Conference hydra::amiga_v1

    I have my original 1000 in shop 4+ years and still going strong. I see 
    more people with 500's having drive problems than 2000 owners. They are 
    supposed to use the same drive. Your friends drive most likely needs to 
    be aligned. I think that many of the copy protection schemes put the 
    floppies through a lot of punishment. Many say no, but the sounds I
    hear when booting some games sure make me think that something is going
    on that surely is not good for the drives. If you are worried I'd buy a
    drive alignment package, before I bought a spare drive,and align the
    drive before it has severe problems. I do not advocate the disk
    cleaners I see on the market. i'd wipe mine with a pad,and disk
    cleaning fluid, but I won't stick what looks a kitchen scrubber into
    one of my drives. Another thing. If you do not have an external drive
    get one. I think the single drive systems see a lot of abuse from the
    need to constantly change disks, and if you can eliminatr some of this
    activity you are better off.
    								bill

    Greg,
    	Quality seems to be a problem with all parts of the A500,
    though things have improved over the last year or so. From what	
    I understand from your mail, your friend is still having problems
    with loading software. If so, I would like to pass on the following
    warning. Read/write errors are not necessarily caused by a faulty drive,
    these can also be caused by faulty RAM, so if he does buy a replacement
    drive he could be wasting his money. 

My A1010 external floppy died also (unreadable disk) after 2 years of normal
use and no booting from it. I tested it on a friend's A1000 and it gives the
same error. It's true also that some disk operations give you a strange 
sentiment because of the 'sinister' sound of some head moves.
I've asked to an Ami dealer about repairing but he told me that it may be better
to buy the new type of unit (cheaper and more silent). So I'm trying to decide
(I would like to wait for an eventual high density unit ... but will it exist?).

Sorin (who feels that Amiga's weakest part are the floppies)

    The internal drive on my two year old A500 has also been rapidly going
    downhill. I've reached the point where everything goes fine until I
    write to a floppy and then that floppy generates READ/WRITE ERRORs some
    time later (but never on the file I just wrote to).
    
    I'm about to arrange a hospital visit for my amy, my local distributer
    offers a very good repair service (ME + SCREWDRIVER = BOOM! ) but
    before I do that, are there any diagnostic software tools out there
    that might help me run a health-check on my beloved (sick) Amiga?
    

    Re: .3

�(I would like to wait for an eventual high density unit ... but will it exist?).

    Applied Engineering is advertising a high-density/"standard"-density
    (1.52Mb/880K) external drive for the Amiga listing for $239. According
    to the ad, this drive has an electronic disk eject; this should be the
    norm for *all* floppy drives.

    Pete

    Re: .4

>    The internal drive on my two year old A500 has also been rapidly going
>    downhill. I've reached the point where everything goes fine until I
>    write to a floppy and then that floppy generates READ/WRITE ERRORs some
>    time later (but never on the file I just wrote to).
    
>    I'm about to arrange a hospital visit for my amy, my local distributer
>    offers a very good repair service (ME + SCREWDRIVER = BOOM! ) but
>    before I do that, are there any diagnostic software tools out there
>    that might help me run a health-check on my beloved (sick) Amiga?


    For memory diagnostics, there's MemDiag on FF214. As far as floppy testing
    goes, there are some commercial utilities available. There's a review of 
    Dr. Ami in the November AmigaWorld, though it wasn't exactly a raving 
    review. 

    -Dom

Re. .5

Thank you, that's good news. The electronic eject is really a necessity also
because accidental disk eject before write operation ending may very easily
happen on Amigas. I'll check for this high density unit.

Sorin 

    I just saw the Applied Engineering drive on for $199.xx in AmigaWorld. 
    I agree the electronic eject is a plus, but which CLI command does
    DISMOUNT/UNLOAD ?  Should we follow the MacIntosh norm of dumping the
    disk icon in the wastebasket?

Re..8

They should add a Dismount command in CLI and the equivalent option in the
Workbench menu. For professional applications it's worth to prevent disk
corruption induced so easily by premature disk eject.
I wonder if the floppy unit suppliers (with electronic eject for Ami) have
not provided an adequate command.

Sorin

    And while we're adding items to the wish-list, how about something
    along the lines of VMS's SHUTDOWN.COM? Dismount all volumes, stop all
    processes, etc. I'd feel a whole better knowing that some background
    task wasn't in the middle of attempting disk access while I'm pulling
    the plug...

    Paranoid-Pete

    re: the last few...
    
    picky, picky, picky :-)
    
    MS-DOS doesn't use any of that stuff, and look at how popular it is.
    
    If I remember the argument on usenet correctly, DISMOUNT is not
    possible because there is no resource tracking, you could only DISABLE
    the drive so that it couldn't be used, but it's resources would still
    there.  One way around that would be to add a "dismount" command to
    the device driver that tells it to free up it's resources and exit.
    
    If you added that functionality to all applications, handlers, device
    drivers, etc., then implementing a shutdown command would be easy.
    
    Dave

Re: .11

I don't think that the DISMOUNT being discussed is the one to cause
the device driver to terminate and remove itself from memory.  Instead,
the DISMOUNT command being discussed simply caused a disk eject.

DISMOUNT Df0: would cause df0: to eject its disk, but the system
device driver would remain loaded so that the next time you inserted
a floppy in df0: the you could use df0:!

    re:.12
    
    Sorry for the confusion, I was more replying to the idea of an orderly
    shutdown, when you have no control over writes to the disk drives
    unless you can DISABLE or DISMOUNT or WRITE PROTECT it!
    
    I assume AE has some sort of EJECT command that comes with their
    drive to do the equivalent of pushing the eject button on the drive.
    But doesn't that run into the same problem that manually pushing
    the button does - on a multitasking system, any program could be
    writing to the drive, so your EJECT command could electronically
    eject the disk in the middle of a bunch of write commands.
    
    Dave

	Since I am looking for a second drive I looked at the AE drive very
    seriously.

	The pictures in the ads show what looks like a standard eject button
    that could just as well serve as an eject requestor as something you have
    to mouse around with on the screen or as a CLI command.  KISS.

	I called AE up and asked about booting from a HD diskette.  They said
    that you could only boot from the internal drive.  TRUE?  (If so, it's a
    pain in the ...)

    Why be paranoid about system shutdown?  What specific background task
    will be attempting disk access at shutdown time that you don't have
    direct control over?
    
    A shutdown procedure makes sense for VMS and UNIX systems because there
    are many processes that aren't directly under the control of the user. 
    Stuff like mail, network communications, audits, virtual memory
    operations, etc.  But what would you run on an Amiga that you couldn't
    turn off?  I can think of "cron," which reads a disk every minute or
    two and executes programs on a weekly schedule, but I think the cron on
    my system can be turned off by a command like "cron -d(isable)."
    
    IMO, a standalone Amiga system is not sophisticated enough to get you
    into trouble as Pete Capriccio worries in .10.
    
    Wes

re .14

1.3 can't boot from an external drive. 2.0 can, as I found out by accident.

As for a shutdown script, I agree with Wes. The Amiga is a single-user machine,
and nothing shipped on the machine will touch the drive unless the user
tells it to. So, unless you installed something that touches the drive
when it wants to, the standard shutdown script for an Amiga would be:

    echo "Ok to shut the system down."

Of course, if you have Cron, or USEnet, or some such, then you may have
problems. But you added those, so can arrange to turn them off in s:shutdown
yourself.

      <mike

    re:.16
    
    How does 2.0 let you pick which floppy drive to boot from?  I assume
    that is a warmboot since I haven't seen anything about nonvolitile ram
    like ibmpc's have in the Amiga to handle a power-on boot.  Unless there 
    is some space in the clock chip?
    
    re:system shutdown
    three fingers works fine for me.  But for the novice user, a CBM
    blessed way to gracefully shutdown the system would help.  Ask
    yourself, "self, after a long session of running a bunch of wonderfully
    multitasking programs that in turn run whatever they need to, how do
    I know when it's safe to power off the system?"
    
    CBM came up with the expansion directory to make installing new
    hardware drivers easier, and has a nice ascii startup-sequence to 
    make it easy to customize the startup.  But there is nothing for
    a "blessed" way to notify all programs that the user would like
    to shutdown the system.  Not even a "s:shutdown-sequence" that
    makes an interesting noise.
    
    A standard way, maybe a "-q" or even an Arexx "bye", so that a 
    shutdown program could grab a list of what's running and send them
    all a shutdown command would be lot better than trying to remember
    what each program needs.
    
    Dave
    

re: shutdown

I believe wb 2.0 has a "close workbench" menu  item. 

While this isn't really a "shutdown" procedure, it does have potential to kill
off tasks in an orderly fashion (?)


FWIW I wrote a reboot program that looks like this:

	reboot -r		wait 5 seconds and reboot amiga
	reboot -c xxx		if avail chip memory is less than xxx,
				wait 5 seconds and reboot

The reboot assembly code was adopted from an AmigaMAIL article by 
Bryce Nesbit (sp?) and does some rather creative 68K hacking :-)

I wrote this mainly as a way to automatically reboot my amiga the just after
power up. At power on (Due to "features" of the Imtronics 68030 board which I
won't go into here) all my gvp hard disk partition handlers loaded into
chip memory. So I have a line like this in my startup-sequence:

	reboot -c 700000

Which reboots my amiga if the available chip memory is less than 700K
(ie hd drivers loaded into chip).

The boot devices are prioritized, and then checked in order for bootable
disks. To boot from the HD, you have no bootable floppies. To boot from
the external floppy, you make sure the internal floppy doesn't have a bootable
disk in it.

Re: shutdown

Note that shutting down the workbench in 2.0 won't shut down any processes;
it just makes the user go close all non-wb windows before it'll continue.
Process without an open window won't be affected at all.

The CBM manuals state that it's safe to shut the system down if the drive
lights aren't blinking. That should re-assure the users, but it isn't true.

Re: startups, etc.

For those of you who haven't seen 2.0, user control of startup software is
much nicer. First, there's a drawer called WBSTARTUP. All icons in that
drawer are launched when the workbench starts up. This is where the commodities
exchange toys (autopoint, screen blanker, nocapslock, leftymouse, etc)
go. You pick up the icon and drop it in the drawer, and it runs when you
start the system. Very nice.

Secondly, the startup-sequence now executes user-startup if it exists. You can
put stuff that won't work right via wbstartup there. Between these two, even
my very hacked enviroment (wshell/fcomp, a console window at startup, etc)
manages to use the startup-sequence provided by CBM.

And this leads up to the problem with shutdowns - CBM doesn't expect
applications to need to be told to shut down, so there's no way for the
system to do that. As a result, there are few (if any) applications that
provide some hook for shutting down, other than the user telling them to
stop.

How does the idea of a wbshutdown drawer that works the opposite of
the wbstartup drawer sound? You copy the shutdown icons into it, and when
you select "Shutdown" from the workbench menu, it runs them all and waits
for them to complete?

	<mike

> The CBM manuals state that it's safe to shut the system down if the drive
> lights aren't blinking. That should re-assure the users, but it isn't true.

    This kinda hits the nail on the head; even though there's no LED
    activity, I can hear hard drive activity at random times (seek
    calibration?). Who's to say that I might power-off during some disk
    activity and possibly cause damage? Okay, this is true paranoia, since
    the power surprise should have enough gumption to let the drive finish
    its current task and park the heads. Speaking of "park", I suppose the
    CBM supplied PARK could be used, but does this still apply to the
    modern drives, especially when several partitions are mounted?

    Re: back a few

    As far as the Amiga falling into the category of single-user machines;
    this IMHO is most certainly not true. While not a true multi-user
    machine, its multitasking capabilities make it a virtual multi-user
    system. For example, with enough resources available, its possible to
    have many screens, CLI's, tasks, etc. active at one time. Say I am
    in DPaint while de-arcing in one CLI, EDiting a file in another and 
    downloading a file in a terminal emulator (not to mention several
    background tasks running). Then the download finishes, I switch
    screens, and logoff the host system and then (since it's three in the
    morning) inadvertantly power-down before cleanly exiting/aborting.
    Now no software is going help the idiotic and feeble-minded souls like
    myself, but it would be nice to be able to double-click on, say, a
    shutdown icon that would list the current tasks, or maybe send a
    "shutdown request" to tasks so equipped, so they could post requestors
    saying "are you sure?" and the like, making sure all open files are
    closed, changes are saved, drives are spun-down (maybe?), etc.

> And this leads up to the problem with shutdowns - CBM doesn't expect
> applications to need to be told to shut down, so there's no way for the
> system to do that. As a result, there are few (if any) applications that
> provide some hook for shutting down, other than the user telling them to
> stop.

    Granted this isn't in existance today, but as I said before, it was
    merely a wish-list wish. As things get more complex and perhaps VM
    enters the OS, I would say a clean shutdown would be a plus.

    Pete (still pleasantly paranoid)

Amiga is multi=user, because WorkBench 1.3 has a console interface CLI to the
serial device.  Someone could be doing VT100 type interfacing out the back
while someone else was doing windows on the front, but you'd need a terminal.
I don't mean the crash debugger ROMWACK, I mean a CLI.
this means you could sign on your Amiga from work, if you have a way of
dailing out.  I don't know how.
Besides, multi-user machines are not what people use any more, while 
they are actual doing 
work.  The mini's are all going to be servers.
And the supers are going, too.  Lawrence Livermore Labs is abandoning the Cray
and giving everyone Suns, because it comes out to more raw power.
hundreds of Suns are more powerful than one Cray.  If 200 people are 
multitasking via a VAX server to the Cray, it is lower perforamnce seen by
one person than if thye  have a Sun at their desk.
Tom

    This kinda hits the nail on the head; even though there's no LED
    activity, I can hear hard drive activity at random times

In that case, there's something wrong with your configuration. That shouldn't
happen.

�    Then the download finishes, I switch
    screens, and logoff the host system and then (since it's three in the
    morning) inadvertantly power-down before cleanly exiting/aborting.

That's a _good_ reason for having a shutdown: to prevent you from doing
something stupid. Have you sent that request to CBM?

Failing that, what features would you like in a shutdown program/icon?

	<mike

I think I'd be satisfied if commodore just supported something like Xoper.
The tool should be able to show what is running and issue BREAK commands
to specific or all tasks.  STATUS just doesn't cut it since it only shows
certain tasks which were started from a cli.  Of course between STATUS and
the new close workbench 2.0 menu item you can probably track everything
important.

Anyone know how commodities exchange shuts down commodities?  Something
like CX (say applications exchange??) would be more user friendly than
looking at exec lists.

-steve

Unfortunetly Xoper is beyond the average user's ability to understand internal
process names etc. 

It would be easy to write a program that was run as the first command in the
startup-sequence which collected a list of the tasks that are running
right at boot time.

Then when you wanted to shutdown, it would look at the current list of running
tasks and request that you (or it) kill off anything that wasn't on the original 
list.

john

That's why I mentioned CX.  For anyone who hasn't used 2.0, this program
monitors the operation of properly written input handlers.  It may be more
general than that, if someone has seen the dev notes and they are allowed
maybe they can give us a better explanation.  In fact, they should probably
start a new topic for this.

The reason I brought it up is that you can go into CX and terminate one of
these programs that are CX "compliant".  I don't know if CX is generic
enough to support real applications.

-steve 

Most of the paranoia about shutting down the system I suspect is due
to the bitmap writeback.

AmigaDOS when writing to a drive waits about 5 seconds after all disk
I/O completes before it updates the block allocation bitmap on the
drive.  (It does this to improve system performance.)  I believe
it's this activity that causes people to suspect that AmigaDOS does
I/O behind their back.  (By the way, it is the only I/O AmigaDOS
itself does behind your back.)

When hard disks rattle when the disk I/O light is not on, the disk
is doing an internal calibration.  There's no danger in turning
off the machine while this is occurring.  In fact, there's no danger
in turning off the machine while there's a read in progress.  Power
down during a write is what kill you.

I've long felt that the Amiga, by not requiring a shutdown procedure,
had gotten things right.  I/O under the Amiga is very predictable,
and unless you ran something that's going to do I/O, nothing
is going to do I/O.   (Contrast this with paging systems or
write-back disk caches.)

In a consumer machine, whether a home computer or an office personal
computer, the interface should be kept stupid.  The way you turn an
appliance off is you push the off button.  Computers for the rest
of us should be designed the same way: because whether proper or
not, that is frequently how they are going to be used.  I've seen
people cycle the power on UNIX workstations because they didn't
know any other way of getting them to shutdown and reboot.

Personally, I think that if a personal computer needs a shutdown
procedure to be performed before being shut off, the system's designers
need to make the off switch send send an interrupt to the system,
rather than cut the power.

I agree with Randy that powering a system off should be as safe as possible.
That doesn't mean that it's safe to power off the system whenever you want
to. Even the Mac (king of the "make it obvious" machines) has a shutdown
sequence, probably for just that reason.

CX is overkill. It's designed for things that want to play with the input
stream. Having something in it just so you can shut yourself down seems
like overkill. Since we're going to change the system, I'd suggest another
IDCMP message "SYS_SHUT" that programs can request if they want to know about
a pending shutdown.

On the other hand, there are only two cases where you have problems. 1)
the bitmap writeback hasn't happened yet; 2) Some process has a file open
for write (I've seen that cause problems; I'm not sure exactly what it is).
A program that wades through the Exec lists ought to be possible; if
it finds any it trys to find a string to send to the user. If it doesn't,
it waits five seconds, then does shuts down/tells the user it's safe/whatever.

	<mike

    I like that suggestion of doing the same thing as the 2.0 startup
    for shutdown.  Since they have to learn how to do that to play with
    the startup, there isn't anything new to learn about how to play
    with the shutdown.  And it creates a flexible standard way for applications
    to do whatever they need to - send an Arexx message, IDCMP message,
    command line -q, etc. from the Workbench.
    
    Gee, that was easy.  Is it "already in there for WB3.0"? :-)
    
    Dave
              

Re: .27

>I agree with Randy that powering a system off should be as safe as possible.
>That doesn't mean that it's safe to power off the system whenever you want
>to. Even the Mac (king of the "make it obvious" machines) has a shutdown
>sequence, probably for just that reason.

Actually, I was thinking of the Mac as I wrote my note.  I believe the
Mac got it wrong: I believe the Mac would have been far more elegant
if the "off" switch, not a menu command, caused the system to shutdown
gracefully.

Supposedly, the reboot key sequence on a Amiga 2000 (but not the 500)
does not do the hardware reset immediately.  Instead, it causes an
interrupt, and then causes the hardware reset after a short delay.
The reason for this feature (I don't know if it is actually used),
is that the interrupt is sent to the trackdisk device so that it
can gracefully shut down.

    Re:.13
    
>    I assume AE has some sort of EJECT command that comes with their
>    drive to do the equivalent of pushing the eject button on the drive.
>    But doesn't that run into the same problem that manually pushing
>    the button does - on a multitasking system, any program could be
>    writing to the drive, so your EJECT command could electronically
>    eject the disk in the middle of a bunch of write commands.
    
    Rereading the AE ad in this month's AmigaWorld indicates that you
    *do* push the physical eject button on the drive - the internal logic
    prevents the floppy from ejecting until any ongoing write operation
    is complete. So it doesn't look like AE supplies any s/w to do the
    equivalent of ejecting a Mac floppy. At least that's my interpretation
    of their ad.

    -Dom 

    
    I believe someone from CBM said on usenet that in 2.0 they've
    tightened up this 3-5sec delay before the additional access
    occurs.  It sure seems that way on my A3000, but then maybe the
    system's just faster :-).
    
    -steve

    Re: .30

>    Rereading the AE ad in this month's AmigaWorld indicates that you
>    *do* push the physical eject button on the drive - the internal logic
>    prevents the floppy from ejecting until any ongoing write operation
>    is complete. So it doesn't look like AE supplies any s/w to do the
>    equivalent of ejecting a Mac floppy. At least that's my interpretation
>    of their ad.

    According to the review in November's Amazing Computing, the push-button
    is just a switch; there's no connection via the traditional mechanical
    linkage to the eject mechanism. Instead, an ejection motor spits the
    disk out only after the spindle motor stops. Also, the drive is
    basically a redesigned Mac drive, since AE has been making Apple
    accessories for a while and fairly new to the Amiga market. They supply
    a modified trackdisk.device which supposedly has the software support
    to auto-eject disks under software control, ala Mac. At the time of the
    article, the author had some problems using Quarterback and CrossDOS
    when in HD mode. Quarterback V4.2 (what happened to V4.1?) fixes the
    problem of not shutting off the spindle motor and according to an
    editor's note, the new driver shipping with current drives fixes both
    the QB and CrossDOS problems. However, they didn't say if the fix
    supports 1.44Mb MS-DOG HD format for CrossDOS or if it works with WB
    V2.0.

>================================================================================
>Note 3.306                        Amiga Survey                        306 of 306
>DPDMAI::HAGEN                                        18 lines   1-NOV-1990 17:35
>                           -< This is what I have. >-
>--------------------------------------------------------------------------------
>
>    AE HD 3.5" floppy drive

    Dom, would you care to give us your own personal review?

    Re: Shutdown stuff...

    Sorry to cause this note to get off track. For once I'm afraid I have
    to disagree with Randy; I'm not saying that CBM has got it wrong, and
    it seems to me that pulling the plug on any ordinary home computer is
    okay, but the Amiga is *not* your ordinary home computer! In fact, I'm
    truly amazed at how such a sophisticated machine is so easy to operate.
    Them poor low-end PC peeple who brag about how inexpensive their
    machines are become true disbelievers when I tell them I can be
    downloading in one window, formatting a disk in another, while at the
    same time working in DPaint, while they are having trouble getting
    their machines to print hardcopy in the background. It's this fact of
    the not-so-obvious complexity of the Amiga that makes me think of
    comparisons to shutdown procedures on multi-user operating systems.
    FWIW, I wasn't whining or complaining (hard to believe!), but just
    being paranoid as usual.

    Pete

    
    re .32
    
    > ... Quarterback 4.2 (what happenend to 4.1?) .......
    
    Quarterback 4.1 was not generally offered.  It was an interim version 
    you got if you called CCS.  It's purpose was to fix:
         - it's colors: it apparently clashed colors with WB 2.0 such that
           you couldn't read its text under 2.0.
         - restore: it added a switch option to the restore menus which
           allows you to restore "empty" subdirectories.  What's funny is
           that it introduced the bug of not being able to save empty
           subdirectories! (has been reported).
    
    Frank
    
     
     It also seems 
           to have lost the ability to save