Re: CIA old/new?

From: Jim Brain <brain_at_jbrain.com>
Date: Tue, 5 Jan 2021 11:25:27 -0600
Message-ID: <ea784f55-5856-f89c-cc7b-d7e4e756ead0_at_jbrain.com>
On 1/5/2021 4:43 AM, smf wrote:
>
> I don't know.
>
> The PLA, VIC2, TED & 7501 had a lot of manufacturing issues. The VIC2 
> had some design issues, for example the 64 clocks per line and 
> requirement of PLL to fix the dot crawl in the early days. Do you 
> think the 6581 filters were well done?
>
Maybe my point was less that other ICs had no issues (the 6522, for 
example had the lost interrupt issue, and the 6551 stopped transmitting 
immediately after seeing DTR low, even if in the middle of the byte) and 
more that the designs were based on solid engineering principles.  The 
"write and then check if it worked" aspect of the VDC is horrible.  It's 
like an IC design anti-pattern, quite honestly.  I can forgive the 
VIC-II design for having issues given there are 2 clock domains (that's 
tricky even today) and the race condition in the CIA (race conditions 
are always sneaky).  The filters in the SID were not the designer's 
fault.  He hated that they were not finished/well done, but the mgmt 
team gave him no more time to finish.  You could either have a SID is 
non linear filters or no SID at all, and I think we can all agree the 
non linear filter SID is a vast improvement over other audio options of 
the day.


> There is an argument that Bil Herd is responsible for the C128 faults 
> as anyone else would have let the product fail in the crucible of 
> fire. I get the impression he feels guilty as he keeps going on about 
> it ;-)
>
He does feel guilty, but I don't think he's responsible.  If folks think 
that they have a very shallow grasp of how things got decided on the 
C128.  The C128 fault of not hiding the extra VIC-IIe registers from 64 
mode is on Bil, sure, but the Z80 and the mess inside the design to 
accommodate it (and the "support all the cart, including the 
cantankerous Magic Voice" mandate) falls squarely on Marketing and 
Management, not Bil.


> I'm not saying that shipping something crappy is a bad strategy, it's 
> perfectly valid to build market share as long as you fix the issues 
> and you don't kill anyone in the meantime.
>
I'd even go so far as noting it's OK to ship something crappy if your 
target audience is OK with that and does not expect the issues to be 
fixed. It's all about target audience expectations.

Jim
Received on 2021-01-05 19:00:03

Archive generated by hypermail 2.3.0.