| Re: An Argument for Big-Endian: Packed Decimal [message #403077 is a reply to message #403075] |
Mon, 21 December 2020 01:01   |
|
|
Originally posted by: J. Clarke
On Sun, 20 Dec 2020 21:51:21 -0800 (PST), Quadibloc
<jsavard@ecn.ab.ca> wrote:
> On Sunday, December 20, 2020 at 5:36:00 PM UTC-7, Peter Flass wrote:
>> Jon Elson <el...@pico-systems.com> wrote:
>>> Peter Flass wrote:
>
>>>> I think you could link your programs so that part of them loaded into
>>>> processor storage and part into LCS.
>
>>> OK, that particular 360/50 had, I think 128 KB of internal storage and 1 MB
>>> of LCS. The OS and one small partition fit in the 128K, everything else
>>> crawled at 1/4 speed in the LCS.
>
>> Well, OK, that was a particularly stupid configuration. Core was expensive
>> in those days and some bean-counter probably only looked at the cost.
>
> Of course, using the computer cost enough per hour that the bean counter didn't
> quite do his job properly,. Unless the computer was used in such a way that 90% of
> what it did took place in that one small partition.
>
> Of course, today's computers, cheap enough to be owned by ordinary people,
> with gigabytes of RAM running at gigahertz speeds (well, sort of, as that's
> bandwidth, not latency, which is in the dozens of cycles)... would have seemed
> like far-future science fiction back then.
>
> With Dennard scaling dead, and Moore's law finally showing signs of slowing down,
> it may be that ten years from now, we will be using 128-core machines, but still stuck
> below 5 GHz. Or five years from now, at least.
I think we are slowly converging on an end-point configuration.
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403090 is a reply to message #402948] |
Mon, 21 December 2020 12:36 |
Jon Elson
Messages: 646
Registered: April 2013
Karma: 0
|
Senior Member |
|
|
Charlie Gibbs wrote:
> Those limits were somewhat arbitrary. IBM decreed that the
> 360/30 had a maximum memory of 64K, but I once used one with
> 128K. A third party added more memory and was able to extend
> the address bus as required. The extra switch and indicator
> added to the front panel had a bit of a homebrewed look, but
> the machine worked just fine. Apparently some outfits expanded
> model 30s to as much as 512K.
>
This one has always interested me. Since the /30 was an 8-bit architecture
(8 bit + parity memory and 8-bit ALU data paths), why would they have
implemented the microcode to do more than 16-bit address arithmetic? Just
so they could detect address overflows in RX instructions? Seems like it
could detect the overflow anyway, without actually processing the 3rd
address byte.
Somebody at IBM must have decided they would take the performance loss to do
the extra add cycle on every RX instruction because they planned that
someday they might make a machine with more than 16 bits of address.
Without this, it would have been VERY difficult to expand the address, as it
would require major alteration of the microcode. I'm pretty sure that the
/30 did not have a separate address adder to add up the base register, the
index register and the offset, it was all run through the main ALU.
Anybody know any details of this?
Jon
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403093 is a reply to message #403067] |
Mon, 21 December 2020 12:44 |
Jon Elson
Messages: 646
Registered: April 2013
Karma: 0
|
Senior Member |
|
|
Peter Flass wrote:
> MFT had fixed partitions. MVT used regions that could (AFAIK) be
> discontinuous.
>
Well, like a few other things in OS/360 development, they really didn't
understand some of the side effects. A HUGE problem in MVT was memory
fragmentation. Even without the regions, if you let ordinary users ask for
storage in arbitrary amounts, and especially if you charged for CPU minutes
times partition size, people would ask for the exact minimum needed to run
their program. If everybody was allocating partitions in 27K, 31K 54K etc.
chunks, quite soon memory would be balkanized into lots of small pieces with
unused spaces scattered between small, long-running jobs, and the machine
bogged down. The fix was to hack the storage allocation code so that no
matter WHAT you asked for, you were given a partition that was in some
integral number of quanta, at least as big as you asked for. That pretty
much solved the issue. So, you couldn't allow users to allocate random
chunks of storage, and then release them when done. This is the age-old
garbage collection issue.
Jon
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403094 is a reply to message #403068] |
Mon, 21 December 2020 12:53 |
Jon Elson
Messages: 646
Registered: April 2013
Karma: 0
|
Senior Member |
|
|
Peter Flass wrote:
> Jon Elson <elson@pico-systems.com> wrote:
>> Peter Flass wrote:
>>
>>
>>> I think you could link your programs so that part of them loaded into
>>> processor storage and part into LCS.
>>
>> OK, that particular 360/50 had, I think 128 KB of internal storage and 1
>> MB
>> of LCS. The OS and one small partition fit in the 128K, everything else
>> crawled at 1/4 speed in the LCS.
>
> Well, OK, that was a particularly stupid configuration. Core was expensive
> in those days and some bean-counter probably only looked at the cost.
>
Stupid for the installation, BRILLIANT for some IBM bean counter. They must
have had a warehouse full of ancient memory cores left over from the 7070's
or 7090's that they were trying to use up! There's no other way that the
LCS made sense to put on any 360 except as some kind of backing store. It
made absolutely NO SENSE to run code out of an LCS.
Now, I can understand some special installation, like maybe airline
reservations, a huge bank or insurance company that wants to have a big
database online. LCS is still 1000 times faster than disk, and the accesses
would be sparse.
But, you are exactly right, having 80% o the main storage as LCS was just a
really bad plan. Of course, this was their first 360, coming from a 7094
system that was all tape. The 7094 was MUCH faster than a 360/50, even with
a good storage configuration, but it didn't multitask, or even have a real
OS. They did batch stuff, with a 1401 making an input tape of card decks
and printing from an output tape of print records, but it was one program at
a time, ran to completion before the next one started. So, really FAST
computer, mostly waiting on mag tape I/O. The 360 must have sounded
wonderful on paper...
Jon
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403096 is a reply to message #403075] |
Mon, 21 December 2020 12:56 |
Jon Elson
Messages: 646
Registered: April 2013
Karma: 0
|
Senior Member |
|
|
Quadibloc wrote:
>
> Of course, using the computer cost enough per hour that the bean counter
> didn't quite do his job properly,. Unless the computer was used in such a
> way that 90% of what it did took place in that one small partition.
That partition was reserved for the accounting department.
12 hour batch job turnaround was quite common, 8 hours was normal when
LIGHTLY loaded. WHAT a pain!
And, it crashed every few hours, and took 45 minutes to IPL and get HASP II
started.
Jon
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403098 is a reply to message #403077] |
Mon, 21 December 2020 12:58 |
Jon Elson
Messages: 646
Registered: April 2013
Karma: 0
|
Senior Member |
|
|
J. Clarke wrote:
>> With Dennard scaling dead, and Moore's law finally showing signs of
>> slowing down, it may be that ten years from now, we will be using 128-core
>> machines, but still stuck below 5 GHz. Or five years from now, at least.
This increase in CPU cores is great for server farms, but doesn't help the
average desktop user. A fewe apps are parallelizable, but most are not, and
probably won't be for a while, if ever.
Jon
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403099 is a reply to message #403090] |
Mon, 21 December 2020 13:08 |
Dan Espen
Messages: 3867
Registered: January 2012
Karma: 0
|
Senior Member |
|
|
Jon Elson <elson@pico-systems.com> writes:
> Charlie Gibbs wrote:
>
>
>> Those limits were somewhat arbitrary. IBM decreed that the
>> 360/30 had a maximum memory of 64K, but I once used one with
>> 128K. A third party added more memory and was able to extend
>> the address bus as required. The extra switch and indicator
>> added to the front panel had a bit of a homebrewed look, but
>> the machine worked just fine. Apparently some outfits expanded
>> model 30s to as much as 512K.
>>
> This one has always interested me. Since the /30 was an 8-bit architecture
> (8 bit + parity memory and 8-bit ALU data paths), why would they have
> implemented the microcode to do more than 16-bit address arithmetic? Just
> so they could detect address overflows in RX instructions? Seems like it
> could detect the overflow anyway, without actually processing the 3rd
> address byte.
>
> Somebody at IBM must have decided they would take the performance loss to do
> the extra add cycle on every RX instruction because they planned that
> someday they might make a machine with more than 16 bits of address.
> Without this, it would have been VERY difficult to expand the address, as it
> would require major alteration of the microcode. I'm pretty sure that the
> /30 did not have a separate address adder to add up the base register, the
> index register and the offset, it was all run through the main ALU.
>
> Anybody know any details of this?
Nope, but it was perfectly legal in address computation to add 2 large
values, subtract a large value to get a valid address. Even when using
the LA (Load address) instruction.
--
Dan Espen
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403103 is a reply to message #403098] |
Mon, 21 December 2020 13:08 |
Ahem A Rivet's Shot
Messages: 4843
Registered: January 2012
Karma: 0
|
Senior Member |
|
|
On Mon, 21 Dec 2020 11:58:30 -0600
Jon Elson <elson@pico-systems.com> wrote:
> This increase in CPU cores is great for server farms, but doesn't help
> the average desktop user. A fewe apps are parallelizable, but most are
> not, and probably won't be for a while, if ever.
Web browsers (and for that matter pages) are heavily multi-threaded
applications and form the majority of the load on average desktops.
--
Steve O'Hara-Smith | Directable Mirror Arrays
C:\>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403107 is a reply to message #403098] |
Mon, 21 December 2020 13:44 |
|
|
Originally posted by: J. Clarke
On Mon, 21 Dec 2020 11:58:30 -0600, Jon Elson <elson@pico-systems.com>
wrote:
> J. Clarke wrote:
>
>
>>> With Dennard scaling dead, and Moore's law finally showing signs of
>>> slowing down, it may be that ten years from now, we will be using 128-core
>>> machines, but still stuck below 5 GHz. Or five years from now, at least.
>
> This increase in CPU cores is great for server farms, but doesn't help the
> average desktop user. A fewe apps are parallelizable, but most are not, and
> probably won't be for a while, if ever.
It really depends on what you're doing. A couple of years back I
turned an "over-the-weekend" task into a "this afternoon" task by
beating VBA into being a scheduler. I was surprised that it worked
but it does and until the IT people switched us over to 64-bit Office
it was stable.
|
|
|
|
| Re: An Argument for Big-Endian: Packed Decimal [message #403132 is a reply to message #403094] |
Tue, 22 December 2020 12:39 |
Peter Flass
Messages: 8375
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
Jon Elson <elson@pico-systems.com> wrote:
> Peter Flass wrote:
>
>> Jon Elson <elson@pico-systems.com> wrote:
>>> Peter Flass wrote:
>>>
>>>
>>>> I think you could link your programs so that part of them loaded into
>>>> processor storage and part into LCS.
>>>
>>> OK, that particular 360/50 had, I think 128 KB of internal storage and 1
>>> MB
>>> of LCS. The OS and one small partition fit in the 128K, everything else
>>> crawled at 1/4 speed in the LCS.
>>
>> Well, OK, that was a particularly stupid configuration. Core was expensive
>> in those days and some bean-counter probably only looked at the cost.
>>
> Stupid for the installation, BRILLIANT for some IBM bean counter. They must
> have had a warehouse full of ancient memory cores left over from the 7070's
> or 7090's that they were trying to use up! There's no other way that the
> LCS made sense to put on any 360 except as some kind of backing store. It
> made absolutely NO SENSE to run code out of an LCS.
I think the reason for LCS was that as more memory was added the signal
path lengths got longer and could no longer support faster memory. Even
running slower I understand that the layout of the boxes was critical to
keep the memory as close to the CPU as possible.
--
Pete
|
|
|
|
| Re: LCS, was An Argument for Big-Endian: Packed Decimal [message #403142 is a reply to message #403132] |
Tue, 22 December 2020 14:38 |
John Levine
Messages: 1405
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
In article <1717356111.630351271.439249.peter_flass-yahoo.com@news.eternal-september.org>,
Peter Flass <peter_flass@yahoo.com> wrote:
>> Stupid for the installation, BRILLIANT for some IBM bean counter. They must
>> have had a warehouse full of ancient memory cores left over from the 7070's
>> or 7090's that they were trying to use up! There's no other way that the
>> LCS made sense to put on any 360 except as some kind of backing store. It
>> made absolutely NO SENSE to run code out of an LCS.
>
> I think the reason for LCS was that as more memory was added the signal
> path lengths got longer and could no longer support faster memory. Even
> running slower I understand that the layout of the boxes was critical to
> keep the memory as close to the CPU as possible.
Pugh et al. say that LCS was slow cost reduced core memory that would
be in the gap between regular fast core and disks and drums. The
original idea was that the /65 and up would only have selector (mass
storage) channels and would use shared LCS to connect to a smaller 360
model that ran slower devices on a multiplexor channel.
But LCS was never enough cheaper than RAM to be cost effective, and
the fact that it used the regular address space made it impossible to
attach to small models. So they added multiplexor channels to the high
end models, and built a channel-to-channel device to connect two systems.
There were a few high profile LCS customers like NASA who used it as a
fast disk.
It did not sound like the LCS core was related to the 709x core.
--
Regards,
John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. https://jl.ly
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406240 is a reply to message #402977] |
Sat, 06 March 2021 22:27 |
|
|
Originally posted by: Robert Boyer
I have always LOVED the beam spring keyboards on my 37XX terminals. I
think I was one of the last one standing when they were replaced with
the newer models that had the model-f and then model-m type
keyboards. Eventually I think I might have been the ONLY device hooked
to a 3745 and kept that alive way after everything else was tn3270/etc I
loved that terminal keyboard even though I used a lot of other stuff
most of the time as time went by and my use of XEDIT etc declined. I had
to keep it just to remind myself of what a keyboard "SHOULD" be
like. Sure the model-f's and model-m type switches were okay, especially
by the standards of mushy and instantly bottomed out keybords that are
the norm...
I am kind of happy with a few of my keyboards but it's been a lot of
experimentation and a bunch of dead ends...
The closest modern keyboard build I have at the moment is probably my
board built with Kailh Box Navy key switches (not perfect) and either my
SA super thick caps or my MT3 /dev/tty (which are claimed to be exact
copies of 3270 profiles) caps. I have two versions with those
caps/switches = the heavy version that will not budge on my desk... and
the light version that can go with me when I take my laptop/cylinder Mac
Pro or believe it or not... increasingly my iPad Pro (the really big
screen version) now that I have blink and a fire-up on demand docker image
that has a perfectly tuned emacs running that remotes to everything else
on one of my cloud servers working properly with Mosh and TMUX that
allows seamless connect/disconnect like I never left... the whole docker
+ Mosh + tmux thing took a while get to work right but it's actually one
of the better dev environments I've ever had for everything from C to
LISP to Haskell to various Assemblers... on a frigging iPad (well not
really on there but it certainly feels like it even over a cell
connection)
Just thought I'd share for anyone else that misses their beam springs
from the late 70's and early 80's. Kind of funny as from the late 80's
on XEDIT was not really used more than a project here or there for me
on/off again.
|
|
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406255 is a reply to message #406242] |
Sun, 07 March 2021 13:15 |
|
|
Originally posted by: Robert Boyer
Wow... thanks, surprised I have not run across this. I will probably get
in on the pre-order just to see if they are actually better than my
current Box Navy switches that are pretty similar in force and
activation curves. Of course none of them will sound like a real 3278/9
with the little relay thing that pongs the side of the idiotically thick
metal keyboard case just incase the beam springs weren't tactile enough
on their own ;-)
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406259 is a reply to message #406240] |
Sun, 07 March 2021 15:49 |
Peter Flass
Messages: 8375
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
Robert Boyer <rwboyer@mac.com> wrote:
>
> I have always LOVED the beam spring keyboards on my 37XX terminals. I
> think I was one of the last one standing when they were replaced with
> the newer models that had the model-f and then model-m type
> keyboards. Eventually I think I might have been the ONLY device hooked
> to a 3745 and kept that alive way after everything else was tn3270/etc
Pretty much the dame for me. I loved my 3290 and kept it long after
everyone else went tp PCs and tn3270. I had both the termonal and PC on my
desc and switched back and forth depending on what I was doing.
>
> I loved that terminal keyboard even though I used a lot of other stuff
> most of the time as time went by and my use of XEDIT etc declined. I had
> to keep it just to remind myself of what a keyboard "SHOULD" be
> like. Sure the model-f's and model-m type switches were okay, especially
> by the standards of mushy and instantly bottomed out keybords that are
> the norm...
I finally got fed up with the cr@p Dell keyboard I had and bought a Unicomp
Model-M. My only complaint is my wife says it’s too loud when she’s trying
to sleep in the next room.
--
Pete
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406261 is a reply to message #406259] |
Sun, 07 March 2021 18:25 |
John Levine
Messages: 1405
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
In article <1624617332.636842575.976079.peter_flass-yahoo.com@news.eternal-september.org>,
Peter Flass <peter_flass@yahoo.com> wrote:
> I finally got fed up with the cr@p Dell keyboard I had and bought a Unicomp
> Model-M. My only complaint is my wife says it’s too loud when she’s trying
> to sleep in the next room.
That is the exact reason I switched from a Unicomp back to something else.
Hmm, which do I love more, my keyboard or my wife? ....
--
Regards,
John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. https://jl.ly
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406262 is a reply to message #406261] |
Sun, 07 March 2021 19:08 |
|
|
Originally posted by: Robert Boyer
Yea, I was unhappy with most of the keyboards I tried either because of
te travel or the feel of the switches or because of the keycaps... still
looking for the "one" but after trying a lot of different things I am
finally happey with the Box Navy switches and a few of the SA or similar
caps with a profile that has a better "feel"/reachability that I prefer
or am trained into with my old fashioned muscule memory. I'm still
looking but my rate of trying things has slowed considerably in the last
few years since I settled on a cap profile or two and the Box Navy's
came out. Prior to that the Gateron Blue's were a reasonable
alternative. For some reason I was never happy with the m-type clones
and I am not sure why?? Maybe because my fav was the 3279 and second the
PC-AT early F-types?? Who knows, I am okay now and I don't have to deal
with PS/2 to USB stuff and finally I have something that's good AND
portable when I choose to use my iPad-term which is fairly asesome with
Blink.
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406263 is a reply to message #406262] |
Sun, 07 March 2021 19:20 |
|
|
Originally posted by: J. Clarke
On Sun, 07 Mar 2021 19:08:14 -0500, Robert Boyer <rwboyer@mac.com>
wrote:
>
> Yea, I was unhappy with most of the keyboards I tried either because of
> te travel or the feel of the switches or because of the keycaps... still
> looking for the "one" but after trying a lot of different things I am
> finally happey with the Box Navy switches and a few of the SA or similar
> caps with a profile that has a better "feel"/reachability that I prefer
> or am trained into with my old fashioned muscule memory. I'm still
> looking but my rate of trying things has slowed considerably in the last
> few years since I settled on a cap profile or two and the Box Navy's
> came out. Prior to that the Gateron Blue's were a reasonable
> alternative. For some reason I was never happy with the m-type clones
> and I am not sure why?? Maybe because my fav was the 3279 and second the
> PC-AT early F-types?? Who knows, I am okay now and I don't have to deal
> with PS/2 to USB stuff and finally I have something that's good AND
> portable when I choose to use my iPad-term which is fairly asesome with
> Blink.
My problem with the Unicomps is that every single one of them that I
have had would start typing "-" signs periodically for no apparent
reason.
I have a couple of Logitech gamer boards with nice feel but the
keycaps are shot on one and there aren't any replacements readily
available, while various keys on the other stick periodically,
generally at the least opportune moment.
So I'm still looking for the perfect keyboard. I'm tempted to try one
of the Keystones (I wish I had known about the Kickstarter on that
one) but not for a while--my AVR just downloaded an update and bricked
itself, so the "fun stuff" reserve is a bit depleted.
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406265 is a reply to message #406263] |
Sun, 07 March 2021 20:00 |
Dan Espen
Messages: 3867
Registered: January 2012
Karma: 0
|
Senior Member |
|
|
J. Clarke <jclarke.873638@gmail.com> writes:
> On Sun, 07 Mar 2021 19:08:14 -0500, Robert Boyer <rwboyer@mac.com>
> wrote:
>
>>
>> Yea, I was unhappy with most of the keyboards I tried either because of
>> te travel or the feel of the switches or because of the keycaps... still
>> looking for the "one" but after trying a lot of different things I am
>> finally happey with the Box Navy switches and a few of the SA or similar
>> caps with a profile that has a better "feel"/reachability that I prefer
>> or am trained into with my old fashioned muscule memory. I'm still
>> looking but my rate of trying things has slowed considerably in the last
>> few years since I settled on a cap profile or two and the Box Navy's
>> came out. Prior to that the Gateron Blue's were a reasonable
>> alternative. For some reason I was never happy with the m-type clones
>> and I am not sure why?? Maybe because my fav was the 3279 and second the
>> PC-AT early F-types?? Who knows, I am okay now and I don't have to deal
>> with PS/2 to USB stuff and finally I have something that's good AND
>> portable when I choose to use my iPad-term which is fairly asesome with
>> Blink.
>
> My problem with the Unicomps is that every single one of them that I
> have had would start typing "-" signs periodically for no apparent
> reason.
>
> I have a couple of Logitech gamer boards with nice feel but the
> keycaps are shot on one and there aren't any replacements readily
> available, while various keys on the other stick periodically,
> generally at the least opportune moment.
>
> So I'm still looking for the perfect keyboard. I'm tempted to try one
> of the Keystones (I wish I had known about the Kickstarter on that
> one) but not for a while--my AVR just downloaded an update and bricked
> itself, so the "fun stuff" reserve is a bit depleted.
As we've discussed here before, I bought a Logitech G815.
The keys are definitely not clicky and the feel is unique.
Sort of an even resistance but you can tell when you've hit the key.
It's an RGB keyboard and it makes me smile every time I look at it.
Logitech redesigned the keys so that the LED is up inside the key
so you don't see the LED light leaking around the keys. I love the
audio keys and the roller volume control.
So it's not a 3279 but I like it a lot.
--
Dan Espen
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406266 is a reply to message #406259] |
Sun, 07 March 2021 19:12 |
|
|
Originally posted by: songbird
Peter Flass wrote:
....
> I finally got fed up with the cr@p Dell keyboard I had and bought a Unicomp
> Model-M. My only complaint is my wife says it’s too loud when she’s trying
> to sleep in the next room.
i typed with mine under the blankets.
they're not really the same keyboard and a plastic
case isn't as good as the metal ones were.
they didn't last and i won't buy again from them.
be wary of using them typing draped across your lap.
that's how i do all my typing now and they both stopped
working shortly after the warranty period expired.
they're probably fine if you use them on a flat desk.
i also had an issue with one keyboard they sent
having a different switch (the sound and the feel were
different) that they would not replace because the key
did work, it just wasn't the same as the rest of the keys
on the keyboard. a simple quality check before sending
the keyboard out would have caught that.
after the warranty period expired i tried to fix them
with no luck, i don't really understand how the
connection ribbon cables work and there must be some kind
of trick to getting them back together. ah well. too
bad.
songbird
|
|
|
|
|
|
| Re: keyboards [message #406274 is a reply to message #406263] |
Mon, 08 March 2021 06:30 |
Niklas Karlsson
Messages: 265
Registered: January 2012
Karma: 0
|
Senior Member |
|
|
On 2021-03-08, J Clarke <jclarke.873638@gmail.com> wrote:
>
> So I'm still looking for the perfect keyboard. I'm tempted to try one
> of the Keystones (I wish I had known about the Kickstarter on that
> one) but not for a while--my AVR just downloaded an update and bricked
> itself, so the "fun stuff" reserve is a bit depleted.
I'm apparently not what you'd call a picky typist. I tend to get used to
pretty much any keyboard that isn't entirely godawful.
I was uncomfortable with the very short travel of the keys on my shiny
new M1 MacBook Pro that I use for work, but I've more or less gotten
used to that, too.
Niklas
--
Perl is one of those languages where any idiot can write a reasonably large
program and it will run and appear to work. A program written in C by a
similarly-incompetent programmer will usually to crash and burn in an obvious
manner and not get as far as deployment. -- Peter Corlett
|
|
|
|
|
|
| Re: keyboards, address sizes, was An Argument for something [message #406301 is a reply to message #406294] |
Tue, 09 March 2021 14:09 |
John Levine
Messages: 1405
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
In article <1995645828.636992207.698788.peter_flass-yahoo.com@news.eternal-september.org>,
Peter Flass <peter_flass@yahoo.com> wrote:
> John Levine <johnl@taugh.com> wrote:
>> In article <1624617332.636842575.976079.peter_flass-yahoo.com@news.eternal-september.org>,
>> Peter Flass <peter_flass@yahoo.com> wrote:
>>> I finally got fed up with the cr@p Dell keyboard I had and bought a Unicomp
>>> Model-M. My only complaint is my wife says it’s too loud when she’s trying
>>> to sleep in the next room.
>>
>> That is the exact reason I switched from a Unicomp back to something else.
>
> What did you switch to?
One like this. I hate it, but it is definitely quiet, and I am still married.
https://satechi.net/products/aluminum-wired-usb-keyboard?var iant=12420646404184
--
Regards,
John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. https://jl.ly
|
|
|
|
|
|
| Re: keyboards, address sizes, was An Argument for Big-Endian: Packed Decimal [message #406307 is a reply to message #406302] |
Tue, 09 March 2021 18:06 |
Peter Flass
Messages: 8375
Registered: December 2011
Karma: 0
|
Senior Member |
|
|
Questor <usenet@only.tnx> wrote:
> On Sun, 7 Mar 2021 23:25:03 -0000 (UTC), John Levine <johnl@taugh.com> wrote:
>> In article <1624617332.636842575.976079.peter_flass-yahoo.com@news.eternal-september.org>,
>> Peter Flass <peter_flass@yahoo.com> wrote:
>>> I finally got fed up with the cr@p Dell keyboard I had and bought a Unicomp
>>> Model-M. My only complaint is my wife says it’s too loud when she’s trying
>>> to sleep in the next room.
>>
>> That is the exact reason I switched from a Unicomp back to something else.
>>
>> Hmm, which do I love more, my keyboard or my wife? ....
>
> "...so she told me it was either her or the ham radio, over."
>
>
"This is Lexi, she's an 8 week-old German Sheppard, I bought Lexi as a
surprise for my wife but it turns out she is allergic to dogs so we are now
looking to find her a new home. Her name is Siobhan, she is 39 years old, a
beautiful and caring woman who drives, is a great cook and keeps a good
house."
--
Pete
|
|
|
|
Members
Search
Help
Register
Login
Home
