🚀 go-pugleaf

On Wed, 14 Jun 1995, John Woods wrote:

> e4315555@student.uq.edu.au (Peter Naylor) writes:
> >Well, the only experience I've had with SCSI on a PC was when I installed 
> >a Sony magneto-optical and Adaptec controller on one of the machines at 
> >work.  It certainly needed a "driver" type file, and it came with the 
> >controller, not built-in.
> 
> Magneto-optical drives probably require a special driver (as do CD-ROMs,
> by the way).  But traditional hard drives are handled by the BIOS.

The BIOS in question actually had quite some trouble allowing automatic 
mounting of the SCSI devices... we couldn't get it to overlook the SCSI 
controller and automatically boot from the EIDE drive.  Every reboot it'd 
say something to the effect of "device <blah> non-system disk, press F1 
to continue...".  Having pressed F1 it'd move on the the other devices, 
but do you think we could find any combination of dip switches on the 
SCSI controller or any setting in the BIOS to fix this?!  NO.  *SIGH*

> >Having to include mounting brackets is rather silly... my machine came 
> >with the brackets inside, so I don't have to worry about that sort of 
> >thing, and it doesn't incur any extra cost on the part of the drive seller.
> 
> Actually, they don't _have_ to include brackets (and the most recent Mac
> drive I bought didn't):  a running Mac already *has* a drive bracket, so
> it would be perfectly possible to pop out the drive (I wish to repeat my
> praise for the designers of the Mac drive bracket system), unscrew it,
> screw it onto the new drive, and pop the new drive in.  However, it saves
> the customer a good bit of time to have the drive already mounted, and the
> cost to the vendor is not a great deal.

This is all very well... but you missed one of the major points in what I 
said... note that I said "brackets".  I have several 3.5" form factor 
positions in my machine, and they all came with a universal bracket that 
fits almost all 3.5" devices.  I find that pretty handy... and sensible.  
What you're saying above is that the 30-45 seconds that it takes me to 
mount a device in a bracket (a job that I do about once a year on each 
machine) justifies including a bracket on every mechanism.  I think it's 
only good sense that the supplier you bought from didn't include a 
prefitted bracket.  In any case... it's really trivial, and we're way off 
the original track ;)

> >I guess the main question is... are there any good prepping and 
> >partitioning packages for the Mac that are freely distributable.
> 
> There are no really _good_ ones available free, in fact I'm not sure there
> are any. 

I think this is just another indication of the rather poor (in comparison 
to other platforms) availability of freely-distributable, quality 
utilities on the Mac.  I've been very disappointed in my net-leeching.  
Where is the community spirit?  The PC can rely on the fact that it has a 
huge user-base... but the Mac people need to get some friendly 
contribution going... and well, the Amiga and UNIX systems are something 
else again... huge volumes of very high quality software, FREE!

> There used to be one (I've forgotten who wrote it), but it went
> commercial.  (I don't think it was actively exterminated so much as just
> allowed to disappear; it wouldn't be compatible with recent System software,
> so there'd be no point now in archiving it...)

Why should anything as standard-driven as software which accesses a SCSI 
device and driver at a low level be made redundant by OS revisions?  It 
doesn't happen to anyone else...

> But the commercial ones are cheap, and (as I said) you get one "free" with
								 ^^^^^^
I'm glad to see you put this in quotes...

> every drive you purchase from a Mac vendor.  And once you have purchased a
> driver (either with a Mac drive, or independantly), you can use it on future
> drives regardless of source.

Not really a good solution though is it... and a bit of a limitation.  
Still, you could always just get a copy of the software from a friend who 
got it "free" with a drive from a Mac dealer.
 
> >I can't think of a single thing that Apple could change in 
> >the ROMs to improve the performance of a SCSI drive with their machines, 
> 
> They change absolutely nothing except identification.  (Well, on very early
> Macs they also changed the SCSI reset / UNIT ATTENTION behavior, due to their
> implementing a SCSI-2 driver before the spec was finished...)

That's what I suspected.  And obviously, if the spec wasn't finished they 
couldn't really implement SCSI-2 in a driver could they?  I suspect they 
simply have a few SCSI-2 type commands implemented for accessing devices 
such as CDROMs without missing out on all the control functions, just as 
my SCSI-1 controller does.  But then, it's all handled by the 
controller... and I have the option of turning them off by simply toggling 
a switch in a small control program, all handled by hardware mind you, no 
special "drivers" necessary.

> >there are only two things I can think of that would justify that strategy...
> >1.  To be able to charge whatever they feel like for a real "Apple" drive, 
> >which is of course "special" for "special" Mac users...
> 
> This is between 99 and 100% of their motivation, that's almost certain.  It
> is possible that they feel this is partially justified by their QA of drives:
> you can bet that when Apple starts shipping a drive, they have tested a fairly
> large number of units and are comfortable with their reliability and function.

As does almost every manufacturer of computer systems... with the 
exception of cheap cruddy PC clones.  I'm not really sure that I see how 
QA fits into this little equation... QA is basically the responsibilty of 
the mechanism manufacturer as I see it, and Apple has only to make sure 
they have some recourse if the quality drops from the expected supply 
level.  I can see that testing is very advantageous, but I don't see that 
Apple has any particularly good claim to being able to charge extra for 
it... that sort of thing is included in the price of a new system, and I 
don't see how it should apply to buying new peripherals for that system, 
given that Apple add basically no technical value to the product.

> (Check out comp.periphs.scsi for a report that a recent Quantum drive (Atlas?)
> doesn't work right on the Mac unless you have the latest microcode ROM!!)

The fact that it only arises on Macs is a little suspicious... but yes, 
there are always problems like this with a small percentage of devices... 
I believe that my controller has some difficulties with one or two CDROM 
drives, but that's solely the fault of the actual SCSI controller chip... 
which is used on controllers through a lot of industry.  This is the sort 
of thing you can pick up by simply hooking up various peripherals and 
testing all the functions... not an overly expensive procedure in the 
overall scheme of designing a system... and something that simply affects 
the brands that a company like Apple will supply as peripherals.

> However, even their QA department is fallible (witness the famous stiction
> problem of early drives, which to be fair bit _every_ computer vendor).

Of course they're fallible... they can't test a large number of drives 
for ALL functions over a length of time approaching the MTBF.  No-one 
can... that's why the drive manufacturer is in the best position to keep 
tabs on quality... and then make assurances to the buyer... that's what 
legal agreements are for, and warranties.

> >2.  To inhibit the use of some SCSI functions and lock the drive into 
> >certain modes, because of some inadequacies with the Apple SCSI controller 
> >when it comes to certain standard SCSI settings.
> 
> The only change they've ever made, as indicated above, was to make drives
> conform to what the SCSI-2 draft said about reset behavior, and they only 
> did that until they started shipping new Mac models with corrected drivers.

Oh, so to avoid simply providing sensible OS provision for device level 
access, they went to the extra cost of having custom ROMs fitted to all 
their drives.  Very clever... and where exactly was the advantage in 
having SCSI-2 provision for hard drives which were for 99% of the cases 
SCSI-1?  NONE of this justifies the extra cost which is invariably passed 
onto the customer, IMHO.

> >I'd be of the opinion that if a drive which adheres well to the SCSI standard 
> >doesn't work with a standard Mac, with the standard system software (ie, 
> >not costing any extra, PD is fine), then Apple has ripped you off when it 
> >comes to supplying a SCSI interface.
> 
> Apple certainly ought to consider adding value to their drives to justify
> their price, rather than a ridiculously ineffective effort to hold users
> hostage.  Even the simple believe that the drive is slightly more likely
> to be trouble-free is worth something to many customers.

I wouldn't believe anything Apple claimed... except if they released a 
statement saying that their controller hardware just doesn't work at all 
with "Brand X, Model Y" device.  The reliability is something I'd base on 
the manufacturer's specifications... and their warranty is my peace-of-mind.

> At least you don't have to worry about whether or not two different drive
> manufacturers' drives can be used on the same cable, like you used to (and
> still occasionally do) with IDE.

Actually, in a very small number of strange setups, there do exist SCSI 
problems which are similar.  But you're right IDE does suck pretty 
badly... that's why I have never and will never use it.  On my platform 
of choice there are few options for using IDE anyway... but at least the 
options exist, and I have complete control.

Long live SCSI!  (But preferably with no hidden limitations and no 
rip-off tactics)