| Tuesday, May
7, 2002
This is it: the column has to be
done tonight. Begin with this:
Dr. Pournelle Although I found these through
Slashdot I still thought that you and your readers might enjoy this. Mr:
Edgar Villanueva Nuñez Congressman of the Republic of Peru is proposing a
bill that would make *free* software the choice for the government of
Peru. Microsoft sent him a letter to try and persuade him (termed by the
Slashdot audience as FUD) not to implement this measure.
The Microsoft letter is at http://www.pimientolinux.com/peru2ms/alt_ms_to_villanueva.html
His response is very informative and reasoned about
what *free* software is and isn't. http://www.pimientolinux.com/peru2ms/villanueva_to_ms.html
P.S. While I'm not an supporter of Richard Stallman,
I wish that our Congress Critters were more like Mr. Nuñez and less like
Senator Hollings.
Sincerely John Weidman john.weidman@atk.com
I will comment on this another time.
Meanwhile, on USB:
Jerry:
My ears are still ringing from David S..'s - uh -
frank expression of his views about USB2.0 last Saturday. I followed his
link to the 3-year-old Apple Royalist page of the same topic and tone (or
even more strident). Perhaps I can add some data into the conversation.
Your USB1 controller has 12Mbps of bandwidth, coming
out of (usually) 2 ports of your root hub. To share that among more
devices, you add external hubs. I don't know exactly what a "simple
hub" is supposed to be; the activity of a USB1 hub is quite complex,
but it's robust because it's been beat up real good in design and in the
field; and it's cheap because they make so many of them. Once the hub has
been instructed (like a router?) to include one its ports in the bus, that
device shares the 12MBps. If it's a WebCam, that can hog up most of it.
If you want more bandwidth, you can add another host
controller and the port(s) of its root hub. There's another bus of 12Mbps.
If a low-speed device [mouse] is sharing the bus, it's taking 8 times as
much of that time for each bit as a full-speed device [printer] does. The
article to which David linked says, and illustrates, that this waste
balloons to 320x with the 480Mbps bandwidth available from USB2.
"Unfortunately, that turns out not to be the
case." Here is how it actually works:
A USB2 controller provides USB2 bandwidth among 1 or
more ports on its root hub. However, if a USB1 device (including a simple
USB1 hub) is plugged into one of those ports, it is connected to a
full-speed Open Host Controller Interface (USB1) controller, built right
into the Enhanced Host Controller Interface (USB2) controller by spec. In
fact, an EHC can have one OHC for each port - and some of them do.
If a (stupid router-like) USB2 hub is connected to a
USB2 port, it shares the USB2 bandwidth. Data always moves at that speed
over USB2 connections. If it is destined for a USB1 device, it's packaged
in envelopes called "split messages" until it hits the last port
of the USB2 tree, and is decompressed into USB1 speed. (Somehow that 40x
improvement becomes a liability, because it doesn't make the old equipment
run any faster. Is that the Revolution of Rising Expectations?)
So to recap: USB2 devices get 40x speed improvement.
A USB2 hub has a full USB1 tree growing out of each port, so your 15
WebCams can all run full-bore (ahem) if your processor, drivers, and apps
can support them. (Incidentally, EHCI offloads much more
schedule/buffer/list work from the host processor onto the controller HW,
compared with the OHCI of 1394 and some USB1 host controllers.)
The original mission of USB was to support a
hot-plug tree for peripherals both medium-fast and slow. (A scanner is
SLOW???) I'm probably too dumb to understand all the problems David sees
in speeding it up to handle more data. The inner workings are another
matter, of course; but what's under the hood of modern automobiles looks
hopeless to most people who can select, drive, and reasonably maintain
them. Users don't need to mess with USB firmware and drivers.
If you're curious about USB, in addition to the
valuable insights offered by Apple enthusiasts, you can go to www.usb.org
and look at their overviews (and look up prices of hubs). The innards are
behind the "developers" tab. Then you'll know enough to feel
fear, uncertainty, and doubt about USB2.0.
Or not.
-- Bill Kilner
P.S.: Jerry, you said USB on laptops is a problem
due to power requirements. Is the USB absorbing that power, or is it the
habit of making peripherals bus-powered, where they would have needed
wallbugs before? If so, while they are frustrating with laptops now,
before USB they would have been simply unavailable.
Thanks.
As usual, the whole FireWire vs USB 2.0 thing boils
down to money, plain and simple:
http://news.com.com/2100-1040-272147.html?legacy=cnet
In short: Apple and others were wanting to charge
$1/port for FireWire. Intel and other PC makers balked and came up with
USB 2.0. Apple and company saw the light and came up with $0.25/system,
which most everyone can live with. USB 2.0 is still in the pipeline
though, and it'll still cost less than FireWire, which means it'll have
traction regardless of how technically superior FireWire is.
Pete Flugstad
The latest high speed motherboards
from Intel have USB2 on the motherboard. As to technical superiority, at
what point are things Good Enough? And
You have been sent this message from a
washingtonpost.com user as a courtesy of the Washington Post (http://www.washingtonpost.com).
To view the entire article, go to http://www.washingtonpost.com/wp-dyn/articles/A39830-2002May6.html
U.S. Pulls Out of International Court Treaty
By Peter Slevin
The United States has renounced formal involvement
in a treaty creating the first permanent war crimes tribunal, Ambassador
Pierre-Richard Prosper said Monday.
Hurrah! I'll have more to say on
this another time.
Jerry,
Saw the posting about the President pulling us out
of the International Criminal Court, and thought I'd send along a couple
of reasons (okay, several reasons) why the ICC is A Bad Thing. These are
not the reasons usually cited by the administration, but are good reasons
to stay away from the ICC even if the administration's public concerns
were addressed.
1) Victim-supported prosecutions. A cornerstone of
modern civilization is state justice over personal justice. A detached,
professional prosecutor screens complaints and brings into court only
those cases where there is legitimate evidence of wrongdoing. This is the
thrust of every modern, fair legal system I am aware of.
The ICC permits victims and victim-advocacy
organizations to underwrite the costs of prosecution. Victims and their
advocates may pay for additional prosecuting attorneys, may sit at the
prosecutor's table in court, and may work closely with the ICC's permanent
prosecutorial staff in prosecuting cases.
In a U.S. court, that is not supposed to happen. For
example, the U.S. Attorney's office did not send Adobe Systems a bill for
filing charges against Dmitry Skylarov. It's a question of fundamental
fairness, and the ICC is a "stacked deck" against the accused
from day one.
2) Broad definitions of crimes. The ICC statutory
definition of genocide is broad enough to cover speech, as well as action,
aimed at any identifiable group. By the statute, genocide includes
"causing serious bodily or mental harm" with "intent to
destroy, in whole or in part, a national, ethnical, racial or religious
group, as such."
In other words, proselytizers could be liable for
genocide. Why? Well, clearly, a proselytizer targets an identifiable
religious group--those not of his religion. He wishes them all to join his
religion--thus "destroying" their old religious body.
Stretching it? Maybe. But considering the heavy
European influence on the Court, and Europe's continued persecution of the
Church of Scientology and the Unification Church, not so unbelievable, eh?
(Full disclosure--I'm a Mormon, so of course I have an interest here . .
.)
3) No free speech protection. Consider that much
speech "causes serious mental harm." In the United States, you
can plead the First Amendment to get out of prosecution unless you were
directly inciting people to violence. There is no First Amendment in the
Rome statute. You can be convicted of genocide for nothing more than
words, even if you did not incite others to action.
Not only is my hypothetical proselytizer at risk; so
too are rabble rousers, race-baiters, and all kinds of people who are,
frankly, jerks, but in the US would have free speech rights. In the ICC,
they will not.
4) Supremacy of the court. The claim is made that
the court only steps in when local courts refuse to prosecute. This
supposedly protects national sovereignty. In fact, who decides whether
local courts are refusing? That's right, the ICC itself. The ICC by treaty
and statute is legally empowered to override the decisions of local
tribunals. So if our intrepid proselytizer is an American, and the
American prosecutors refuse to prosecute because proselytizing is
protected by the First Amendment, the ICC can take custody and try the
prisoner.
5) Purported universal jurisdiction. The ICC statute
purports to have jurisdiction over an accused if (a) the accused is a
national of a ratifying state, or (b) the crime affected somebody in a
ratifying state. So, suppose that our intrepid proselytizer puts up a web
page, in the USA, hosted on a US-based ISP, preaching his personal gospel.
And suppose somebody in a signatory state reads that page and is offended
by it--"seriously mentally harmed" even. Boom, the ICC now
claims to have jurisdiction despite the President's repudiation. (Note:
the only crime for which nations have historically exercised universal
jurisdiction was piracy. There is no justification in international law
for this treaty's sweep.)
I'm glad the President repudiated it. However, we
may eventually find that more serious action is required to convince the
Court not to prosecute our citizens. Force often works.
Steve Setzer
And that is just for openers.
And it's clear I don't known
EVERYTHING...
I take it from that comment [in the current column]
that you knew TECO on the PDP/8. Too bad. Apparently it was a well kept
secret that TECO was a magnificent full screen, GUI based editor when run
on it's native platform, the PDP/12. The '12 was a hybrid machine with a
PDP-8i and a Link-8 spliced together with mode-shift instructions that let
you switch from one instruction set to the other. The LINK had a 15 inch
graphics scope (X/Y D-A driven) and A/D channels built in. TECO threw text
on the screen and would do cursor position from the A/D lines. Sometime
about 1974, I replaced the teletype with a Keytronics keyboard, adding a
couple of high quality wire-wound pots on the top. The thing worked at
least as well as most of the mush-pads we suffer with on today's
notebooks.
The TECO version that DEC shipped with OS/8 had all
the LINK instructions still active. It just ran slow as the PDP/8 just
NOP'd the Link instructions. Someone once commented to me that half of the
PDP/12 production was used internally at Maynard. The people doing system
development on the PDP/8 new a good thing when they saw it. It was always
amusing to see the look on the face of a PDP/8 TECO user when they saw it
running on the '12.
Whoever it was at OMSI (http://www.omsi.edu/) that
wrote TECO deserves credit toward putting us where we are today with
graphical editors.
Dan Walker
That's really all I did see TECO on.
Bob Forward told me once that TECO was wonderful, and he wrote his first
novel on it, but then he was at Hughes where they had serious iron, and I
only had old Zeke the Z-80...
TOP
CURRENT
VIEW |
CLICK HERE for
a Special Request. 
