Grass roots VOIP - telephone network?
on
VoIP at $15 a Pop
·
· Score: 2
[...] OpenH323 has [...] a VoIP -> PSTN gateway [...]
Perhaps a bunch of volunteers or entrepreneurs could set up home servers to allow incoming H323 connections to make local phone calls.
There was an effort to allow for free fax transmissions this way a few years ago. It used
DNS as the mechanism for keeping track of which
servers could make local calls to which phone number prefixes. However, I haven't heard about that project in a long time.
AT command to dial phone and hold line open
on
VoIP at $15 a Pop
·
· Score: 2
There must be a million little utilities for generating touch tones, but you probably don't
even have to bother with that. If your computer
is connected to the phone line by a modem that understands the AT command set, ending the phone number with a ";" will tell your modem to just dial a phone number and leave the line open without trying to negotiate a data connection. For example, "ATDT16505551212;" will dial 1-650-555-1212 and hold the line open. "ATH" will hang up the line.
Open Source access to Innosphere network?
on
VoIP at $15 a Pop
·
· Score: 2
Perhaps Innosphere would actually be supportive of open source and other third party software to connect to their network. They are, after all, just charging for use. The more usage the network gets, the more money they should make. At only $0.15/minute to call China, I'd sure like to use their service.
G.723.1 voice compression is apparently covered by
software patents.
I assume the VOIP blaster comes with a license to run G.723.1 inside the device, but these patents would impede the development of free software that could use the data stream from VOIP blasters for purposes other than talking to other G.723.1 hardware devices.
It sure would be nice if the VOIP blaster had a mode where it could just transmit and receive raw audio samples (preferably by the standard USB audio class interface).
I did not say that authors would make more money with a shorter copyright. I said the public would be better off.
The part of your response that addresses my thesis that the public would be better off with a five year copyright is the scenario of a work being "discovered" much later. You need to argue that this scenario occurs with statistically significant frequency, not just an anecedotal case that apparently came about by a legal error anyhow, and that the incentive of a longer monopoly actually effects authors' production. The second part is particularly important. For example, if most of the works that are "discovered" later are the early works of authors that are now popular, then you are only talking about rewarding authors if and when they are already making a lot of money from their success.
By the way, a five year copyright expiration might even make an author more likely to be "discovered" in time to hit the big time with his or her next book while he or she is still has enough years to enjoy the money. This would be against the publisher's interest, since there is no guarantee that the author will use the same publisher. So you won't see a lot of release from five year copyright releases without a change in copyright law, since scales of efficiency will result in a small number of high volume publishers and a large number of authors, giving publishers greater negotiating power with unknown authors.
At least in the United States, copyright's
purpose is to promote the progress of science and the useful arts. If book publishers want to renegotiate the copyright bargain, then we should take a look at moving the balance in other direction, that of reducing the scope and duration of traditional copyright.
Computers have greatly reduced the time involved in writing, editing, typesetting and printing books since the days of writing a book with a typewriter. Distribution, sales and shipping of books have also been accelerated by technology (printing in more than one location, nearly realtime sales information across entire store chains, etc.).
Technology also means that the opportunities that copyright impedes have greatly increased. Being able to freely copy material online means that many people do not have to chop down trees to store information. Physical storage of books in digital forms is much more compact. Searching and sharing of free online information is orders of magnitudes easier.
There is even a secondary opportunity cost to authors in long copyrights: the development of derivative works is greatly limited by copyright when they are outside of "fair use." For example,
I think that, given how much time has elapsed, Richard Hatch should be allowed to make his
Battlestar Galactica sequel, and the
rewrite of Gone with the Wind from a black
perspective (The Wind Done Gone) should be allowed whether or not the book qualifies as a parody. The opportunities lost by impeding this sharing are increased when the efficiency with which these derivative works can be made is increased (i.e., more potential derivative works that otherwise would be produced are lost during each year of the copyright).
The costs of creating a book have dropped. The rate at which that investment can be recovered has accelerated, and opportunities that we lose during each year of copyright have increased. In my view, the balance point at which the public benefit of copyright is maximized has been greatly reduced.
I believe that it would maximize public benefit to
accelerate copyright expiration to about five years, maybe even less.
The web pages for djvu talk about the GPL'ed version including a decoder and a "simple bitonal (black-and-white) encoder" and a "very simple color encoder." This makes me wonder how the free encoders are and how they compare in compression and speed to other free encoders,
such as those for jpeg and png.
A detailed positive answer to this question might make it worthwhile to more people to risk a couple hundred dollars of their time to try to install DjVu.
Oops. I meant to links to existing
products based on IBM's 20.8" 2048x1536 ITQX20 module. Here they are.
Raintree Systems
IN 2080-50 and
National Display's
Nova have the standard IBM dual LVDS digital inputs. I think there was an outfit in the UK called "Gemini Electonics" that was going to produce a similar device, but I could not find a link to them right now. RealVision has a dual LVDS
video card for driving these monitors, although they
promote it more for the electrically identical 6144x1536 grey scale version of the ITQX20. Finally, IBM's
T-210 apparently uses the ITQX20 module, but only allows analog input, and only at something like 30Hz if you want to full resolution.
I have no financial relationship to any of these vendors.
IBM has had a ~$2k
20.8 inch 2048x1536 ITQX20 TFT display with very deep blacks
and great viewing angle for about three years already. I have seen it in real life. It is very nice. Unfortunately, everyone who makes a monitor around this card wants ~$6k for the monitor and video card. This is understandable since quantities are low at that price point and it is necessary to stock and ship a bunch of these displays.
What I wish someone would make is a kit consisting of everything except display module, which one could buy directly from IBM. That is: a plastic housing for the display, a power supply and a video card. The video card needs to be driven by 4 24-bit LVDS transmitter chips, which were $8 each when I looked in it--I think the chip may have beeen the
National Semiconductor DS90CF581. 4 x $8 = $32 chip cost x 5X markup for retail = $160 additional cost of video card plus cost of two LVDS connectors (using the hokey rule of thumb that I've heard that total electronic component costs are typically 20% of the suggested retail prices of the resulting product).
Without the need to stock the display component, there would not have to have such a big mark up to cover storage, damage, etc. I think the kit without the display could easily cost under $1k.
The video card would be the significant engineering task. The two LVDS streams have to be kept in sync, so you cannot just use two of the LVDS cards that can drive an SGI-1600SW display. It looks like you really do have a make a new video card. The big question that I had not gotten around to researching was whether were was a VGA chip or chip set that could deliver the digital output for two screens without convering the signal to analog,that is deliver two 24-bit parallel streams (the display interface is basically that of two 1024x1536 LVDS flat panels, side by side). There are a number of dual head
VGA cards. I just never got around to looking into
whether it was possible to get digital output from their chips.
I think that if a kit like this were available, some computer retailers would assemble it and the panels to offer the finished product, and that would reach untapped section of the flat panel market that I think there should be significant deamnd for (a $3k 20.8" TFT for CAD, engineeringing, graphic art, etc.).
One minor drawback that might slightly impede the popularity of
this display as it gets closer to the consumer
range is that the housing for it is currently not
very thin. The housing is about four inches thick, making it a bit less sexy looking from the side than most other flat panels.
I am not a lawyer, so don't rely on this as legal advice.
Generally, my experience with DSL providers has been that they do not prohibit this kind of sharing. Indeed Covad's sales staff touts this kind of unrestrictedness as selling point. (I called Covad's sales line to confirm this a while ago.)
This is not so surprising when you consider that the places where wireless service is expected to be most useful are "hot spots" where people gather and sit down, such as resturaunts, coffee shops, retailers, and various waiting areas, typically places that are likely not to order broadband otherwise.
It is difficult to prove a negative. So, maybe
those who believe their acceptable use policy prohibits this should provide relevant excerpts. It would be interesting to see if those AUP's also, with the same stretch of interpretation, prohibit attaching a wireless access point, using Network Address Translation to connect more than one computer, or using the connection for work.
By the way, my impression is that DSL people seem to be more positively disposed toward this sort of thing than the cable modem providers.
Disclaimer and plug: I am involved in
LANRoamer,
a GPL'ed wireless roaming network that allows people to get paid for providing wireless service.
The back end is also GPL-compatible open source.
Dr. Beltrami and his colleagues from the University of Michigan found that more than half of the land's heat gain over the past 500 years came during the 20th century, and 30% since 1950.
Sorry, I blew it, so I'll publicly take responsibility for my error. As the responses by "dhogaza" and "PhuCknuT" correctly point out, I misread that sentence. Beltrami et al claim that 30% of the temperature increase over that past 500 years happened since 1950, not just 30% of the temperature increase since 1900. Although I obviously puzzled over that sentence for before posting about it, I will endeavor to be more diligent against making this kind of mistake in the future. Sorry for wasting everyone's time with a simple misreading.
Dr. Beltrami and his colleagues from the University of Michigan found that more than half
of the land's heat gain over the past 500 years came during the 20th century, and 30% since 1950.
So, they believe the rate of warming for 1951-2000 was less than half what is was for 1901-1950. I don't have much basis for an opinion on the meaningfulness of these researchers' results, but I would sure like to know how they explain this apparently levelling off.
A more cost effective strategy may be to promote the political careers of those who agree with us. (Granted, this is not a mutually exclusive with basically buying votes. It's just the other end of a continuum.)
We should actively help the careers of any politician who agrees with us, even if that allegiance would only be useful if they achieve some higher office later or can trade a favor.
Although money helps a campaign (I have contributed to EFF), money isn't necessarily the only currency. Campaign money eventually buys media and staffing. These are things which can be contributed in kind. To those sympathetic slashdotters who have time but not money and want to help: maybe you could find some college students who want to volunteer to help Rick Boucher's reelection campaign, finagle some ad banners, answer email, etc. On a larger scale, maybe you could make a web site to collect information on who is our friend, and to connect large numbers of volunteers with campaign organizations. That might be helpful enough to an election to make our support worth something in terms of election results.
Here is some
additional VNC-related software that you may find useful. It comes from a project at Wyse a couple of years ago.
vncd spawns VNC X-windows sessions for incoming connections. It can be run from inetd or by itself. This is handy if you to deploy a bunch of thin clients around a Linux system. It also has a little protocol for negotiating things like display geometry and color depth. If you just sent "DONE" following a line feed to it, it will then start the VNC protocol. See the source code for commands for setting display geometry, color scheme and environment variables in the session.
remote-audio is a client and server for having audio programs send their output to a remote client. On the Linux machine that you are logged into, it intercepts the open() symbol in the C library to catch attempts to open the/dev/dsp audio device, and then it serializes the I/O to a your display device, which runs a server to accept connections for this purpose. There is a facility for a simple password check in this protocol, but there is currently no encryption in it, so you probably want to firewall it and only access it from outside of your firewall by a VPN scheme (or extend this software).
vnc-3.3.3.patch is source code for some VNC optimizations, such as local cursor, gzip compression of the link, special encoding for 1x1 rectangles (which, if I recall correctly, were nearly 50% of rectangles in some tests).
Wyse also shipped a "regular" X server that ran on an X terminal that could also accept incoming VNC connections to allow remote operation of that terminal. I believe the product was called WT5000. I started working on putting this into XFree86-4.2 and I've put a source snapshot
here,
but I have no idea if it works (it adds a "-vnc" argument to the X command line to allow incoming VNC connections).
LANRoamer and NoCatAuth appear to have started around the same time.
The two people who started NoCat gave a talk at
the Bay Area Wireless User Group about a week after they started development, and I talked
to them there. They (or at least one of them) said that they knew about LANRoamer when they started but thought that the LANRoamer back end was proprietary (we had publicly released it as free software by that time, but there was a period of about two weeks from when we announced LANRoamer to when we decided to free the back end, so I understand how they got that impression).
LANRoamer is a GPL'ed system
that has been doing this for a while. We gave presentations on it at
Bay Area Wireless User Group and
Sbay.org back in June, I believe, before even the NoCat project
started.
If you're into "bazaar" style software development, one thing you should note is that
LANRoamer does network booting and upgrade reboots.
So, if you contribute a useful feature to LANRoamer, it can be widely deployed quickly
(based on our stability labels and the
stability level each gateway owner has selected).
Also, in addition to free accounts and revenue sharing to our access point providers, we also offer free courtesy accounts for people who run open access points (not just during a free beta), partly in an effort to thank the developers and "evangelists", but also to get them involved.
The network boot floppy currently requires that the first ethernet
card be compatible with 3COM 3c59x, 8139too, Ether Express Pro
100, NE2000 PCI cards, Via Rhine, Tulip cards and PC-Net PCMCIA
ethernet (the 802.11 card or the ethernet connection
to your access point can be just about any card that Linux supports).
Unlike NoKat (the last time I checked), LANRoamer can work behind firewalls, including NAT
routers, even ones that distribute IP addresses that LANRoamer would
otherwise use. Once your gateway is up, client machines can obtain
addresses from your wireless gateway by DHCP and are taken to an
SSL-based login page when they try to go anywhere on the web until
they log in.
Obfuscated translation is copyright infringement
on
Abusing the GPL?
·
· Score: 3, Interesting
I'm not a lawyer, so don't use this as legal advice. Instead, you (the author of this slashdot article) may want to show it to your company's lawyer and suggest that he track this down.
According to
this link, there is a case called "Whelan" that established that duplicating the detailed structure of a program was copying of expression rather than ideas, and therefore copyright infringement.
Also, I remember reading a very good article about ten years ago by law professor Pamela Samuelson, I think in Communications of the ACM or some other ACM publication, that talked about this decision and mentioned "detailed structure and flow", which would make the case for infringement even stronger.
Finally, I recall reading somewhere, perhaps in that same article, that there is some common law rule that the standard of similarity by which copyright infringement should be determined is supposed to correspond to how much access the alleged infringer had to the original work. In other words, if the alleged infringer had easy access to the original work (e.g., had carefully read the original GPL'ed source), then the standard for proving infringement is supposed to be easier.
Again, I'm just a layman. Don't use this as real legal advice.
If anyone is thinking of making a campaign
contribution,
you might want to look at whoever runs against
Ted Stevens in his November 2002 senate reelection in Alaska, for the following reasons.
While most
Replublicans seem skeptical of SSSCA, Ted Stevens
cosponsored it. Without
Stevens' support, this kind of bill will seem much more partisan, reducing its chances unless the democrats manage to get control of both houses of congress and the white house.
Alaska has a small population, so a little
money should go a long way in an election. Stevens
has
$1.5 million, including PAC contributions, an amount within the fund raising capabilities of a bunch of slashdot's wealthier nerds, less than half of the
1996 national average of
$3.6 million.
Also, Stevens might not get critical support from his fellow Republicans if he finds himself in need, as he has
embarassed his party with his pork barrel politics, netting Alaskans $766 per capita of federal money (making the state #1 in this category), in comparison to the national average of $25.52, as calculated
here.
By the way, I'm not trying to pick on a Republican. It's just that the Democratic sponsor of the bill, Fritz Hollings,
does not run for reelection until
2004.
Please note that I really have not done much research on this. Other more researched specific suggestions are welcome.
Until there is reliable encryption that takes prohibitively long periods to break
(remember, WEP is broken, and the break is a relatively quick one), this technology is simply unsecure, particularly
for corporate use.
You can two parties can use
Diffie-Hellman key exchange to
agree on a key even when all traffic is being watched.
Also, there is plenty of "reliable encryption that takes prohbitibitively long periods to break", such as triple DES (Data Encryption Standard), and any of
the
the Advanced Encryption Standard finalists,
at least in the sense that a lot of very qualified people have tried hard to break them for a long time in a very open process and so far failed.
(Rijndael won the AES endorsement, but, not to my knowledge, because of a vulnerability discovered
in any of the other finalists.) Granted, these algorithms are
not mathematically proven to require a substantial number of cycles to break or even to be as difficult as some other famous problem (like Michael Rabin's public key algorithm), but, if that is your standard of security, then you also should not be sending even your encrypted traffic over any internet backbone links that are not known to you to be physically secure.
I understand that digital flat panel interface
cables can be extended over great lengths with
fancy hardware. However, that probably isn't
worthwhile if you're trying to find the lowest
possible cost per display.
I agree that cabling
and reliability issues would probably make it
unoptimal to run all of the displays in a class
room from one computer. However, you could still
get considerable cost savings by, say, running
four or five displays per computer, where those
displays were in adjacent carels or clustered
together on a table, or even in four cubicles in a commercial environment in a "+" shaped arrangement.
Nobody is stopping anyone from writing a kernel
patch management system. Perhaps, after a dozen
or attempts, somebody will write a system that
is simple enough, fast enough and good enough
by various other criteria so that enough
people will gradually start using it in large
numbers, just like sourceforge, slashcode, cvs or diff and patch.
The likelihood that anyone will get it right on the first attempt is low. So, I think it would
be courting disaster for the Linux kernel development
community to commit to such a system in advance.
I just switched my DSL service from SBC to Covad
today. Although it was not my reason for leaving
SBC (they don't seem to prohibit NATing either),
one reason for my choice of Covad was that the salesperson I spoke to was actively pushing all of their DSL services as NAT-friendly. I told him I was going to put up a wireless NAT box for my neighbors to use and my salesperson told me that they think that's perfectly OK.
Going off topic, I feel compelled to warn anyone
who follows my recommendation
that if you use Covad's "TeleSurfer" DSL service, you'll need to use PPP-over-ethernet (requiring a patched version of PPP which I believe is already in some distributions), and your PPP login will be username@covad.net, as opposed to just
username.
No space-based expatration system is going to ship enough people off this
planet to make the slightest bit of difference. There are 250,000 new people on this planet, ever single
day. That is net of deaths, by the way.
250,000 people per day is 91.25 million people per year.
According to
this slide, European air travel was 541 million passengers in 1998, almost six times your figure for world population growth.
Wrong. Although copyrights on books and software both found in title 17, there are lots of restrictions of copyright law that specifically state that they only apply to software. Just look at the federal supremacy provision for just one example.
Not everyone agrees that the restrictions on books
are acceptable terms. There is lots of controversy about the ever-shrinking definition of fair use in books. For a good history of these restrictions from the origin of copyright as a way of controlling publication of the Christian bible in England, you might want to read
The Nature of Copyright: A Law of User's Rights, by L. Ray Patterson and Stanley W. Lindberg, with a forward by Robert W. Kastenmeier, who chaired the House subcomittee that created the 1976 Copyright Act.
Roxio, Inc. (Nasdaq: ROXI - news), the Digital Media
Company, and Gracenote today announced the signing of a multi-year license that provides Gracenote CDDB as the
exclusive CD recognition service to current and future Roxio customers
In other words, Roxio agreed to cease offering freedb service, even as an option.
Slashdot Mirror
[...] OpenH323 has [...] a VoIP -> PSTN gateway [...]
Perhaps a bunch of volunteers or entrepreneurs could set up home servers to allow incoming H323 connections to make local phone calls.
There was an effort to allow for free fax transmissions this way a few years ago. It used DNS as the mechanism for keeping track of which servers could make local calls to which phone number prefixes. However, I haven't heard about that project in a long time.
There must be a million little utilities for generating touch tones, but you probably don't even have to bother with that. If your computer is connected to the phone line by a modem that understands the AT command set, ending the phone number with a ";" will tell your modem to just dial a phone number and leave the line open without trying to negotiate a data connection. For example, "ATDT16505551212;" will dial 1-650-555-1212 and hold the line open. "ATH" will hang up the line.
Perhaps Innosphere would actually be supportive of open source and other third party software to connect to their network. They are, after all, just charging for use. The more usage the network gets, the more money they should make. At only $0.15/minute to call China, I'd sure like to use their service.
I assume the VOIP blaster comes with a license to run G.723.1 inside the device, but these patents would impede the development of free software that could use the data stream from VOIP blasters for purposes other than talking to other G.723.1 hardware devices.
It sure would be nice if the VOIP blaster had a mode where it could just transmit and receive raw audio samples (preferably by the standard USB audio class interface).
I did not say that authors would make more money with a shorter copyright. I said the public would be better off.
The part of your response that addresses my thesis that the public would be better off with a five year copyright is the scenario of a work being "discovered" much later. You need to argue that this scenario occurs with statistically significant frequency, not just an anecedotal case that apparently came about by a legal error anyhow, and that the incentive of a longer monopoly actually effects authors' production. The second part is particularly important. For example, if most of the works that are "discovered" later are the early works of authors that are now popular, then you are only talking about rewarding authors if and when they are already making a lot of money from their success.
By the way, a five year copyright expiration might even make an author more likely to be "discovered" in time to hit the big time with his or her next book while he or she is still has enough years to enjoy the money. This would be against the publisher's interest, since there is no guarantee that the author will use the same publisher. So you won't see a lot of release from five year copyright releases without a change in copyright law, since scales of efficiency will result in a small number of high volume publishers and a large number of authors, giving publishers greater negotiating power with unknown authors.
Computers have greatly reduced the time involved in writing, editing, typesetting and printing books since the days of writing a book with a typewriter. Distribution, sales and shipping of books have also been accelerated by technology (printing in more than one location, nearly realtime sales information across entire store chains, etc.).
Technology also means that the opportunities that copyright impedes have greatly increased. Being able to freely copy material online means that many people do not have to chop down trees to store information. Physical storage of books in digital forms is much more compact. Searching and sharing of free online information is orders of magnitudes easier.
There is even a secondary opportunity cost to authors in long copyrights: the development of derivative works is greatly limited by copyright when they are outside of "fair use." For example, I think that, given how much time has elapsed, Richard Hatch should be allowed to make his Battlestar Galactica sequel, and the rewrite of Gone with the Wind from a black perspective (The Wind Done Gone) should be allowed whether or not the book qualifies as a parody. The opportunities lost by impeding this sharing are increased when the efficiency with which these derivative works can be made is increased (i.e., more potential derivative works that otherwise would be produced are lost during each year of the copyright).
The costs of creating a book have dropped. The rate at which that investment can be recovered has accelerated, and opportunities that we lose during each year of copyright have increased. In my view, the balance point at which the public benefit of copyright is maximized has been greatly reduced. I believe that it would maximize public benefit to accelerate copyright expiration to about five years, maybe even less.
The web pages for djvu talk about the GPL'ed version including a decoder and a "simple bitonal (black-and-white) encoder" and a "very simple color encoder." This makes me wonder how the free encoders are and how they compare in compression and speed to other free encoders, such as those for jpeg and png.
A detailed positive answer to this question might make it worthwhile to more people to risk a couple hundred dollars of their time to try to install DjVu.
Oops. I meant to links to existing products based on IBM's 20.8" 2048x1536 ITQX20 module. Here they are.
Raintree Systems IN 2080-50 and National Display's Nova have the standard IBM dual LVDS digital inputs. I think there was an outfit in the UK called "Gemini Electonics" that was going to produce a similar device, but I could not find a link to them right now. RealVision has a dual LVDS video card for driving these monitors, although they promote it more for the electrically identical 6144x1536 grey scale version of the ITQX20. Finally, IBM's T-210 apparently uses the ITQX20 module, but only allows analog input, and only at something like 30Hz if you want to full resolution.
I have no financial relationship to any of these vendors.
What I wish someone would make is a kit consisting of everything except display module, which one could buy directly from IBM. That is: a plastic housing for the display, a power supply and a video card. The video card needs to be driven by 4 24-bit LVDS transmitter chips, which were $8 each when I looked in it--I think the chip may have beeen the National Semiconductor DS90CF581. 4 x $8 = $32 chip cost x 5X markup for retail = $160 additional cost of video card plus cost of two LVDS connectors (using the hokey rule of thumb that I've heard that total electronic component costs are typically 20% of the suggested retail prices of the resulting product).
Without the need to stock the display component, there would not have to have such a big mark up to cover storage, damage, etc. I think the kit without the display could easily cost under $1k.
The video card would be the significant engineering task. The two LVDS streams have to be kept in sync, so you cannot just use two of the LVDS cards that can drive an SGI-1600SW display. It looks like you really do have a make a new video card. The big question that I had not gotten around to researching was whether were was a VGA chip or chip set that could deliver the digital output for two screens without convering the signal to analog,that is deliver two 24-bit parallel streams (the display interface is basically that of two 1024x1536 LVDS flat panels, side by side). There are a number of dual head VGA cards. I just never got around to looking into whether it was possible to get digital output from their chips.
I think that if a kit like this were available, some computer retailers would assemble it and the panels to offer the finished product, and that would reach untapped section of the flat panel market that I think there should be significant deamnd for (a $3k 20.8" TFT for CAD, engineeringing, graphic art, etc.).
One minor drawback that might slightly impede the popularity of this display as it gets closer to the consumer range is that the housing for it is currently not very thin. The housing is about four inches thick, making it a bit less sexy looking from the side than most other flat panels.
I am not a lawyer, so don't rely on this as legal advice.
Generally, my experience with DSL providers has been that they do not prohibit this kind of sharing. Indeed Covad's sales staff touts this kind of unrestrictedness as selling point. (I called Covad's sales line to confirm this a while ago.)
This is not so surprising when you consider that the places where wireless service is expected to be most useful are "hot spots" where people gather and sit down, such as resturaunts, coffee shops, retailers, and various waiting areas, typically places that are likely not to order broadband otherwise.
It is difficult to prove a negative. So, maybe those who believe their acceptable use policy prohibits this should provide relevant excerpts. It would be interesting to see if those AUP's also, with the same stretch of interpretation, prohibit attaching a wireless access point, using Network Address Translation to connect more than one computer, or using the connection for work.
By the way, my impression is that DSL people seem to be more positively disposed toward this sort of thing than the cable modem providers.
Disclaimer and plug: I am involved in LANRoamer, a GPL'ed wireless roaming network that allows people to get paid for providing wireless service. The back end is also GPL-compatible open source.
Dr. Beltrami and his colleagues from the University of Michigan found that more than half of the land's heat gain over the past 500 years came during the 20th century, and 30% since 1950.
Sorry, I blew it, so I'll publicly take responsibility for my error. As the responses by "dhogaza" and "PhuCknuT" correctly point out, I misread that sentence. Beltrami et al claim that 30% of the temperature increase over that past 500 years happened since 1950, not just 30% of the temperature increase since 1900. Although I obviously puzzled over that sentence for before posting about it, I will endeavor to be more diligent against making this kind of mistake in the future. Sorry for wasting everyone's time with a simple misreading.
Dr. Beltrami and his colleagues from the University of Michigan found that more than half of the land's heat gain over the past 500 years came during the 20th century, and 30% since 1950.
So, they believe the rate of warming for 1951-2000 was less than half what is was for 1901-1950. I don't have much basis for an opinion on the meaningfulness of these researchers' results, but I would sure like to know how they explain this apparently levelling off.
A more cost effective strategy may be to promote the political careers of those who agree with us. (Granted, this is not a mutually exclusive with basically buying votes. It's just the other end of a continuum.)
We should actively help the careers of any politician who agrees with us, even if that allegiance would only be useful if they achieve some higher office later or can trade a favor.
Although money helps a campaign (I have contributed to EFF), money isn't necessarily the only currency. Campaign money eventually buys media and staffing. These are things which can be contributed in kind. To those sympathetic slashdotters who have time but not money and want to help: maybe you could find some college students who want to volunteer to help Rick Boucher's reelection campaign, finagle some ad banners, answer email, etc. On a larger scale, maybe you could make a web site to collect information on who is our friend, and to connect large numbers of volunteers with campaign organizations. That might be helpful enough to an election to make our support worth something in terms of election results.
Here is some additional VNC-related software that you may find useful. It comes from a project at Wyse a couple of years ago.
vncd spawns VNC X-windows sessions for incoming connections. It can be run from inetd or by itself. This is handy if you to deploy a bunch of thin clients around a Linux system. It also has a little protocol for negotiating things like display geometry and color depth. If you just sent "DONE" following a line feed to it, it will then start the VNC protocol. See the source code for commands for setting display geometry, color scheme and environment variables in the session.
remote-audio is a client and server for having audio programs send their output to a remote client. On the Linux machine that you are logged into, it intercepts the open() symbol in the C library to catch attempts to open thevnc-3.3.3.patch is source code for some VNC optimizations, such as local cursor, gzip compression of the link, special encoding for 1x1 rectangles (which, if I recall correctly, were nearly 50% of rectangles in some tests).
Wyse also shipped a "regular" X server that ran on an X terminal that could also accept incoming VNC connections to allow remote operation of that terminal. I believe the product was called WT5000. I started working on putting this into XFree86-4.2 and I've put a source snapshot here, but I have no idea if it works (it adds a "-vnc" argument to the X command line to allow incoming VNC connections).
LANRoamer and NoCatAuth appear to have started around the same time.
The two people who started NoCat gave a talk at the Bay Area Wireless User Group about a week after they started development, and I talked to them there. They (or at least one of them) said that they knew about LANRoamer when they started but thought that the LANRoamer back end was proprietary (we had publicly released it as free software by that time, but there was a period of about two weeks from when we announced LANRoamer to when we decided to free the back end, so I understand how they got that impression).
LANRoamer is a GPL'ed system that has been doing this for a while. We gave presentations on it at Bay Area Wireless User Group and Sbay.org back in June, I believe, before even the NoCat project started.
If you're into "bazaar" style software development, one thing you should note is that LANRoamer does network booting and upgrade reboots. So, if you contribute a useful feature to LANRoamer, it can be widely deployed quickly (based on our stability labels and the stability level each gateway owner has selected). Also, in addition to free accounts and revenue sharing to our access point providers, we also offer free courtesy accounts for people who run open access points (not just during a free beta), partly in an effort to thank the developers and "evangelists", but also to get them involved.
Anyhow, here is the software, including the latest LANRoamer network boot floppy or CD-ROM.
The network boot floppy currently requires that the first ethernet card be compatible with 3COM 3c59x, 8139too, Ether Express Pro 100, NE2000 PCI cards, Via Rhine, Tulip cards and PC-Net PCMCIA ethernet (the 802.11 card or the ethernet connection to your access point can be just about any card that Linux supports). Unlike NoKat (the last time I checked), LANRoamer can work behind firewalls, including NAT routers, even ones that distribute IP addresses that LANRoamer would otherwise use. Once your gateway is up, client machines can obtain addresses from your wireless gateway by DHCP and are taken to an SSL-based login page when they try to go anywhere on the web until they log in.
I'm not a lawyer, so don't use this as legal advice. Instead, you (the author of this slashdot article) may want to show it to your company's lawyer and suggest that he track this down.
According to this link, there is a case called "Whelan" that established that duplicating the detailed structure of a program was copying of expression rather than ideas, and therefore copyright infringement.
Also, I remember reading a very good article about ten years ago by law professor Pamela Samuelson, I think in Communications of the ACM or some other ACM publication, that talked about this decision and mentioned "detailed structure and flow", which would make the case for infringement even stronger.
Finally, I recall reading somewhere, perhaps in that same article, that there is some common law rule that the standard of similarity by which copyright infringement should be determined is supposed to correspond to how much access the alleged infringer had to the original work. In other words, if the alleged infringer had easy access to the original work (e.g., had carefully read the original GPL'ed source), then the standard for proving infringement is supposed to be easier.
Again, I'm just a layman. Don't use this as real legal advice.
If anyone is thinking of making a campaign contribution, you might want to look at whoever runs against Ted Stevens in his November 2002 senate reelection in Alaska, for the following reasons.
While most Replublicans seem skeptical of SSSCA, Ted Stevens cosponsored it. Without Stevens' support, this kind of bill will seem much more partisan, reducing its chances unless the democrats manage to get control of both houses of congress and the white house.
Alaska has a small population, so a little money should go a long way in an election. Stevens has $1.5 million, including PAC contributions, an amount within the fund raising capabilities of a bunch of slashdot's wealthier nerds, less than half of the 1996 national average of $3.6 million. Also, Stevens might not get critical support from his fellow Republicans if he finds himself in need, as he has embarassed his party with his pork barrel politics, netting Alaskans $766 per capita of federal money (making the state #1 in this category), in comparison to the national average of $25.52, as calculated here. By the way, I'm not trying to pick on a Republican. It's just that the Democratic sponsor of the bill, Fritz Hollings, does not run for reelection until 2004.
Please note that I really have not done much research on this. Other more researched specific suggestions are welcome.
Until there is reliable encryption that takes prohibitively long periods to break (remember, WEP is broken, and the break is a relatively quick one), this technology is simply unsecure, particularly for corporate use.
You can two parties can use Diffie-Hellman key exchange to agree on a key even when all traffic is being watched.
Also, there is plenty of "reliable encryption that takes prohbitibitively long periods to break", such as triple DES (Data Encryption Standard), and any of the the Advanced Encryption Standard finalists, at least in the sense that a lot of very qualified people have tried hard to break them for a long time in a very open process and so far failed. (Rijndael won the AES endorsement, but, not to my knowledge, because of a vulnerability discovered in any of the other finalists.) Granted, these algorithms are not mathematically proven to require a substantial number of cycles to break or even to be as difficult as some other famous problem (like Michael Rabin's public key algorithm), but, if that is your standard of security, then you also should not be sending even your encrypted traffic over any internet backbone links that are not known to you to be physically secure.
I agree that cabling and reliability issues would probably make it unoptimal to run all of the displays in a class room from one computer. However, you could still get considerable cost savings by, say, running four or five displays per computer, where those displays were in adjacent carels or clustered together on a table, or even in four cubicles in a commercial environment in a "+" shaped arrangement.
Nobody is stopping anyone from writing a kernel patch management system. Perhaps, after a dozen or attempts, somebody will write a system that is simple enough, fast enough and good enough by various other criteria so that enough people will gradually start using it in large numbers, just like sourceforge, slashcode, cvs or diff and patch.
The likelihood that anyone will get it right on the first attempt is low. So, I think it would be courting disaster for the Linux kernel development community to commit to such a system in advance.
I just switched my DSL service from SBC to Covad today. Although it was not my reason for leaving SBC (they don't seem to prohibit NATing either), one reason for my choice of Covad was that the salesperson I spoke to was actively pushing all of their DSL services as NAT-friendly. I told him I was going to put up a wireless NAT box for my neighbors to use and my salesperson told me that they think that's perfectly OK.
Going off topic, I feel compelled to warn anyone who follows my recommendation that if you use Covad's "TeleSurfer" DSL service, you'll need to use PPP-over-ethernet (requiring a patched version of PPP which I believe is already in some distributions), and your PPP login will be username @covad.net, as opposed to just username.
No space-based expatration system is going to ship enough people off this planet to make the slightest bit of difference. There are 250,000 new people on this planet, ever single day. That is net of deaths, by the way.
250,000 people per day is 91.25 million people per year. According to this slide, European air travel was 541 million passengers in 1998, almost six times your figure for world population growth.
Wrong. Although copyrights on books and software both found in title 17, there are lots of restrictions of copyright law that specifically state that they only apply to software. Just look at the federal supremacy provision for just one example.
Not everyone agrees that the restrictions on books are acceptable terms. There is lots of controversy about the ever-shrinking definition of fair use in books. For a good history of these restrictions from the origin of copyright as a way of controlling publication of the Christian bible in England, you might want to read The Nature of Copyright: A Law of User's Rights, by L. Ray Patterson and Stanley W. Lindberg, with a forward by Robert W. Kastenmeier, who chaired the House subcomittee that created the 1976 Copyright Act.
From the Roxio press release at Yahoo:
Roxio, Inc. (Nasdaq: ROXI - news), the Digital Media Company, and Gracenote today announced the signing of a multi-year license that provides Gracenote CDDB as the exclusive CD recognition service to current and future Roxio customers
In other words, Roxio agreed to cease offering freedb service, even as an option.