Is computer history taught now? - alt.folklore.computers

Message from discussion Is computer history taught now?

Frank McCoy

View profile

More options Mar 20 2007, 10:48 pm

Newsgroups: alt.folklore.computers

From: Frank McCoy <mcc...@millcomm.com>

Date: Tue, 20 Mar 2007 21:48:06 -0500

Local: Tues, Mar 20 2007 10:48 pm

Subject: Re: Is computer history taught now?

  Reply to author
  |
  Forward
  | Print
  | View thread
  | Show original
  | Report this message
  | Find messages by this author
  

In alt.folklore.computers Brian Inglis

<Brian.Ing...@SystematicSW.Invalid> wrote:
>fOn Tue, 20 Mar 2007 15:53:41 -0400 in alt.folklore.computers, krw
><k...@att.bizzzz> wrote:

>>In article <CL6dnVKzQ-aepZ3bnZ2dnUVZ8qKvn...@bt.com>,
>>am.swal...@btopenworld.com says...
>>> jmfbah...@aol.com wrote:
>>> > In article <MPG.2069a32ad465a2ff98a...@news.individual.net>,
>>> > krw <k...@att.bizzzz> wrote:
>>> >> In article <4sydnRnoSdN7tGLYRVny...@bt.com>,
>>> >> am.swal...@btopenworld.com says...

>>> >>> From its behaviour in 1980 IBM was still believed mainframes should
>>> >>> sell for a million.
>>> >> An ES9000 full of processors, channels, and memory went over $20M in
>>> >> the early '90s. There was no shortage of customers.

>>> > The poster doesn't seem to have any concept about the real
>>> > computing world out there.

>>> He does but the buyers may not have realised the price war had started.

>>...and you can't see that there is more than price at work.

>>> Having said that in the early 1980s microprocessors were only fighting
>>> minicomputers. Mainframes were safe until the chips with 32 bit *data*
>>> buses came along.

>>...and minicomputers were fighting mainframes.

>Loaded minis were getting up there near mainframe prices, and you often
>needed more than a single mini to handle the workload of any non-small
>company.

>Current enterprise class RISC servers and storage are in the same cost
>range as mainframes and provide similar capabilities. However (at least
>IBM) mainframe systems and peripherals are *designed* to have zero
>customer-visible hardware outages during normal customers' ownership
>timeframes (five years?), and the software attempts to attain that goal.
>Not aware if any other systems are designed to meet that goal.

>Remember Lynn's story about virtual device handling causing more than
>the expected number of device errors (IFCCs?) to be reported in a survey
>of large customers' systems, and the product manager followup to find
>the cause, resulting in Lynn changing his code to *not do that*?
>How many companies provide that level of customer service in such cases
>where each customer remains blissfully unaware of any issue?

When I worked for Seagate, they found out the competition was *hiding*
disk-errors; doing automatic retries and never reporting anything wrong
until pretty much completely unrecoverable. This, in spite of their
*specifications* saying that retries and recoveries were all being
reported so that overhead software could keep track of soft errors, and
if necessary avoid that spot on the disk. Only: the competitor's
disk-drives weren't as good as ours; so to *look* good they just didn't
report retries and soft errors.

So Marketing, instead of trumpeting this fact as a good reason to buy
*our* disks instead of the competition; because *we* got the same "error
rate" without lying, they decide that *we* have to do the exact same
lying-through-the-interface to remain competitive!!!!

Gaaahhhh!
Worse yet, they made it so you couldn't disable same through the
interface, so you had *no* idea if the disk was going to crash until it
did.

This sort of shit is what lead to S.M.A.R.T.; which in theory allowed
software on the drive to keep track of such soft errors (but never
allowing the customer to see it); and supposedly notifying said customer
when things start to go bad.

Trouble is: even more specsmanship of that took place; hiding even MORE
errors, until S.M.A.R.T. is pretty much completely useless as well;
never EVER reporting failures or incipient troubles until the drive is
already crashing or crashes. ;-{

This isn't just for Seagate drives (which IMO are still better than
most) but in *all* disk drives, no matter which company makes them.
;-{

--
_____
/ ' / ™
,-/-, __ __. ____ /_
(_/ / (_(_/|_/ / <_/ <_

Reply to author Forward