If you really think that you can prove to the
satisfaction of a court that the net impact of software patents is negative, and that they can and
should be distinguished from other sorts of patents (which flies in the face of established
US legal history where software patents are
permitted primarily because they are considered
technically indistiguishable from a black box that
performs the same function as the software program), then why not go ahead and appropriate a few US patents.
You should be able to use any software patents with impugnity. Apple in particular seems litigous toward individuals who borrow their IP.
Personally I think it would be both difficult and problematic to prove. The best chance for repealing software patents IMO is to lobby congress to specifically change the law to make software patents illegal. If on the whole they cause more harm then good, they should be willing to listen... of course major corporations might disagree with you on which way the balance should go. And in the US both the courts and congress tend to believe that business knows what it is talking about when determining what has a net positive impact on the economy.
You don't think people have already thought of using that approach to bugger the system? It is illegal for two parties not genuinely at odds to use the courts in that way, and if they figure out what is going on you'll soon be in the middle of a criminal rather than civil proceeding.
"Microsoft said it had not been served with the lawsuit and could not comment."
Those are of course two disconnected statements. After they've been served they will
say:
"Microsoft representatives were unable to comment on the active lawsuit."
And after settlement they will write:
"The Microsoft.v AT&T patent infringement lawsuit has been resolved in a private settlement. Microsoft couldn't comment on the details of the
settlement."
Part of the reason it has been difficult to challenge new types of patents is because of a ruling by the Supreme Court that specific patents are not subject to this constitutional restriction, only that the patent office on the whole must do more to promote business than to harm it.
I'm not saying there isn't a case to be made, but you have to balance both sides. The patent system is
there solely 'to promote progress in science and the useful arts' (as the US Constitution puts it), so any
granting of patents on algorithms must pass this test.
So this is unfortunately only true if you can argue that the whole system of patents fails the test, if I understand correctly.
Special letter forms don't need to be coded into unicode to be viewable. SVG, Postscript
and other languages do a perfectly good level
of presentation. So unless you can convince me
that a Korean/Chinese person will be trying to
do a word search through an historical Japanese/Taiwanese/Vietnamese document and will
always inadvertently find the Korean ACK/Chinese
SPOO when what he was really looking for was the Japanese FOOFLE/Taiwanese FLUM.
Personally I can't understand why anyone in the world would want to search in a character set of more than 60,000 characters. I'd personally be pissed off if the UNICODE committee started adding special letter forms for US product trademarks (so they would render correctly) when as a user I'd rather just have them be findable.
Really, the author needs to understand the use of the ALT tag.
Noise cancellation depends on being able to generate a waveform with the opposite amplitude of the noise around you that is in perfect synch with that noise. Putting your noise canceller through an amplifier and speakers would just give you a nice reverb loop, and would be quite difficult to synchronize even if you somehow were able to solve the crossed i/o.
Yeah, I used to have an Atari Lynx. Great in a dark movie theater before the movie starts. Useless in strong sunlight.
From what I've seen though (from people carrying around the new Motorola cell phones) these things are pretty usable even in direct sunlight, partially because when they are off they really absorb light quite well so contrast is very good. I doubt that the OLED's would become
reflective when 'lit' so a sunny beach might still be out of the question, but at least they aren't as hard to read as the first LED digital watches.
Although it is true that the Linux TCP stack will allow you to construct raw TCP packets, you can't do that without root privs. Windows basically has the same TCP implementation as an
unprivileged UNIX login.
That said, I do think it would be helpful if routers checked the return IP address to see if they have arrived over a valid channel as that would essentially eliminate spoofing.
Remember when much of the internet backbone used to run routed?
And what do you have against hypocrisy, pray tell?:)
Seriously though, just because you see them as
incompatible doesn't mean that they are. Of course to be compatible you have to be able to
make a value judgement, which I know is terribly out of
vogue these days. These days we send out doctors
to heal the wounded in war torn regions and
end up giving aid and comfort to the militia,
thereby allowing them to prolong their reign of
terror... because we don't want to seem
hypocritical.
I think that this will be the predominant political issue in the coming decade or two (and
I think John McCain's showing in the republican presidential primary was in large part an effect of his stance on campaign finance reform, which is closely tied to all of these issues.)
Ah, good point. But is Microsoft actually going to win the console wars? True they have
billions of dollars...
But against them, they don't understand the
market, and they aren't even using the tried and true
methods of succeeding in that market, but have instead decided to forge their own new path,
and ultimately their hardware is more expensive
than their competition.
Microsoft hasn't exactly made good on WebTV if you'll recall, and that was their best bet at a
grandmother can use it device that I've seen
in recent years. It's too sad.
I do think that the console market is a rich oligopoly ripe for upsetting, I just don't think that Microsoft has a guarantee of winning.
Well... I don't claim to have a very good sense of time, so you may be right.
I started out on Caldera 1.0, then jumped to
Redhat 4.2 when Caldera dropped support for their 1.0 series and moved to an incompatible distribution. XFree86 was usable all along, but had drivers that were far behind Xinside. I had a Matrox Millenium II back then and XFree only supported my old ATI card. As I recall it wasn't until Redhat 5.0 that Matrox support was included at all, and at that point I was getting about 14 Xmarks from AccelleratedX but only about 5 Xmarks from XFree86. I didn't actually end up getting rid of AccelleratedX until Redhat released 6.1 (I skipped 6.0).
It also might depend on how much free time you had (versus how much free money.) In the SF Bay Area most of the Linux users I knew had more of the latter, but I speak only from personal memory. We all started using Linux well after college, so money wasn't our constraining factor. (I still have a thousand dollar copy of Solaris x86 that I bought on special (for $500) and never installed
since I found Linux the same month.)
This presumes that you know what the next desktop OS will look like. If instead the industry completely misses the boat, heads off in one direction and interest of the
public runs off in another then Linux could
very easily cross the finish line first.
Consider TIVO for example. I'm not a big
fan of the device, but suppose it caught the
imagination of the american public in a major
way; it could easily result in a complete
displacement of Windows on the consumer
desktop.
There are other examples, but that isn't
the point. Open source also means that a
thousand companies, without any significant
value or staff size, are able to innovate
however the mood strikes them. And that, I
claim, is the very core of invention.
Microsoft's can be displaced just as easily
as Nokia surprised and overwhelmed the
cell-phone market; all it takes is a blind
spot. And right now that blind spot (for
Microsoft is in recognizing the
value of open source. (Who knows what it will
be next year.)
Isn't it amazing how out of touch the author was. Even a year ago when the story was first posted you could tell that Motif was dying, and the past 12 months haven't made his assertions seem exactly full of insight.
But I suppose it is to be expected, considering the author's investment in Motif.
In addition to making a good developers desktop, Linux is pretty good when supporting users who have absolutely no intention of doing their own system administration.
My Mom calls me all the time asking how to fix something or another on her Windows desktop that I could fix much more easily if she were running
Linux and I could simply ssh to her
machine.
And for the software that she likes to run
Linux is really almost there. She just needs a
good mail client, a web browser, and a word
processor. I need her to use tools that don't
do unexpected things like open unclosable windows
(like Netscape & IE), propagate mail viruses (like Outlook & Word), or mess up her registry (like Windows 98).
Personally I think that Linux is getting close to satisfying all of those. Windows doesn't even seem to be trying.
Especially true when you compare where Linux was two years ago and where it is today.
By my recollection about two years ago XFree86 was just starting on 4.0, performance and video card support were so slow to arrive that almost everyone running X had to buy either Metro-X or Accellerated-X to get good performance and
reasonable hardware support.
Two years ago Motif was still the only complete and standard toolkit for X11 based software development, it looked like crap, had infinitely many bugs, and added another $100 to the cost of a system.
Two years ago was when I really despaired of Linux ever finding success on the Desktop, but not today. I finally gave in a month ago and installed GNOME (replacing fvwm as my favorite
desktop), and while the keyboard accellerators
aren't all quite the same it is a much better
environment overall.
The architecture of XFree86 now allows new video hardware to be added much more quickly. Motif is dead and unmourned; even Mozilla has
moved on. Speaking of which Mozilla has finally
become usable to my level of tolerance, meaning
it has fewer annoyances than Netscape does now.
Linux on the desktop has never looked as good
before as it does today.
Maybe it is dead according to someone's expectations, but it looks like it is
on an exponential growth curve to me.
Yeah, I remember reading about IP/SMTP back when it came out. The ping times are supposed to be 30s or so, but with the IT infrastructure at
my company I think they would end up being closer
to five minutes or more.
In any case TCP isn't going to run well over that kind of latency, you'll just fill up the pipe with retransmit requests until the TCP connection times out.
To really make use of this kind of application you need some protocols that aren't oriented toward interactive latency. IP datagrams over avian carrier are all well and good, but at an
hour per packet they had damned well better be
carrying complete messages.
I could see the Palm Pilot application synchronization being done over non-interactive protocols. Store and forward messaging would
work fine, and of course the web would be okay
if a bit slow (over SMTP.)
So there are plenty of potential uses, the question is really whether those uses will become prevalent, or the latencies of handheld wireless devices will drop to where they are no longer needed.
I don't even think that it is that far out. Wireless, disconnected devices, and packet radio connected subnets are all examples of systems that need a similar solution. I for one would quite like to see an HTTP implementation that is hosted on a message oriented transport protocol instead
of TCP. E-mail would also translate well of course, as would usenet.
HTTP would have to be modified slightly to bundle subrequests automatically (a wget protocol?) and I haven't yet read far enough to see how the draft RFC proposes establishing a
route between the source and destination. If
those routes change very rapidly (as for example
might be true over packet radio) I can imagine
some difficulty in finding a path to the destination.
I just picked up working with RPC in C after several years working with CORBA, RMI, and other more recent protocols, and I have to agree completely. RPC really is a very nice and simple distributed invocation protocol. See my sig to check out how I am using it.
In case you are interested, I've also written an XML stream for XDR (the data translation layer of RPC) so that structures can
be deserialized and reserialized easily to XML, but without giving up XML's loose coupling. I wrote up an article on the idea here.
Well... in this case Worlds has been around for quite a while, and no one else that I'm aware of had implemented anything like shared 3d VR over the internet back when they started it.
Worlds chat was released in 1994, so it does follow Snowcrash by about two years. Can you claim fiction as prior art?
Getting code to compile is one thing. Getting your linker not to spew out 150 copies of the generated template class is quite another, and the pragmas required for doing so without hiding the implementation class vary from compiler to compiler (except MSVC which handles the problem in its linker.)
Moreover, templated code is extremely annoying to debug.
Despite its lack of type safety I prefer abstract base types (like void*) for container implementations rather than STL like type safe
containers. The overhead of STL templates is
getting a new copy of their whole btree algorithm for each new type you'd like to build a map from.
I think in his post to comp.compression Mike
wrote that his random data had come from random.org. Random doesn't use a PRN generation algorithm; they get random data from atmospheric radio noise (tune your AM radio to a non-station and you'll hear what that means.) To remove any bias in the data they postprocess the noise by removing high order bits, and the low order bits are shifted to normal using a bit discarding algorithm.
Slashdot Mirror
If you really think that you can prove to the satisfaction of a court that the net impact of software patents is negative, and that they can and should be distinguished from other sorts of patents (which flies in the face of established US legal history where software patents are permitted primarily because they are considered technically indistiguishable from a black box that performs the same function as the software program), then why not go ahead and appropriate a few US patents.
You should be able to use any software patents with impugnity. Apple in particular seems litigous toward individuals who borrow their IP.
Personally I think it would be both difficult and problematic to prove. The best chance for repealing software patents IMO is to lobby congress to specifically change the law to make software patents illegal. If on the whole they cause more harm then good, they should be willing to listen... of course major corporations might disagree with you on which way the balance should go. And in the US both the courts and congress tend to believe that business knows what it is talking about when determining what has a net positive impact on the economy.
You don't think people have already thought of using that approach to bugger the system? It is illegal for two parties not genuinely at odds to use the courts in that way, and if they figure out what is going on you'll soon be in the middle of a criminal rather than civil proceeding.
Those are of course two disconnected statements. After they've been served they will say:
And after settlement they will write:
Just a guess ;)
Part of the reason it has been difficult to challenge new types of patents is because of a ruling by the Supreme Court that specific patents are not subject to this constitutional restriction, only that the patent office on the whole must do more to promote business than to harm it.
I'm not saying there isn't a case to be made, but you have to balance both sides. The patent system is there solely 'to promote progress in science and the useful arts' (as the US Constitution puts it), so any granting of patents on algorithms must pass this test.
So this is unfortunately only true if you can argue that the whole system of patents fails the test, if I understand correctly.
IANAL by the way.
Prince is again Prince. He got his name back when the music industry contract that prohibited him from using his own name expired.
Special letter forms don't need to be coded into unicode to be viewable. SVG, Postscript and other languages do a perfectly good level of presentation. So unless you can convince me that a Korean/Chinese person will be trying to do a word search through an historical Japanese/Taiwanese/Vietnamese document and will always inadvertently find the Korean ACK/Chinese SPOO when what he was really looking for was the Japanese FOOFLE/Taiwanese FLUM.
Personally I can't understand why anyone in the world would want to search in a character set of more than 60,000 characters. I'd personally be pissed off if the UNICODE committee started adding special letter forms for US product trademarks (so they would render correctly) when as a user I'd rather just have them be findable.
Really, the author needs to understand the use of the ALT tag.
Noise cancellation depends on being able to generate a waveform with the opposite amplitude of the noise around you that is in perfect synch with that noise. Putting your noise canceller through an amplifier and speakers would just give you a nice reverb loop, and would be quite difficult to synchronize even if you somehow were able to solve the crossed i/o.
Yeah, I used to have an Atari Lynx. Great in a dark movie theater before the movie starts. Useless in strong sunlight.
From what I've seen though (from people carrying around the new Motorola cell phones) these things are pretty usable even in direct sunlight, partially because when they are off they really absorb light quite well so contrast is very good. I doubt that the OLED's would become reflective when 'lit' so a sunny beach might still be out of the question, but at least they aren't as hard to read as the first LED digital watches.
Although it is true that the Linux TCP stack will allow you to construct raw TCP packets, you can't do that without root privs. Windows basically has the same TCP implementation as an unprivileged UNIX login.
That said, I do think it would be helpful if routers checked the return IP address to see if they have arrived over a valid channel as that would essentially eliminate spoofing.
Remember when much of the internet backbone used to run routed?
And what do you have against hypocrisy, pray tell? :)
Seriously though, just because you see them as incompatible doesn't mean that they are. Of course to be compatible you have to be able to make a value judgement, which I know is terribly out of vogue these days. These days we send out doctors to heal the wounded in war torn regions and end up giving aid and comfort to the militia, thereby allowing them to prolong their reign of terror... because we don't want to seem hypocritical.
Bullshit.
There is a huge amount of information on the net about the increasing corporatization of America. From the WTO protests (and presidential primary demonstrations), to activism around Biotechnology, copyright extensions, trademark and trade secrets litigation, patents on software, and on DNA.
I think that this will be the predominant political issue in the coming decade or two (and I think John McCain's showing in the republican presidential primary was in large part an effect of his stance on campaign finance reform, which is closely tied to all of these issues.)
Ah, good point. But is Microsoft actually going to win the console wars? True they have billions of dollars...
But against them, they don't understand the market, and they aren't even using the tried and true methods of succeeding in that market, but have instead decided to forge their own new path, and ultimately their hardware is more expensive than their competition.
Microsoft hasn't exactly made good on WebTV if you'll recall, and that was their best bet at a grandmother can use it device that I've seen in recent years. It's too sad.
I do think that the console market is a rich oligopoly ripe for upsetting, I just don't think that Microsoft has a guarantee of winning.
Well... I don't claim to have a very good sense of time, so you may be right.
I started out on Caldera 1.0, then jumped to Redhat 4.2 when Caldera dropped support for their 1.0 series and moved to an incompatible distribution. XFree86 was usable all along, but had drivers that were far behind Xinside. I had a Matrox Millenium II back then and XFree only supported my old ATI card. As I recall it wasn't until Redhat 5.0 that Matrox support was included at all, and at that point I was getting about 14 Xmarks from AccelleratedX but only about 5 Xmarks from XFree86. I didn't actually end up getting rid of AccelleratedX until Redhat released 6.1 (I skipped 6.0).
It also might depend on how much free time you had (versus how much free money.) In the SF Bay Area most of the Linux users I knew had more of the latter, but I speak only from personal memory. We all started using Linux well after college, so money wasn't our constraining factor. (I still have a thousand dollar copy of Solaris x86 that I bought on special (for $500) and never installed since I found Linux the same month.)
This presumes that you know what the next desktop OS will look like. If instead the industry completely misses the boat, heads off in one direction and interest of the public runs off in another then Linux could very easily cross the finish line first.
Consider TIVO for example. I'm not a big fan of the device, but suppose it caught the imagination of the american public in a major way; it could easily result in a complete displacement of Windows on the consumer desktop.
There are other examples, but that isn't the point. Open source also means that a thousand companies, without any significant value or staff size, are able to innovate however the mood strikes them. And that, I claim, is the very core of invention.
Microsoft's can be displaced just as easily as Nokia surprised and overwhelmed the cell-phone market; all it takes is a blind spot. And right now that blind spot (for Microsoft is in recognizing the value of open source. (Who knows what it will be next year.)
Isn't it amazing how out of touch the author was. Even a year ago when the story was first posted you could tell that Motif was dying, and the past 12 months haven't made his assertions seem exactly full of insight.
But I suppose it is to be expected, considering the author's investment in Motif.
In addition to making a good developers desktop, Linux is pretty good when supporting users who have absolutely no intention of doing their own system administration.
My Mom calls me all the time asking how to fix something or another on her Windows desktop that I could fix much more easily if she were running Linux and I could simply ssh to her machine.
And for the software that she likes to run Linux is really almost there. She just needs a good mail client, a web browser, and a word processor. I need her to use tools that don't do unexpected things like open unclosable windows (like Netscape & IE), propagate mail viruses (like Outlook & Word), or mess up her registry (like Windows 98).
Personally I think that Linux is getting close to satisfying all of those. Windows doesn't even seem to be trying.
Especially true when you compare where Linux was two years ago and where it is today.
By my recollection about two years ago XFree86 was just starting on 4.0, performance and video card support were so slow to arrive that almost everyone running X had to buy either Metro-X or Accellerated-X to get good performance and reasonable hardware support.
Two years ago Motif was still the only complete and standard toolkit for X11 based software development, it looked like crap, had infinitely many bugs, and added another $100 to the cost of a system.
Two years ago was when I really despaired of Linux ever finding success on the Desktop, but not today. I finally gave in a month ago and installed GNOME (replacing fvwm as my favorite desktop), and while the keyboard accellerators aren't all quite the same it is a much better environment overall.
The architecture of XFree86 now allows new video hardware to be added much more quickly. Motif is dead and unmourned; even Mozilla has moved on. Speaking of which Mozilla has finally become usable to my level of tolerance, meaning it has fewer annoyances than Netscape does now.
Linux on the desktop has never looked as good before as it does today.
Maybe it is dead according to someone's expectations, but it looks like it is on an exponential growth curve to me.
Yeah, I remember reading about IP/SMTP back when it came out. The ping times are supposed to be 30s or so, but with the IT infrastructure at my company I think they would end up being closer to five minutes or more.
In any case TCP isn't going to run well over that kind of latency, you'll just fill up the pipe with retransmit requests until the TCP connection times out.
To really make use of this kind of application you need some protocols that aren't oriented toward interactive latency. IP datagrams over avian carrier are all well and good, but at an hour per packet they had damned well better be carrying complete messages.
I could see the Palm Pilot application synchronization being done over non-interactive protocols. Store and forward messaging would work fine, and of course the web would be okay if a bit slow (over SMTP.)
So there are plenty of potential uses, the question is really whether those uses will become prevalent, or the latencies of handheld wireless devices will drop to where they are no longer needed.
I don't even think that it is that far out. Wireless, disconnected devices, and packet radio connected subnets are all examples of systems that need a similar solution. I for one would quite like to see an HTTP implementation that is hosted on a message oriented transport protocol instead of TCP. E-mail would also translate well of course, as would usenet.
HTTP would have to be modified slightly to bundle subrequests automatically (a wget protocol?) and I haven't yet read far enough to see how the draft RFC proposes establishing a route between the source and destination. If those routes change very rapidly (as for example might be true over packet radio) I can imagine some difficulty in finding a path to the destination.
And just when I'm out of moderation points too... Your prosecuted had me laughing out loud once I managed to read it correctly.
Gawd I hate the trend toward trying to XML everything. Why in the world would you want to use XML as a protocol? For message content, fine.
But for encapsulation? MIME is much better at framing a message.
For a protocol? SMTP and HTTP are fine protocols...
Is there some reason that this is more legibile than SMTP?
<xml>
<error>550</error>
<msg>We don't accept mail from spammers</msg>
</xml>
Yuck!
I just picked up working with RPC in C after several years working with CORBA, RMI, and other more recent protocols, and I have to agree completely. RPC really is a very nice and simple distributed invocation protocol. See my sig to check out how I am using it.
In case you are interested, I've also written an XML stream for XDR (the data translation layer of RPC) so that structures can be deserialized and reserialized easily to XML, but without giving up XML's loose coupling. I wrote up an article on the idea here.
Well... in this case Worlds has been around for quite a while, and no one else that I'm aware of had implemented anything like shared 3d VR over the internet back when they started it.
Worlds chat was released in 1994, so it does follow Snowcrash by about two years. Can you claim fiction as prior art?
Getting code to compile is one thing. Getting your linker not to spew out 150 copies of the generated template class is quite another, and the pragmas required for doing so without hiding the implementation class vary from compiler to compiler (except MSVC which handles the problem in its linker.)
Moreover, templated code is extremely annoying to debug.
Despite its lack of type safety I prefer abstract base types (like void*) for container implementations rather than STL like type safe containers. The overhead of STL templates is getting a new copy of their whole btree algorithm for each new type you'd like to build a map from.
I think in his post to comp.compression Mike wrote that his random data had come from random.org. Random doesn't use a PRN generation algorithm; they get random data from atmospheric radio noise (tune your AM radio to a non-station and you'll hear what that means.) To remove any bias in the data they postprocess the noise by removing high order bits, and the low order bits are shifted to normal using a bit discarding algorithm.