Path: utzoo!utgpu!attcan!uunet!tank!uwvax!rutgers!apple!bloom-beacon!think!ephraim
From: ephraim@think.COM (Ephraim Vishniac)
Newsgroups: comp.sys.mac
Subject: Re: DiskTimer II results for verious disks wanted.
Message-ID: <30582@think.UUCP>
Date: 7 Nov 88 14:15:44 GMT
References: <30248@think.UUCP> <31@eplrx7.UUCP>
Sender: news@think.UUCP
Reply-To: ephraim@vidar.think.com.UUCP (Ephraim Vishniac)
Distribution: comp
Organization: Thinking Machines Corporation, Cambridge MA, USA
Lines: 65

In article <31@eplrx7.UUCP> lad@eplrx7.UUCP (lad) writes:
>Disktimer NEVER used the Device Manager, and it NEVER used the File
>Manager, something ALL drivers have to do to talk to the disk.

Read the source code.  DiskTimer uses Read and Write calls.  These are
executed by the Device Manager, which selects the appropriate driver
and calls it.

Also, read Inside Mac.  The File Manager uses the Device Manager uses
the device drivers which use the SCSI manager or hardware.  Claiming
that the device drivers call the File Manager only shows that you've
never read or written a device driver.

>>Just within the
>> past month, there were a series of articles here (comp.sys.mac)
>> discussing how someone mistakenly formatted his disk at 1:1 instead of
>> 2:1 and got dismal performance -- plainly shown by DiskTimer!

>I read those articles but they have nothing to do with what we're
>talking about here.  I'm talking about real tests done by real people
>on real disks, which, by the way, were formatted at 1:1, 2:1, and
>3:1.  Disktimer consistantly favored disks formatted 1:1 no matter
>which machine they're run on.

Could one of the people involved in that earlier discussion please
speak up?  Are you a real person?  Are you a zombie?  An android?  An
Alien Life Form?  Did you run the tests on real disks or on imaginary
ones? 

>What's nonsense is how you continue to praise Disktimer as a real
>performance benchmark.

I didn't, Larry.  Calm yourself, and reread what I wrote.  There's no
praise of DiskTimer there, just criticism of your criticism. 

>By the way, you forget too easily that Reeks did extensive testing
>and even published an article on Disktimer.  I'm going to track it
>down and post it to the net.  I remember you flaming Reeks about that
>article, using the same arguments you're using now.  Those arguments
>were wrong then, and they're wrong now.

Of course I remember Jim Reekes' bizarre tirade.  In it, he made such
amazing claims as "the Mac file system only does single-sector
transfers, so multi-sector transfers are irrelevant."

>Steve Brecher has even said many times before that Disktimer, the
>program *he* wrote, dosen't reflect real-world disk performance.
>It's even in the program's initial screen!!

I agree that DiskTimer's got problems.  But the only way we'll get a
better benchmark is through accurate, intelligent criticism.

>You stop the garbage, Ephraim.  It's seems you have some vested
>interest in seeing that Disktimer somehow succeeds.  I wish you luck.

DiskTimer has already "succeeded," unfortunately - it's the most
widely accepted benchmark of add-on storage for the Mac.  My interest
is in seeing that people understand its failings.  And as one of
Micah's owners you're perfectly disinterested, right?

Ephraim Vishniac					  ephraim@think.com
Thinking Machines Corporation / 245 First Street / Cambridge, MA 02142-1214

     On two occasions I have been asked, "Pray, Mr. Babbage, if you put
     into the machine wrong figures, will the right answers come out?"