quite happy
with it. I did have to get a DataFrame though instead of the old
clunker I was using (shame that :-) because the clunker died when the
power went out. I've now been using my mac a couple of times when we
had a power outage and everything kept right on chugging, just like
it's supposed to. Even when the power company pulled what is a common
trick around here of taking the power out, waiting 30 seconds, and
turning it back on for a half second or so.
For the number and review freaks amongst you it's a Relax Technologies BPS 1.
Works like a champ.
David
------------------------------
From: DWB (13391)
Subject: RE: Line Conditioners/AC power (Re: Msg 13389)
Date: 3-OCT-01:31: Hardware & Peripherals
Oh yeah, it cost under $300.
David
------------------------------
From: NAKMAN (13400)
Subject: RE: Line Conditioners/AC power (Re: Msg 13360)
Date: 3-OCT-03:47: Hardware & Peripherals
If you need the UPS just for the Mac, check out the article by Jay James in the
Fall 1985 BMUG NL. It describes how to build an uninterruptible power supply
for less than $100. It puts out square waves, and can't give much power, and
simply involves a battery charger, a car battery, and an inverter, but, hey, it
does the job!
-- Raines / Team BMUG
------------------------------
From: JEFFS (13407)
Subject: RE: Line Conditioners/AC power (Re: Msg 13402)
Date: 3-OCT-18:28: Hardware & Peripherals
I have a Topaz Powermaker UPS. I have the 400VA model which lasts 35
minutes under full load and 75 under half. It puts out nice sine wave
60 Hz current for my Mac and all peripherals. It has a 4 millisecond
(typical with 10 mill max) switching time. It set me back $800 when I
bought it.
Jeff
------------------------------
From: ROWLAND (13349)
Subject: MonsterMac/Switcher ?
Date: 2-OCT-20:39: Business Mac
I suddenly started having a problem that puzzles me : switcher stopped working
on my MonsterMac(with 128K ROM). It starts up, draws the box that is to contain
the applications, but doesn't fill anything in. It then hangs - the mouse works
but nothing else does; I have to reboot.
Details: Switcher 5.0b3 , Sys/Finder 3.2 : I've replaced both the system file
and switcher from virgin copies so I think they are OK. Switcher does work when
I use that copy (and system - but with HD20 file) on a 64K Rom normal Mac. I
have used successfully exactly the system that fails (though of course not
identical - there was a period of a month or so that I didn't use switcher, and
did the usual changes to things).
If there are any suggestions or other experiences, I'd be glad to hear them. (
I'm anxiously waiting for Servant, but till then ..).
Thanks Mike Burns
------------------------------
From: HSTARR (13405)
Subject: RE: MonsterMac/Switcher ? (Re: Msg 13349)
Date: 3-OCT-18:12: Business Mac
That behaviour is normally Appletalk opening -- do yuou have Appletalk
on in the Control panel?? If you have, and you have anything but an
Appletalk pod connected to the printer port -- it is usually Bye,Bye!!
------------------------------
From: RONB (13366)
Subject: floating pt boards
Date: 2-OCT-23:21: Hardware & Peripherals
Does anyone have any experience/information about floating point add-in boards?
I'm particularly interested in speeding up fortran code. (I'm aware of the
recent Macworld article). Thanks in advance...
Ron.
------------------------------
From: MACINTOUCH (13367)
Subject: another 800K dies
Date: 2-OCT-23:21: Hardware & Peripherals
Another 800K internal drive just croaked without warning. It won't read or
initialize any disks successfully. On the _other_ Mac, today, a disk got stuck
in the internal 800K drive. No extra labels, no obvious problems, but it
wouldn't come out except with a paper-clip boost (gives you that cold sweat
feeling ...). It may have been a C Itoh disk, as it was dark red.
Ric
------------------------------
From: PEABO (13378)
Subject: RE: User Interface Problem: Macintosh Alert Boxes
Date: 3-OCT-00:30: Network Digests
>Date: Wed, 1 Oct 86 13:42:52 PDT
>From: <DAVEG@SLACVM.BITNET>
>Reply-to: DAVEG%SLACVM.BITNET@Forsythe.Stanford.Edu
>Subject: User Interface Problem: Macintosh Alert Boxes
>I've had a gripe about the Mac user interface for quite a while. The specific
>problem is alert boxes which only have an OK button. These boxes are typically
>used to warn the user that something has gone wrong during the processing
>of some request by the user. A good example is the Finder. You drag
>6 folders from one disk to another and go off to do something else while
>the Mac is busy doing the copying. You return 1/2 hour later assuming that
>the files have been copied, only to find that there is an ALERT box on
>the screen which says " The file DAVEG.GOODIES cannot be written" or something
>to that effect.
Not only that, but suppose you are sitting there watching the Finder copy a
selection of files, and you get several of these alerts (this has happened to
me on more than one occasion). What are you supposed to do, write down the
names of the files??? (1) I got a computer so I wouldn't have to use paper
and pencil anymore (ha!) and (2) under HFS, writing down file names can be
pretty tedious. It seems to me that what is needed at least in this specific
case is a notion of a task composed of a number of subsidiary tasks, which
can be completed independently, and for which at the end there is some
summary of results, such as a list of files that didn't make it. You might
ask for example to select all files that appear in the list, or just scroll
through them. Another gripe of mine is what happens when I select a random
subset of files in a folder to copy somewhere. After the Finder is done, the
files I copied are no longer selected, so I have to go though and select
them again, should I want to throw them in the trash for example. Hope I don't
make a mistake!
I'm not sure what the Finder should do in this case (I don't remember either,
which may be because my idea of the right thing is to keep the files selected
for another operation). Perhaps what it does now is dictated by some other
user interface principle that doesn't come to mind at the moment. It may boil
down to a limitation in understanding what the intent of the user is when
performing a sequence of operations on a bunch of items. The more assumptions
a program makes about what the user probably intends, the more the program is
suffering from modal operation.
peter
------------------------------
From: BRECHER (13393)
Subject: Re: Misc. programming problems
Date: 3-OCT-01:53: MUGS Online
To: Mark Nodine <mnodine@labs-b.bbn.com>
Subject: Misc. programming problems
> ... if you call SetVol with the number returned as the volume reference
> number by SFGetFile, it returns an error that there is no such volume.
This should work. Perhaps you are not clearing the ioNamePtr field of
the parameter block? ioNamePtr takes precedence over ioVRefNum, so if
you have garbage in ioNamePtr it will fail. If you're using the high
level SetVol rather than PBSetVol, you should pass Nil for the
StringPtr argument (do not pass a pointer to a zero-length string).
------------------------------
End of Delphi Mac Digest
************************
-------