Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.misc > #1741 > unrolled thread

Re: The Hex concept

Back to article view | Back to comp.misc

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: The Hex concept "Dave" <dwickford@yahoo.com> - 2014-03-17 20:46 +0000
    Re: The Hex concept RS Wood <rsw@therandymon.com> - 2014-03-18 08:59 +0000
      Re: The Hex concept Paul Sture <nospam@sture.ch> - 2014-03-18 13:42 +0100
        Re: The Hex concept RS Wood <rsw@therandymon.com> - 2014-03-18 19:02 +0000
          Software Bloat Oregonian Haruspex <bob_davis_retired@yahoo.com> - 2014-03-18 19:19 +0000
            Re: Software Bloat Johny B Good <johnny-b-good@invalid.ntlworld.com> - 2014-03-18 20:51 +0000
              Re: Software Bloat Oregonian Haruspex <bob_davis_retired@yahoo.com> - 2014-03-18 14:49 -0700
                Re: Software Bloat Johny B Good <johnny-b-good@invalid.ntlworld.com> - 2014-03-18 22:05 +0000
                  Re: USB interface RS Wood <rsw@therandymon.com> - 2014-03-20 20:09 +0000
                    Re: USB interface Huge <Huge@nowhere.much.invalid> - 2014-03-20 20:24 +0000
                    Re: USB interface Johny B Good <johnny-b-good@invalid.ntlworld.com> - 2014-03-21 04:31 +0000
                      Re: USB interface RS Wood <rsw@therandymon.com> - 2014-03-21 09:39 +0000
                        Re: USB interface RS Wood <rsw@therandymon.com> - 2014-05-01 21:19 +0000
                          Re: USB interface Mike Spencer <mds@bogus.nodomain.nowhere> - 2014-05-01 23:33 -0300
                          Re: USB interface Tim Watts <tw_usenet@dionic.net> - 2014-05-02 08:49 +0100
                            Re: USB interface Bob Eager <news0002@eager.cx> - 2014-05-02 07:56 +0000
                              Re: USB interface Tim Watts <tw_usenet@dionic.net> - 2014-05-02 10:39 +0100
                                Re: USB interface Bob Eager <news0002@eager.cx> - 2014-05-02 10:21 +0000
                                  Re: USB interface Tim Watts <tw_usenet@dionic.net> - 2014-05-02 19:30 +0100
                            Re: USB interface Huge <Huge@nowhere.much.invalid> - 2014-05-02 08:23 +0000
                            Re: USB interface Tim Streater <timstreater@greenbee.net> - 2014-05-02 09:58 +0100
                              Re: USB interface RS Wood <rsw@therandymon.com> - 2014-05-02 09:14 +0000
                                Re: USB interface Tim Streater <timstreater@greenbee.net> - 2014-05-02 11:06 +0100
                                  Re: USB interface Bob Eager <news0002@eager.cx> - 2014-05-02 10:21 +0000
                                    Re: USB interface Paul Sture <nospam@sture.ch> - 2014-05-02 22:35 +0200
                            Re: USB interface Will Honea <whonea@whonea.net> - 2014-05-02 12:25 -0600
                          Re: USB interface Anssi Saari <as@sci.fi> - 2014-05-02 13:23 +0300
                            Re: USB interface Dag-Erling Smørgrav <des@des.no> - 2014-05-02 15:29 +0200
                              Re: USB interface Anssi Saari <as@sci.fi> - 2014-05-06 16:47 +0300
                                Re: USB interface Dag-Erling Smørgrav <des@des.no> - 2014-05-08 16:55 +0200
                      Re: USB interface Rich <rich@example.com> - 2014-03-21 15:20 +0000
                        Re: USB interface Johny B Good <johnny-b-good@invalid.ntlworld.com> - 2014-03-22 00:41 +0000
                          Re: USB interface Dave Garland <dave.garland@wizinfo.com> - 2014-03-21 22:44 -0500
                            Re: USB interface Oregonian Haruspex <bob_davis_retired@yahoo.com> - 2014-03-22 00:49 -0700
                              Re: USB interface Paul Sture <nospam@sture.ch> - 2014-03-22 09:46 +0100
                                Re: USB interface Dave Garland <dave.garland@wizinfo.com> - 2014-03-22 12:42 -0500
                              Re: USB interface Johny B Good <johnny-b-good@invalid.ntlworld.com> - 2014-03-22 15:17 +0000
                Re: Software Bloat RS Wood <rsw@therandymon.com> - 2014-03-18 22:07 +0000
                  Re: Software Bloat Oregonian Haruspex <bob_davis_retired@yahoo.com> - 2014-03-18 15:38 -0700
                    Re: Software Bloat Spalls Hurgenson <spallshurgenson@gmail.com> - 2014-03-19 10:30 -0400
                      Re: Software Bloat Huge <Huge@nowhere.much.invalid> - 2014-03-19 15:40 +0000
                        Re: Software Bloat Tim Streater <timstreater@greenbee.net> - 2014-03-19 16:27 +0000
                    Re: Software Bloat Paul Sture <nospam@sture.ch> - 2014-03-19 16:59 +0100
                  Re: Software Bloat polygonum <rmoudndgers@vrod.co.uk> - 2014-03-18 22:55 +0000
                    Re: Software Bloat Whiskers <catwheezel@operamail.com> - 2014-03-19 12:27 +0000
                      Re: Software Bloat Tim Streater <timstreater@greenbee.net> - 2014-03-19 16:26 +0000
                      Re: Software Bloat RS Wood <rsw@therandymon.com> - 2014-03-19 20:07 +0000

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

#2598 — Re: USB interface

In article <7avb3b-via.ln1@squidward.local.dionic.net>, Tim Watts
<tw_usenet@dionic.net> wrote:

> On 01/05/14 22:19, RS Wood wrote:
> 
> > Resurrecting this old thread because there's been an interesting post at
> > Hacker News and the comments in that thread are pretty interesting. The
> > article is interesting, too.  Never thought I'd hear such a vigorous
> > defense for RS -232!
> >
> > USB Sucks (2005):
> > http://www.technozeal.com/topic2.html
> >
> > Thread on Hacker News:
> > https://news.ycombinator.com/item?id=7649063
> >
> 
> Meanwhile, whilst potentially simple to make electronics for, RS232 was 
> in practise the biggest dogs breakfast of all time.
> 
> Many many speeds (fair enough),
> 
> different parity, random combinations of handshaking signals, stop bits,
> 
> and worst of all - probably the most use and abuse of connectors known 
> to man. I have personally seen:
> 
> DB9, DB25 (good), RJ45 (well, handy for patching, but, ewww), 
> PS2/mini-din (whyFFS), 3.5mm(1/8") stereo jack (yuk yuk yuk)...
> 
> Sure there were more, but WTF...

And then there was the DCE/DTE bullshit. The removal of RS232 ports is
no loss whatever. And if you are one of the vanishingly small number of
people who has to visit a PoP, and while there needs to connect their
lappy to a router or other large bit of kit, there are very reliable
USB to RS232 converters about. I bought a two-porter one of these to
connect my Mac Mini to a 68000 SBC and never had a problem with it.

-- 
Tim

"That excessive bail ought not to be required, nor excessive fines imposed,
nor cruel and unusual punishments inflicted"  --  Bill of Rights 1689

[toc] | [prev] | [next] | [standalone]

#2599 — Re: USB interface

 And if you are one of the vanishingly small number of
> people who has to visit a PoP, and while there needs to connect their
> lappy to a router or other large bit of kit, there are very reliable
> USB to RS232 converters about. I bought a two-porter one of these to
> connect my Mac Mini to a 68000 SBC and never had a problem with it.


Yes, when I was mucking about with a serial modem a year or two ago
(still, far and away the easiest way to get a Linux box connected over
a phone line), I used this Trendnet cable with a Trendnet external
serial modem, with extremely good results.

http://www.amazon.com/TRENDnet-RS-232-Serial-Converter-TU-S9/dp/B0007T27H8/ref=sr_1_1?ie=UTF8&qid=1399021541&sr=8-1&keywords=usb+serial+adaptor

I still kind of miss watching those flashing lights on the modem.  Not
enough to go back to ~56K though.  Anyway, you're right - as long as
you buy carefully, a USB-Serial adaptor is often good enough.  The only
device I ever failed to connect was a Psion 5MX to a Mac, and I never
figured out if it was the cable, the Mac software, or just bad luck,
and pretty soon I got tired of trying.

[toc] | [prev] | [next] | [standalone]

#2605 — Re: USB interface

In article <20140502091429.b7a3519f.rsw@therandymon.com>, RS Wood
<rsw@therandymon.com> wrote:

> I still kind of miss watching those flashing lights on the modem.

I miss them on the computer, too.

-- 
Tim

"That excessive bail ought not to be required, nor excessive fines imposed,
nor cruel and unusual punishments inflicted"  --  Bill of Rights 1689

[toc] | [prev] | [next] | [standalone]

#2607 — Re: USB interface

On Fri, 02 May 2014 11:06:16 +0100, Tim Streater wrote:

> In article <20140502091429.b7a3519f.rsw@therandymon.com>, RS Wood
> <rsw@therandymon.com> wrote:
> 
>> I still kind of miss watching those flashing lights on the modem.
> 
> I miss them on the computer, too.

Not to mention the loudspeaker....could debug quite a bit by listening.

-- 
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
 http://www.mirrorservice.org

[toc] | [prev] | [next] | [standalone]

#2616 — Re: USB interface

On 2014-05-02, Bob Eager <news0002@eager.cx> wrote:
> On Fri, 02 May 2014 11:06:16 +0100, Tim Streater wrote:
>
>> In article <20140502091429.b7a3519f.rsw@therandymon.com>, RS Wood
>> <rsw@therandymon.com> wrote:
>> 
>>> I still kind of miss watching those flashing lights on the modem.
>> 
>> I miss them on the computer, too.
>
> Not to mention the loudspeaker....could debug quite a bit by listening.

I never worked out whether the RK05 disks we had on our PDP had separate
speakers or the sound effects were produced by the disk head movement itself,
but the pattern of sounds was very distinctive for certain jobs.  We did
have a deadly embrace which occurred every couple of months or so and
that had its own rhythm.  The flashing lights confirmed it and we knew
which program to kill to unjam it.

(I tried in vain to track the offending deadly embrace down, but I never
managed to do so - it was faster to write a program which recovered from
killing the offending process, which catered for recovery from system
crashes too)


-- 
Paul Sture

The final step of #heartbleed recovery is to call your mother, and advise
her to change her maiden name                                  -- @gojomo

[toc] | [prev] | [next] | [standalone]

#2611 — Re: USB interface

Tim Watts wrote:

> On 01/05/14 22:19, RS Wood wrote:
> 
>> Resurrecting this old thread because there's been an interesting post at
>> Hacker News and the comments in that thread are pretty interesting. The
>> article is interesting, too.  Never thought I'd hear such a vigorous
>> defense for RS -232!
>>
>> USB Sucks (2005):
>> http://www.technozeal.com/topic2.html
>>
>> Thread on Hacker News:
>> https://news.ycombinator.com/item?id=7649063
>>
> 
> Meanwhile, whilst potentially simple to make electronics for, RS232 was
> in practise the biggest dogs breakfast of all time.
> 
> Many many speeds (fair enough),
> 
> different parity, random combinations of handshaking signals, stop bits,
> 
> and worst of all - probably the most use and abuse of connectors known
> to man. I have personally seen:
> 
> DB9, DB25 (good), RJ45 (well, handy for patching, but, ewww),
> PS2/mini-din (whyFFS), 3.5mm(1/8") stereo jack (yuk yuk yuk)...
> 
> Sure there were more, but WTF...

I had to acquire an RS232 dongle to supplement USB on my laptop as soon as 
they had dropped the RS232 port.  You would be amazed at the amount of 
legacy gear still in the field that requires RS232 interface, especially in 
the industrial control world.  Many systems which require regulatory 
certification - such as nuclear plants - are still governed by 40 year old 
regs.  USB? WTF is USB?  Ethernet? Never heard of it.

-- 
Will Honea

[toc] | [prev] | [next] | [standalone]

#2608 — Re: USB interface

RS Wood <rsw@therandymon.com> writes:

> Resurrecting this old thread because there's been an interesting post
> at Hacker News and the comments in that thread are pretty interesting.
> The article is interesting, too.  Never thought I'd hear such a
> vigorous defense for RS -232!

I guess the most amusing thing is that Intel themselves are even today
making motherboards with all (or most) the legacy ports included. In the
guise of cheapies with slow Atom processors though. So what happened to
those people at Intel who were pushing those "legacy free PCs" when USB
first appeared?

[toc] | [prev] | [next] | [standalone]

#2610 — Re: USB interface

Anssi Saari <as@sci.fi> writes:
> I guess the most amusing thing is that Intel themselves are even today
> making motherboards with all (or most) the legacy ports included. In the
> guise of cheapies with slow Atom processors though. So what happened to
> those people at Intel who were pushing those "legacy free PCs" when USB
> first appeared?

Legacy free does not mean no serial port, it means no i8042 keyboard
controller, no A20 gate, and usually a very minimal BIOS.

(If you know what the above means, you can stop reading here)

In order to address up to 1 MB of RAM using 16-bit registers, the i8086
and its descendents (when running in real mode) have an addressing
scheme based on segments and offsets.  All memory access is relative to
one of four segment registers (CS, DS, ES, SS) and the physical address
is computed by left-shifting the segment by four and adding the offset.
With only 20 address lines, the carry is lost and addresses beyond 1 MB
wrap around (65535 * 16 + 65535 = 1114095 which wraps around to 65519).

About three minutes after the first IBM PC shipped, idi^H^H^Hprogrammers
started writing code that relied on this feature.  So when the 80286
arrived with its 24-bit bus, address line 20 had to be disabled at boot
so addresses beyond 1 MB would still wrap around.  Obviously, software
that wanted access to the full address space needed a way to re-enable
the A20 line.  Someone noticed that the i8042 keyboard controller had an
unused pin and decided that it made perfect sense to use that pin to
control the gate that enabled the A20 line.

(Another unused pin on the i8042 was hooked up to the CPU's reset pin to
allow the CPU to hard-reset itself, but that's a whole other story.)

In the mid-2000s, I worked on a Linux-based thin client OS which was
designed to run on pretty much anything you could lay your hands on
(which made it very popular with cash-strapped schools).  We had to
patch the kernel to ignore a failure to flip the A20 gate so it would
work on "legacy-free" thin clients.  These days, Linux checks whether
the A20 line is enabled before trying to flip the gate.

If memory serves, Apple was the first major manufacturer to ship x86
computers with no i8042 (whether discrete or emulated) and no A20 gate.

DES
-- 
Dag-Erling Smørgrav - des@des.no

[toc] | [prev] | [next] | [standalone]

#2647 — Re: USB interface

Dag-Erling Smørgrav <des@des.no> writes:

> Anssi Saari <as@sci.fi> writes:
>> I guess the most amusing thing is that Intel themselves are even today
>> making motherboards with all (or most) the legacy ports included. In the
>> guise of cheapies with slow Atom processors though. So what happened to
>> those people at Intel who were pushing those "legacy free PCs" when USB
>> first appeared?
>
> Legacy free does not mean no serial port, it means no i8042 keyboard
> controller, no A20 gate, and usually a very minimal BIOS.

Strange view. I meant the common variety of legacy-free as explained by
Wikipedia at http://en.wikipedia.org/wiki/Legacy-free_PC.

[toc] | [prev] | [next] | [standalone]

#2700 — Re: USB interface

Anssi Saari <as@sci.fi> writes:
> Dag-Erling Smørgrav <des@des.no> writes:
> > Legacy free does not mean no serial port, it means no i8042 keyboard
> > controller, no A20 gate, and usually a very minimal BIOS.
> Strange view. I meant the common variety of legacy-free as explained
> by Wikipedia at http://en.wikipedia.org/wiki/Legacy-free_PC.

Wikipedia's definition simply says that certain components are
physically absent or inaccessible.  That's a useless definition,
especially since modern PCs _do_ have serial and paralell ports and
expansion buses, even if they aren't visible, as well as integrated
peripherals which behave as if they were on an ISA bus.

DES
-- 
Dag-Erling Smørgrav - des@des.no

[toc] | [prev] | [next] | [standalone]

#1858 — Re: USB interface

Johny B Good <johnny-b-good@invalid.ntlworld.com> wrote:
> On Thu, 20 Mar 2014 20:09:54 +0000 (UTC), RS Wood
> <rsw@therandymon.com> wrote:

> >On 2014-03-18, Johny B Good <johnny-b-good@invalid.ntlworld.com> wrote:
> >>  We all know it will never come to that. However, assuming such a
> >> scenario, by that logic the windows 98 and the windows XP
> >> developers will be in exactly the same boat (as will those two
> >> patsies at Intel responsible for 'inventing' the USB interface).
> >
> >Incidentally, what's your gripe about the USB interface?  Honestly
> >curious - is it the fact that 99% of the time, you wind up sticking
> >the plug in the wrong way, or something more fundamental that that?
> >

>  Oh, it was much more fundamental than that. It's sole purpose seems
> to have been to help push the sales of intel's next generation of
> money spinning processors since it was such a profligate consumer of
> cpu cycles for so little return.

> >I'd think compared to what it replaced for common use peripherals -
> >PS/2, parallel, and serial - USB is an improvement, though there's
> >room to grow.

>  It's only in recent years that cpu grunt finally reached a state
> where USB2's voracious appetite for cpu cycles could finally be
> satisfied without too much detriment to the rest of the system's
> performance.

Keep in mind USB's original design parameters (I'm talking USB 1.0). 
Very low bandwidth async and synchronous comm. for things like
keyboards, mice, audio speakers, etc.  1.5mbit and 12mbit.  At those
rates, even when 1.0 was released, designing it as PIO was not
necessarily "wrong".  One could argue that designing in DMA would have
been superior, and one would be correct.  But at a max of 12mbit
(seldom actually ever achieved anyway) the CPU load for doing PIO
transfers just was not that huge of a hit.  And for its design space
(keyboards, mice, speakers) none of those items would have even
saturated the 12mbit rate.  Keyboards, while they might transmit at the
12mbit rate, will ship a short packet each time you hit a key, then
have a significant pause (from the CPU's perspective) before the next
PIO load hits.  Mice are the same.  Even vigorous mousing wouldn't
significantly load the CPU.  Speakers, well, assuming you ship out
48Khz 16-bit samples, two channel you are looking at right about the
1.5MBit slow rate for all of the data.  Add framing/packet overhead and
maybe the stream is at most 2Mbit.  Also not a serious CPU time sink.  

And lack of DMA, in that environment, meant the interface chips were
significantly cheaper to design and build.

>  USB3 has had to be redesigned to use dma to work at its much higher
> data rates so is finally what USB should have been based upon in the
> first place, rather than using PIO methods designed to artificially
> generate a need to upgrade cpu performance to suit Intel's purpose.

One can make a good argument that when USB 1 was extended to USB 2,
that DMA should have been added at that point.  And you'd be right.  So
the vitrol should be directed towards the USB 2 designers, not at USB
in general.  For it's initial environment, it was not so bad a design. 
Could it have been better, yes, but at the cost of extra expense in the
chips and extra effort in the interface drivers.  Given the low speed
of 1.0, that extra cost/effort probably did not justify itself.

[toc] | [prev] | [next] | [standalone]

#1875 — Re: USB interface

On 21 Mar 2014 15:20:10 GMT, Rich <rich@example.com> wrote:

>Johny B Good <johnny-b-good@invalid.ntlworld.com> wrote:
>> On Thu, 20 Mar 2014 20:09:54 +0000 (UTC), RS Wood
>> <rsw@therandymon.com> wrote:
>
>> >On 2014-03-18, Johny B Good <johnny-b-good@invalid.ntlworld.com> wrote:
>> >>  We all know it will never come to that. However, assuming such a
>> >> scenario, by that logic the windows 98 and the windows XP
>> >> developers will be in exactly the same boat (as will those two
>> >> patsies at Intel responsible for 'inventing' the USB interface).
>> >
>> >Incidentally, what's your gripe about the USB interface?  Honestly
>> >curious - is it the fact that 99% of the time, you wind up sticking
>> >the plug in the wrong way, or something more fundamental that that?
>> >
>
>>  Oh, it was much more fundamental than that. It's sole purpose seems
>> to have been to help push the sales of intel's next generation of
>> money spinning processors since it was such a profligate consumer of
>> cpu cycles for so little return.
>
>> >I'd think compared to what it replaced for common use peripherals -
>> >PS/2, parallel, and serial - USB is an improvement, though there's
>> >room to grow.
>
>>  It's only in recent years that cpu grunt finally reached a state
>> where USB2's voracious appetite for cpu cycles could finally be
>> satisfied without too much detriment to the rest of the system's
>> performance.
>
>Keep in mind USB's original design parameters (I'm talking USB 1.0). 
>Very low bandwidth async and synchronous comm. for things like
>keyboards, mice, audio speakers, etc.  1.5mbit and 12mbit.  At those
>rates, even when 1.0 was released, designing it as PIO was not
>necessarily "wrong".  One could argue that designing in DMA would have
>been superior, and one would be correct.  But at a max of 12mbit
>(seldom actually ever achieved anyway) the CPU load for doing PIO
>transfers just was not that huge of a hit.  And for its design space
>(keyboards, mice, speakers) none of those items would have even
>saturated the 12mbit rate.  Keyboards, while they might transmit at the
>12mbit rate, will ship a short packet each time you hit a key, then
>have a significant pause (from the CPU's perspective) before the next
>PIO load hits.  Mice are the same.  Even vigorous mousing wouldn't
>significantly load the CPU.  Speakers, well, assuming you ship out
>48Khz 16-bit samples, two channel you are looking at right about the
>1.5MBit slow rate for all of the data.  Add framing/packet overhead and
>maybe the stream is at most 2Mbit.  Also not a serious CPU time sink.  
>
>And lack of DMA, in that environment, meant the interface chips were
>significantly cheaper to design and build.
>
>>  USB3 has had to be redesigned to use dma to work at its much higher
>> data rates so is finally what USB should have been based upon in the
>> first place, rather than using PIO methods designed to artificially
>> generate a need to upgrade cpu performance to suit Intel's purpose.
>
>One can make a good argument that when USB 1 was extended to USB 2,
>that DMA should have been added at that point.  And you'd be right.  So
>the vitrol should be directed towards the USB 2 designers, not at USB
>in general.  For it's initial environment, it was not so bad a design. 
>Could it have been better, yes, but at the cost of extra expense in the
>chips and extra effort in the interface drivers.  Given the low speed
>of 1.0, that extra cost/effort probably did not justify itself.

 Well, I pretty much concur with most of what you've said except I
think USB 1 came to be abused by those 'feckin' cheap and nasty
external USB dial up modems (amongst other "USB Gadgets") which now
demanded a Pentium 166MX chip as a minimum (then, latterly P233MX).

 A  33MHz 386SX could have easily done the job with a standard (but
more expensive) serial port connected 56K V90 modem. I'm pretty sure
this sort of abuse was being practised prior to the advent of USB2
(please correct me if I'm wrong).

 I think Intel decided against 'doing the right thing' when they
developed USB2 since the benefit to them of USB1 creating extra demand
for faster CPUs must have been an irresistable 'driver' to avoid the
'needless expense' of DMA.

 Although the cpu grunt required to service the PIO demands of USB2
arrived about a decade ago, the PCs in question would struggle to use
a twin tuner DVB-T adapter to allow 2 simultaneous recordings plus
allow watching another live stream if available from the muxes already
tuned or a recorded one from off the hard disk yet this was a trick
well within the capabilities of any decent twin tuner PVR such as
those made by Humax and Topfield.

 I bought a Topfield 5800 about 5 or 6 years back for the XYL's use. I
was already using my desktop PC (and a laptop to resolve scheduling
clashes) to watch and record Freeview broadcasts by then.

 Occasionally, I'll schedule a recording or two on the Toppy when even
the desktop and laptop isn't enough to resolve a 3 way conflict so
that I can later transfer the recording over the USB2 interface built
into the Toppy.

 The data transfer rate is somewhat execrable to say the least
(obviously faster than USB1 but not by an awful lot - and only after
invoking "Turbo Mode" which bypassess the remote command interpreter
loop in the PVR to release more cpu grunt to the USB interface).

 The situation is even worse for the Humax owner who has to resort to
a software fix that essentially uses a duplicate packet sending error
correction protocol to overcome the even greater lack of cpu grunt
compared to the Toppy (I believe both machines were based on the same
reference design where the Toppy designers elected to use a more
powerful CPU - the evidence certainly supports this 'extra cpu grunt'
hypothesis).

 When you consider that these 'under powered' boxes can achieve what
none of the contempory PC systems could do, it _really_ brings it home
to me what a shite interface USB2 is.

 If the manufacturers had left out the USB interface altogether and
replaced it with a fast ethernet port instead, they would have had an
irresistable product that could have easily transferred the data at a
non-taxing 10MB/s second instead of struggling to get a mere 1 or 2
MB/s out of the USB port (and without any need for a "turbo mode"
kludge solution either).

 It's this knowledge, whenever I have to leave the transfer job
running for a half hour or more just to get a couple of hours worth of
recordings transfered to the laptop (which would normally only take 4
or 5 minutes between an external USB connected HDD and a PC), that
reminds me of just how shitty the USB concept is in its second and all
too ubiquitous incarnation.

 I have every reason to remain totally unimpressed by Intel's USB2
'bodge' (only a 'beancounter' could be impressed by Intel's
'Chutzpah').
-- 
Regards, J B Good

[toc] | [prev] | [next] | [standalone]

#1879 — Re: USB interface

On 3/21/2014 7:41 PM, Johny B Good wrote:

>  Well, I pretty much concur with most of what you've said except I
> think USB 1 came to be abused by those 'feckin' cheap and nasty
> external USB dial up modems (amongst other "USB Gadgets") which now
> demanded a Pentium 166MX chip as a minimum (then, latterly P233MX).
> 
>  A  33MHz 386SX could have easily done the job with a standard (but
> more expensive) serial port connected 56K V90 modem. I'm pretty sure
> this sort of abuse was being practised prior to the advent of USB2
> (please correct me if I'm wrong).

My recollection (perhaps wrong) is that by the time of 386s, serial
ports mostly had UARTs that could keep up, and such ports had become
standard. So it's not the expense of the port that was the issue.

Such abuses appeared even before the appearance of USB, with the
introduction of the internal PCI bus. That was soon followed by
internal Windows-only modems generically known as "winmodems", which
reduced part count (and cost) by offloading a lot of the work to the
CPU. They were cheap crap. But since they did, in fact, cost less,
they quickly took much of the market.

When USB came in, it was attractive to people who didn't want to open
the case to install the modem. But I never did see many USB modems,
although cheap DSL boxes often did provide for a USB connection, as it
was not a given that there would be a Ethernet connection on any given
computer.

[toc] | [prev] | [next] | [standalone]

#1880 — Re: USB interface

On 2014-03-22 03:44:18 +0000, Dave Garland said:

> On 3/21/2014 7:41 PM, Johny B Good wrote:
> 
>> Well, I pretty much concur with most of what you've said except I
>> think USB 1 came to be abused by those 'feckin' cheap and nasty
>> external USB dial up modems (amongst other "USB Gadgets") which now
>> demanded a Pentium 166MX chip as a minimum (then, latterly P233MX).
>> 
>> A  33MHz 386SX could have easily done the job with a standard (but
>> more expensive) serial port connected 56K V90 modem. I'm pretty sure
>> this sort of abuse was being practised prior to the advent of USB2
>> (please correct me if I'm wrong).
> 
> My recollection (perhaps wrong) is that by the time of 386s, serial
> ports mostly had UARTs that could keep up, and such ports had become
> standard. So it's not the expense of the port that was the issue.
> 
> Such abuses appeared even before the appearance of USB, with the
> introduction of the internal PCI bus. That was soon followed by
> internal Windows-only modems generically known as "winmodems", which
> reduced part count (and cost) by offloading a lot of the work to the
> CPU. They were cheap crap. But since they did, in fact, cost less,
> they quickly took much of the market.
> 
> When USB came in, it was attractive to people who didn't want to open
> the case to install the modem. But I never did see many USB modems,
> although cheap DSL boxes often did provide for a USB connection, as it
> was not a given that there would be a Ethernet connection on any given
> computer.

Even before Winmodems were a thing, there were inkjet printers which 
were basically a motor and some nozzles, with all the logic controlled 
from the PC.  No in-printer rasterizer, just a stepper motor and some 
ink droplet control.  I seem to recall the actual name was the 
'Winprinter.'

[toc] | [prev] | [next] | [standalone]

#1882 — Re: USB interface

On 2014-03-22, Oregonian Haruspex <bob_davis_retired@yahoo.com> wrote:
> On 2014-03-22 03:44:18 +0000, Dave Garland said:
>
>> On 3/21/2014 7:41 PM, Johny B Good wrote:
>> 
>>> Well, I pretty much concur with most of what you've said except I
>>> think USB 1 came to be abused by those 'feckin' cheap and nasty
>>> external USB dial up modems (amongst other "USB Gadgets") which now
>>> demanded a Pentium 166MX chip as a minimum (then, latterly P233MX).
>>> 
>>> A  33MHz 386SX could have easily done the job with a standard (but
>>> more expensive) serial port connected 56K V90 modem. I'm pretty sure
>>> this sort of abuse was being practised prior to the advent of USB2
>>> (please correct me if I'm wrong).
>> 
>> My recollection (perhaps wrong) is that by the time of 386s, serial
>> ports mostly had UARTs that could keep up, and such ports had become
>> standard. So it's not the expense of the port that was the issue.
>> 
>> Such abuses appeared even before the appearance of USB, with the
>> introduction of the internal PCI bus. That was soon followed by
>> internal Windows-only modems generically known as "winmodems", which
>> reduced part count (and cost) by offloading a lot of the work to the
>> CPU. They were cheap crap. But since they did, in fact, cost less,
>> they quickly took much of the market.
>> 
>> When USB came in, it was attractive to people who didn't want to open
>> the case to install the modem. But I never did see many USB modems,
>> although cheap DSL boxes often did provide for a USB connection, as it
>> was not a given that there would be a Ethernet connection on any given
>> computer.
>
> Even before Winmodems were a thing, there were inkjet printers which 
> were basically a motor and some nozzles, with all the logic controlled 
> from the PC.  No in-printer rasterizer, just a stepper motor and some 
> ink droplet control.  I seem to recall the actual name was the 
> 'Winprinter.'

I remember these being called "CGI printers" which I think was something like
Computer Generated Images, though might be Imagery there?  Not sure after all
this time but I made a conscious effort to avoid these.

<http://en.wikipedia.org/wiki/Computer-generated_imagery>

My memory is saying cheap laser printers as well as inkjets, running
off parallel ports in the days before USB came along.

-- 
Paul Sture

"In that case," said Napoleon, "let us wait twenty minutes; when the enemy
is making a false movement we must take good care not to interrupt him."

[toc] | [prev] | [next] | [standalone]

#1898 — Re: USB interface

On 3/22/2014 3:46 AM, Paul Sture wrote:
> On 2014-03-22, Oregonian Haruspex <bob_davis_retired@yahoo.com> wrote:

>> Even before Winmodems were a thing, there were inkjet printers which 
>> were basically a motor and some nozzles, with all the logic controlled 
>> from the PC.  No in-printer rasterizer, just a stepper motor and some 
>> ink droplet control.  I seem to recall the actual name was the 
>> 'Winprinter.'
> 
> I remember these being called "CGI printers" which I think was something like
> Computer Generated Images, though might be Imagery there?  Not sure after all
> this time but I made a conscious effort to avoid these.
> 
> <http://en.wikipedia.org/wiki/Computer-generated_imagery>
> 
> My memory is saying cheap laser printers as well as inkjets, running
> off parallel ports in the days before USB came along.

My printer today (an HPLJ6P that will still be cranking out hard copy
long after I'm dead) runs off a parallel port, you insensitive clod!

/.

[toc] | [prev] | [next] | [standalone]

#1893 — Re: USB interface

On Sat, 22 Mar 2014 00:49:18 -0700, Oregonian Haruspex
<bob_davis_retired@yahoo.com> wrote:

>On 2014-03-22 03:44:18 +0000, Dave Garland said:
>
>> On 3/21/2014 7:41 PM, Johny B Good wrote:
>> 
>>> Well, I pretty much concur with most of what you've said except I
>>> think USB 1 came to be abused by those 'feckin' cheap and nasty
>>> external USB dial up modems (amongst other "USB Gadgets") which now
>>> demanded a Pentium 166MX chip as a minimum (then, latterly P233MX).
>>> 
>>> A  33MHz 386SX could have easily done the job with a standard (but
>>> more expensive) serial port connected 56K V90 modem. I'm pretty sure
>>> this sort of abuse was being practised prior to the advent of USB2
>>> (please correct me if I'm wrong).
>> 
>> My recollection (perhaps wrong) is that by the time of 386s, serial
>> ports mostly had UARTs that could keep up, and such ports had become
>> standard. So it's not the expense of the port that was the issue.
>> 
>> Such abuses appeared even before the appearance of USB, with the
>> introduction of the internal PCI bus. That was soon followed by
>> internal Windows-only modems generically known as "winmodems", which
>> reduced part count (and cost) by offloading a lot of the work to the
>> CPU. They were cheap crap. But since they did, in fact, cost less,
>> they quickly took much of the market.
>> 
>> When USB came in, it was attractive to people who didn't want to open
>> the case to install the modem. But I never did see many USB modems,
>> although cheap DSL boxes often did provide for a USB connection, as it
>> was not a given that there would be a Ethernet connection on any given
>> computer.
>
>Even before Winmodems were a thing, there were inkjet printers which 
>were basically a motor and some nozzles, with all the logic controlled 
>from the PC.  No in-printer rasterizer, just a stepper motor and some 
>ink droplet control.  I seem to recall the actual name was the 
>'Winprinter.'

 External usb winmodems, usb adsl modems and those 'dumb' printers are
all excellent examples of "USB Abuse" that clearly demonstrated the
shortcomings of USB.

 If such abominations hadn't been invented, USB could have been better
appreciated for what it was 'supposed' to replace, i.e. those special
multi serial port adapters with breakout box for use in office
environments making use of a few dozen dumb terminals to reduce
equipment costs by effectively mimicing a "MainFrame + dumb terminal
set up" in miniature (they actually existed - I saw one such set up in
a GP surgery about 15 years ago!).

 For a home user needing to control 3 or more peripherals via serial
port connections, it represented an improvement over the classic com
port solution. The problems arose when the bandwidth requirements of
up to 63 serial ports was aggregated into one super channel for such
kit as winprinters and winmodems, often requiring additional cpu
cycles to do the donkeywork normally handled within a standard printer
or modem.

 Back in the day, around 15 years ago when Quake2 and Unreal
Tournament servers started appearing on the internet, the early NTL
cable modems included a USB socket as well as a fast ethernet port.

 Any self respecting gamer would add a cheap network card rather than
put up with the cpu sapping built in USB ports in order to avoid being
disadvantaged. The difference on gaming PCs with marginal CPU grunt
when playing networked games over the internet was like night and day.

 USB2 had to wait for Pentium 2 cpus before it became a practical
reality (cacheless pentium 2s aka pre-Celeron300As) just couldn't cut
it - they were notably less powerful than the Pentium 233MMXs that had
preceded Pentium 2).

 Intel weren't about to step in and put a stop to the nonsense
represented by the manufacturers of these winmodem, winprinter
abominations taking advantage of the consumers' desire to "Have a
bargain at any cost" since it simply wasn't in their best interest to
interfere with a process that ensured an even greater demand for their
latest cpu product (the 'any cost' part of that consumer equation).

 It's been a good 5 years or more since most PC users have come into
possession of a suitable 'sledgehammer' (sheer brute force quantities
of CPU cycles) to crack the USB2 flavoured 'nut', so the question of
USB2's shiteness has been largely rendered academic now on reasonably
current PCs.

 The original pretense that USB would provide a superior alternative
way to connect a bunch of low bandwidth peripherals that would
otherwise have used a com port each has long been swept aside (USB2
certainly saw to that) so the 3rd generation of even faster USB is
such a boost in speed that it can only work using proper DMA mediated
data transfers, no amount of CPU performance boost is going to
overcome the PIO bottleneck.

 Intel have had no choice in this particular matter. Intel aren't
'doing the right thing' this time around because there simply is no
cheap alternative to allow them to 'do the wrong thing' as they had
done with USBs 1 and 2.

 I'm not familiar with the specifics of USB3 but it looks like there
might be an opportunity to utilise the benefit of DMA in the backwards
compatability support (for usb2 at least) which should provide an
improvement with older USB2 peripherals, notably USB2 only external
disk drives. This is one item of improvement I can hang fire upon
until it appears on my next major upgrade a year or so from now.
-- 
Regards, J B Good

[toc] | [prev] | [next] | [standalone]

#1781 — Re: Software Bloat

On 2014-03-18, Oregonian Haruspex <bob_davis_retired@yahoo.com> wrote:
> If there is ever a Nuremburg-type trial at which the Excel developers
> are defendants, the judge will not accept the defense that they were
> just following orders.
>

How about the guys who imposed Microsoft's unloved "Ribbon" toolbar on
us?  I consider myself a power user, but my efficiency was crippled by
that stupid "innovation."

[toc] | [prev] | [next] | [standalone]

#1783 — Re: Software Bloat

On 2014-03-18 22:07:04 +0000, RS Wood said:

> On 2014-03-18, Oregonian Haruspex <bob_davis_retired@yahoo.com> wrote:
>> If there is ever a Nuremburg-type trial at which the Excel developers
>> are defendants, the judge will not accept the defense that they were
>> just following orders.
>> 
> 
> How about the guys who imposed Microsoft's unloved "Ribbon" toolbar on
> us?  I consider myself a power user, but my efficiency was crippled by
> that stupid "innovation."

The woman who foisted that piece of crap off on the world was the same 
person who was in charge of the Windows 8 user interface.  I'd have her 
cleaning toilets if I was in charge at MS, but then again if I was in 
charge at MS I'd have closed down operations years ago, auctioned off 
the assets, and paid back the shareholders for the public good.

[toc] | [prev] | [next] | [standalone]

#1795 — Re: Software Bloat

On Tue, 18 Mar 2014 15:38:46 -0700, Oregonian Haruspex
<bob_davis_retired@yahoo.com> wrote:

>On 2014-03-18 22:07:04 +0000, RS Wood said:

>> How about the guys who imposed Microsoft's unloved "Ribbon" toolbar on
>> us?  I consider myself a power user, but my efficiency was crippled by
>> that stupid "innovation."

>The woman who foisted that piece of crap off on the world was the same 
>person who was in charge of the Windows 8 user interface.  I'd have her 
>cleaning toilets if I was in charge at MS

CLOGGED toilets.

All day, and nothing but.

I might even hire some "big eaters" just to go around the campus to
make sure she has enough "work" to keep her busy all day.

(This may be why I am not on charge of Microsoft...)

[toc] | [prev] | [next] | [standalone]

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

Back to top | Article view | comp.misc

csiph-web