[Date Prev][Date Next] [Chronological] [Thread] [Top]

RE: Memory Card Corruption Problems



That's great that we're going at this with resolve and cool heads.  And I
understand the physics.  But software can stress hardware.  Perhaps we've
pushed this design farther than it was intended to go.  Like running Windows
NT on a 286 AT computer.

I still need answers to my questions below.  I was hoping Tab or Guy might
be able to answer them.

Ian

-----Original Message-----
From: owner-rcr@dieboldes.com [mailto:owner-rcr@dieboldes.com]On Behalf Of Jeffrey
W. Dean
Sent: March 30, 2001 3:08 PM
To: 'rcr@dieboldes.com'
Subject: RE: Memory Card Corruption Problems


Thanks for gathering this info - it adds greatly to the mix.  Remember that
software, of and in itself, does not physically transfer electrical signals
from the bus to the card.  Software does in fact have a large influence on
data bus traffic and sequencing.  We're absolutely not going off half cocked
and ASSuming anything.  Any design changes or retrofits will be based upon a
rational analysis and a very high probability of a permanent solution.
That's our track.

Jeff

> -----Original Message-----
> From:	Ian S. Piper [SMTP:ian@dieboldes.com]
> Sent:	Friday, March 30, 2001 1:03 PM
> To:	rcr@dieboldes.com
> Subject:	RE: Memory Card Corruption Problems
>
> There's one discrepancy in this theory.  Minnesota has experienced no
> memory card corruptions using 1.95e firmware.
>
> I've spoken again with Cathi Smothers and Barry Herron inquiring about any
> memory card corruption events that may have occurred in Minnesota during
> their November elections.  They were the on-site Global support during
> those elections.  They still state that there were no memory card
> corruptions.  They are sure that if Washington, Dakota, or Anoka were to
> experience any problems like this, they would have been notified of it
> that day.  All of the 348 units in Minnesota are Visible Light.  With that
> many units, you would expect at least 1% (3 to 4 units) to experience this
> problem.
>
> Some accounts with 1.94w firmware have experienced memory card corruptions
> upwards of 10%. (e.g., Fresno in November 1999).  Most others have
> experienced an average of approximately 5% (this is a conservative
> estimate on my part).
>
> How is 1.95e different from 1.94w?  Is 1.95 code based on 1.94 structure?
> Or is 1.95 a complete re-write?
>
> Does it perform less multi-threading, and therefore have less data
> traffic?
>
> I'd like to have this discrepancy explained before we start retrofitting
> 15,000 AccuVote-OS units in the field.
>
> Ian
>
>
>
> -----Original Message-----
> From: owner-rcr@dieboldes.com [ <mailto:owner-rcr@dieboldes.com>]On Behalf Of
> Jeffrey
> W. Dean
> Sent: March 29, 2001 4:41 PM
> To: 'rcr@dieboldes.com'
> Subject: Memory Card Corruption Problems
>
>
> I've posted to this list to make sure everyone with an interest can see
> what's going on.  Please read the attachment - we will provide timely
> updates over the next few days as additional info becomes available.
>
> Jeff Dean
>
>  <<PA04.doc>>
>
>