From: adam@address.invalid
Moe Trin wrote:
> ]Subject: Statistics for alt.os.linux.mandriva
> ]Date: Fri, 11 May 2012 21:15:03 +0100
>
> ]Total posts considered: 18 over 7 days
>
> Sheesh - I'm gone for a week of travel, and the newsgroup dies?
Must be your fault, then. ;-)
I'm still running Mandriva 2010.0, at least until Mandriva's future
is more definite, so I'm reluctant to ask questions here about
things that may have been changed or fixed in newer releases.
> On Mon, 30 Apr 2012, in the Usenet newsgroup alt.os.linux.mandriva, in
article
> , Adam wrote:
>> May Day! May Day! Yes, Tuesday is May Day. I don't see why that by
>> itself should constitute a USAF emergency. :-)
>
> It's not USAF - it's ITU (International Telecommunications Union)
They also consider May 1st an emergency???
>> I wasn't clear originally -- this had nothing to do with the speed
>> of the code. It was the speed at which the CPU fetched opcodes (but
>> not data) from RAM. The early Model 4s gave the CPU 2 or 1
>> additional clock cycles during each opcode fetch.
>
> Long way around - it's actually a function of memory bus speed verse
> CPU speed.
I'd call it a consequence of the speed difference, or rather of the
hardware kludge to add wait states for memory access so the early
Model 4s would run reliably. I don't remember what changed so that
they were able to dispense with the wait states during the
production run. I know I already mentioned how that other guy's
Model 4 had an effective speed of 83% of mine, at least for my test
program.
> One way that shows up is if you're using a faster clock
> speed to make up for plodding code.
But then that would mean all the components would have to be rated
for the faster clock speed. That might also cause havoc, as the FDC
required a minimum delay (in usec) after giving it a command.
[digitizing]
>> I had a schematic of it years ago and it had maybe a half dozen
>> chips and not much more. For each channel, one chip to recognize
>> its port number to trigger a latch to hold the values being output,
>> and another chip with resistors as a primitive DAC.
>
> In 1985??? There were several eight bit DAC chips that were quite
> cheap by then - top of the head, the DAC08 from several vendors, or
> the Motorola MC1408
I believe the hardware was designed a few years before that, and
IIRC the resistors for each channel were all in one DIP package.
> Trying to
> drive the ladder directly from something like a CMOS latch is going to
> cost substantial accuracy
Everything was TTL. Given the power of a TRS-80, I think anything
more would have been overkill.
>> I just got a new one this afternoon, known (I found out afterward)
>> as the AMMYY scam. The caller tried to get me to run some commands
>> on my computer which would have allowed them remote access.
>
> news://comp.risks/ Risks Digest 26.81 (4 May 2012) item:
> "Canadians hit by bogus Microsoft Help calls" (Gene Wirchenko)
>
> Geez, how come you're so lucky.
This one gave some other company name, but were insistent on having
_whoever answered_ (I'd said "Adam" wasn't home) run some commands
on the computer. The second time they called, I mentioned the
do-not-call list and filing a complaint (which I did), and I think
they've stopped calling me.
> While I have three computers that
> have windoze decals on the exterior, none of them have it installed
> any more. I suppose I could dig up the backups if I had to.
I kept the OEM Windows (in a shrunken partition) and the recovery
partition on this box until the warranty expired, and the sticker's
still on it.
Meanwhile Bobbie here mentioned UEFI which I hadn't heard about, but
it looks like keeping both OEM Windows (until the warranty expires)
and Linux on my next computer is going to be a challenge. I know
that if there are any problems in warranty, they'll want me to run
some diagnostics that were included with their OEM Windows. That
happened with this box.
[property tax]
> HOLY HOPPY TOADS!
Is that any relation to "Jumping Jehoshaphat!"?
> the reason for the astronomical prices is simply
> "supply and demand". In the case of Sunnyvale, there aren't that many
> build-able lots left, and yet there are a lot of people in the Bay
> area so they gotta live somewhere, and there go the prices.
My cousin in Los Gatos wanted to buy the house they were renting and
it was on the market ($1M+), but they just couldn't afford it. My
cousin in San Jose does own his house, which I gather isn't common
there. Around here it's the norm.
> (At
> NASA, six of the guys I worked with had commutes over 45 miles, and
> one had a commute over 70 miles. Commute costs verses housing costs -
> you're screwed either way.)
That house in Los Gatos had extra rooms they rented out, and one of
them was basically just a bedroom for a guy who lived in Redding but
worked in Silicon Valley.
That also explains why "commuting distance" to NYC has expanded from
Westchester into Putnam and now southern Dutchess County (up to
Poughkeepsie, the last stop on the commuter trains), and why real
estate prices have gone up accordingly.
>> I could get into quite a discussion about large schools vs. small
>> schools but I'll just say that I'm glad that I graduated from a
>> small (enrollment ~2500) college. (In Arlington, NY, as opposed to
>> my two dismal years in ZIP 21218.)
>
> Never mind ZIP 48824 with 47000 students (or was that ZIP 48109 with
> 42000). I only had "big" schools - but that seems to be the way it
> goes. Even the community colleges (here and in the Bay area) are 5k+
48824 (msu.edu) is larger than 48109 (umich.edu). I spent almost
two semesters at 48109 as a grad student, although my home address
was 48104. When I wanted to impress companies, I'd give the
university for my address, as most people assumed that meant
faculty. They had trays for incoming mail for grad students, so I
obviously wasn't the only one doing it.
Adam
--
Registered Linux User #536473
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
