Re:Patents Are Not a Problem...
on
Perens on Patents
·
· Score: 2, Insightful
My point was that that patents are allowed in almost all technical areas so why not for software?
Because software is different. Software is just a bunch of logical/mathematical constructs in a language a computer can understand. It's maths. As an example of how this is different from traditional technological fields, let's take a very simple physical universe: that of Lego bricks.
If you place one Lego brick on the ground, it will just sit there. Put one or two on top, and you'll still have no problem. If you put 100 Lego bricks right on top of each other, you get a very unstable structure. So even in this very simple physical environment, things that behave one way in the small, may behave quite differently when used in a more complex whole.
With software, this is not the case. One small, sound logical reasoning remains just as valid in a larger whole. Software development by its very nature consists entirely of combining all kinds of small logical constructs into a, non-obvious or not, bigger whole. And software is being innovated every day, because if you don't innovate, your competition will and you will lose whatever edge you had.
Please read this study, carried out by the Max Planck Institute and the Fraunhofer institute (by no means anti-swpat establishments) for more on why software is different and how software development works differently. For a more
philosophical approach, in case you want to understand why those empirical results are what they are, see this page.
Broadness of initial patents is just a passing phase and open-source fanatics are damaging the overall purpose of software patents which is to develop a repository of knowledge
But software patents don't work that way, as shown by the recent FTC study (link is to a summary of all swpat related stuff, link to the full study is available at the top for you to verify, should you think that the person who created that page misrepresented the facts).
Some quotes:
One panelist stated that "the [patent] system discourages you from looking very hard [at patent disclosures] because... simply by virtue of poking around to find out what patents exist you expose yourself to willfulness claims which can triple the amount of damages and exposure to attorney's fees."
The panelist summed up the problem with the statement "there's too much information and it is no longer meaningful."
Computer scientists do not go to patent databases to look up information on how to do something.
which would otherwise be locked up in the vaults of giant corporations.
I guess that's why these same giant corporations are arguing that they need software patents, because it's supposedly so easy to reverse engineer and reimplement their precious techniques. They just want it all: very strong protection for abstract things which do not require such guarantees in order for them to be made public.
How come? Because you publish software, you don't manufacture it (just like you don't publish an industrial valve or a chemical reaction, you manufacture or conduct those). Publishing by definition is "making public". It's the same with business methods: either you use a business method and it automatically becomes public, or you don't use it.
Software patents are a giant corporation's wet dream, because they pretty much only benefit the patent holders.
Re:Loophole never tried?
on
Perens on Patents
·
· Score: 2, Insightful
In case of "methods of technology that use these algorithms", you do not get a patent on the algorithm that is implemented in the technical apparatus, but on the way you implemented it technically. In software, you get copyright protection for the way you implement an algorithm (because it's more like writing a story based on an idea, than on constructing a valve based on some physical properties of the materials it's made of).
Software patents however, give you patent protection for an algorithm, which is a mixture (from hell) of both principles. As such, they lead to enormously broad (you get exclusive usage of some principle) and conflicting (copyright grants me the right to publish my program, but a patent owner forbids it because it uses a patented technique) protections.
In the European pro-swpat camp, the reasoning is that a computer executing a computer program is no longer a computer program as such and thus patentable, at least if the computer program allows you to make more efficient use of the computer than without it (or with other known programs). It's really stupid of course... What else can you do with a computer program than run it on a computer? Use it as wall paper?
The problem here is that patents are increasingly and increasingly not about ways to solve a problem and about problems
This is indeed a very important point that's often overlooked. Take for example the claims of this European (granted) patent from Siemens. It's a patent on guessing the word you are trying to type on a mobile phone (e.g., if you type 843, it will show "the" instead of "tgd"). They developed one algorithm to do this, but most of the patent claims are about the problem, not the solution (even the more specific claims).
If you make a processing device coupled to a memory, input device and screen guess the word you want to type, you violate that patent. The more specific claims simply add different kinds of keyboards you can use, that you can also use the frequency of chosen words to guide suggestions, allowing people to turn on/off this guiding algorithm, showing a list of possible words if there's ambiguity and let the user choose from those etc... Not really things that narrow the scope of the claims very much.
They don't take the right in their own hands, i.e. they don't launch attacks on ISP's, don't hack spammer websites, don't try to crash open proxies,... They simply publish a list of IP-address of what they claim are spam-friendly ISP's.
(I'm not SPEWS and don't know anyone at SPEWS). That said:
dslreports.com has address 209.123.109.175. That address only appears in a level 2 listing. Very few people use level 2 listings, the "real" SPEWS are the level 1 addresses. What level 2 really means, is explained in their FAQ (Q22).
SPEWS did not add dslreports.com to their blacklist (search the linked page for dslreports, it's not mentioned). This does not make it less annoying for the owners of dslreports.com obviously, but there are differences. E.g., if a spammers moves, the blacklisting will be moved too, for dslreports.com it obviously wouldn't (no, that doesn't mean I think dslreports should simply move and shut up, I know things like that cost money).
The blacklist that SPEWS publishes is an *opinion*. Everyone is free to follow their opinion or not and use it to (over-)protect their property or not. If an ISP uses it (or any other blacklist) and doesn't clearly inform its customers about that fact, then this ISP is at fault.
Nevertheless, I completely agree it's sad that the spammer situation has gotten so much out of hand that people resort to this kind of carpet-blacklisting to try to force ISP's to stop their spam support (as larger ip-blocks are only added when an ISP refuses to remove its spammers, or starts moving them around to non-blacklisted IP-addresses).
It's however pretty much the last resort that other people have to do anything about it. If an ISP does not experience any significant harm from hosting spammers (and in facts profits largely from it) and does not want to remove them because it's the right thing to do, what else can you do to tell the ISP to FOAD if you don't want to become a vigilante?
I think it's fairly widespread, as Apple's CFO mentioned the unusually high number of warranty expenses for Powerbooks (due to initial white spots problem with the 15" AlBooks, now completely resolved according to them) and iBooks (not further specified) as the reason that their margins were lower than expected during the previous fiscal quarter.
I mentioned this above, but I'll repeat myself... the end user often has no clue what their ISP or company's anti-spam service is doing.
Then the ISP/company is at fault, not the blacklist.
Subscribing to a blacklist service is sort of like hiring a bouncer for your party. He probably does a good job of keeping out anyone uninvited. But what happens when you order a pizza and he punches the pizza guy in the face?
Blacklists are indeed by no means flawless, and I'll never claim that. It's just that spammers made the situation so bad, that a lot of people are willing to put up with the negative side effects of blacklists in order to keep their email (system) usable and/or to try to do something against pro-spam ISP's (like when they use lists like SPEWS).
Stop being stupid and blocking IPs. It's counterproductive. You're throwing napalm on an ant hill and most of the time, the ant you're going after has already moved somewhere else.
If that were the case, almost nobody would use blacklists.
It should be a punishable offense for such idiots who block legitimate IP from sending e-mails.
The recipient has full rights to determine for himself who is legitimised to send him mail.
If a company is using an ISP that happened to get a spammer onboard and the company ends up being blacklisted and losing business, the makers of the blacklist should rightfully be sued for lost business and denial of services.
Blacklists don't block anyone's mail. Its users do, and those users have the full rights to decide who gets to make use of their property or the services they pay for.
Imaine if a citizen set up a road block on a highway just because they heard criminals used it.
Blacklists merely advise you to avoid a highway. They can not and do not block anything.
I'd like to see lawsuits start being brought against blacklist runners and won. Their method is in many cases as reported on slashdot but not called as such, criminal.
Voicing an opinion is actually one of the rights protected in pretty much every constitution around the world.
Maybe it is indeed a manufacturing instead of a design defect. However, I mainly took issue with your assertion that "If people with failed logic boards did have AppleCare they'd be up and running again in under a week." People should not need to buy AppleCare to cover for design or more or less systematic manufacturing defects.
The standard 1 year warranty is meant to cover those and it's true that if you have bad luck, your machine may die 1 day after your warranty expires. However, if it turns out that an unusually large amount of machines show these failures after the standard warranty has expired, the manufacturer should take up responsibility for that. All imho, of course.
compare serial numbers, it would be possible to determine if the problem is, indeed, limited to sequential or near-sequential batches of serial numbers. That would be genuinely useful information, not only for the lawsuit, but also for Apple, who I think would succumb to a repair/replacement if they had that information.
Apple does have all serial numbers of machines that broke, at least of those that have been sent in for repair (which is the majority, I assume).
The situation is similar for the iBooks. Having a major hardware component die is absolutely no fun. However this sort of failure is something that would be covered and has been covered under the hardware warranty.
This is not about whether it's fun or not. There are systematic failures of the same components (motherboard, video cables, video chip) of different iBooks from different people all over the world, happening after roughly the same time (about one year). There is clearly a design error, not some damage that occurs from bad handling or reasonably expectable wear and tear (unless you think it's normal that a laptop should need a new logic board or two every year).
There's the commotion exactly because Apple users are not used to this. In our family there are a IIsi, beige G3, G4/400, PB G3/500, tibook/667, iMac FP 15"/700 and an iBook/800. The only machine that has ever needed repairs is the iBook (two new motherboards in one month). Since my 1-year warranty runs out in less than a month, I'm very inclined to buy AppleCare, which would be a first.
Well, the warranty runs out in 3 months actually, because the last repair is only a week ago and you get 90 days warranty after a repair. But you can't buy AppleCare anymore in those extra months after the first year, afaik...
FWIW, I think the iPod lawsuit is completely bogus, but the people suing over the iBook most certainly do have a case imho. I guess the main problem for Apple is all those schools that bought iBooks...
...over the years I've recieved exactly TWO Norwegian spams - from "Trondelag Teater" and "freewave.no" Of course, I'm pretty careful with my "official" mail, I keep various other junk accounts for other stuff. But the US spam (presumably) keeps coming in, viagra, 411 scams, mortgages, gambling, whatever.
OTOH, this does prove it's a good law though, and that anti-spam laws do help (regardless of what a lot of people claim). The problem is simply that we need such laws also in other countries.
Well, I heard only a week or so ago that the European Union was going to make sending spam illegal in the near future, or has already done so.
The EU directive has been approved already for quite a while, and several countries have already implemented it in national law (e.g. Belgium did so in March 2003).
Unfortunately, as this article on the Register points out, most spam comes from outside of the EU, or turns out to be untraceable anyway... so the question is if this new legislature would have any noticeable effect.
At the very least it will prevent us from getting in a situation like the US is in currently: virtually all spam is sent by or on behalf US "companies" (see Spamhaus' ROKSO list), you can easily prove that (not by tracing the emails through open proxies, but using the spamvertised websites), yet you can't do anything against them.
That's what they do already: the version of gcc that ships with Mac OS X contains several optimizations (30+ last time I heard) which are not in the main gcc. They try to keep that at a minimum though, since every time they sync they have to verify whether nothing broke their code.
Exactly, this is the problem. I want the ability the force all (non-critical, non-realtime) processes not associated with whatever window is currently in focus to freeze completely, until the process associated with the window in focus is idle, waiting for neither CPU nor disk.
The problem with this scheme is that you can get deadlocks. Suppose the high and low priority processes share a block of memory, and that the high priority process is waiting for a spinlock in that block of memory to be cleared (basically a variable which is currently 1 and which has to be set to 0 by the process/thread that owns the lock) by the low priority process.
In your scheme, the low priority process will never get a chance to run, because the high priority process is running all the time. So the high priority process keeps running forever doing nothing useful.
In other words I am suggesting a system option to do this: if the process associated with the window that's in focus is not in the idle state, make everything else freeze.
So if your email client was checking mail in the background, its connection will time out, file sharing will completely die, background printing fails due to communication time-outs,...
Or even this: an option to freeze everything else unless the process associated with the window that's in focus has been idle for at least x seconds (user settable). This latter would prevent swapping out while the user is typing.
These kind of things were possible in older systems (e.g. Mac OS 9 and below froze every other application while you were dragging a window around), but modern operating systems cannot work that way anymore. You don't want downloads to stop when they're in the background, firewalls to stop working because they're never the frontmost application etc...
outlaw the act itself (spam, booze) not the crimes that result from the act (bandwidth theft, drunk driving
The difference is that spamming per definition consists of -among other things- bandwidth theft, while alcohol by no means always results in drunk driving.
I have found nothing about this in my warranty document. Are you certain of this? Apple doesn't support Linux (in the sense that if you have some kind of problem which only occurs under Linux, they won't give you support for it), but running Linux shouldn't void your warranty.
getting a patent or filing a patent, can at least provide some tangible property/proof of a concept, that can enable a startup or small firm to get investors, thus building their business. Its the business standpoint that most people here on slashdot don't consider
If there are no software patents, investors cannot require them. It's really as simple as that. Software patents are the.com hype all over again: companies are not supposed to have a good product, good customer relationship management or even visionary management. No, they have to have something called "intellectual property" (be it in the form of patents or anything else).
How long till that bubble bursts as well? Unless you're a company like Eolas, you don't build a business on software patents. They also don't really help, except if you are attacked by someone else for infringement on their patents (which again is not possible if there are no swpats in the first place). For a business to be succeed, it needs actual products, customers, etc, not some pieces of paper that let it appropriate some general principles when implemented on a computer.
People may complain about abuse, but isn't it better that via the patent system, people disclose their inventions instead of hiding behind trade secrets, thus allowing others to improve upon the initial invention?
There are several problems with this argument when applied to software patents, as nicely highlighted in the recent FTC study (the link is a summary to the swpat-related stuff, a link to the full study is available at the top of that page as well).
First of all, a lot of software patents are simply business methods implemented in a computer program. By their very nature, if such business methods are put to use, they become public (since you use them to interact with the public).
Secondly, most companies forbid their employees from reading any patents, for the fear that it will result in them being sued for willful infringement (making them risk tripple damages and attorneys fees)! Besides, in general the developers can put very little of the patents to use because of all the legalese and the absence of source code.
As someone in the FTC study says "there's too much information and it's no longer meaningful".
That doesn't matter, because this is about patents and not about copyright. The patents actually are on specific features, so if you take them away, you don't infringe the patents anymore.
Slashdot Mirror
If you place one Lego brick on the ground, it will just sit there. Put one or two on top, and you'll still have no problem. If you put 100 Lego bricks right on top of each other, you get a very unstable structure. So even in this very simple physical environment, things that behave one way in the small, may behave quite differently when used in a more complex whole.
With software, this is not the case. One small, sound logical reasoning remains just as valid in a larger whole. Software development by its very nature consists entirely of combining all kinds of small logical constructs into a, non-obvious or not, bigger whole. And software is being innovated every day, because if you don't innovate, your competition will and you will lose whatever edge you had.
Please read this study, carried out by the Max Planck Institute and the Fraunhofer institute (by no means anti-swpat establishments) for more on why software is different and how software development works differently. For a more philosophical approach, in case you want to understand why those empirical results are what they are, see this page.
But software patents don't work that way, as shown by the recent FTC study (link is to a summary of all swpat related stuff, link to the full study is available at the top for you to verify, should you think that the person who created that page misrepresented the facts).Some quotes:
- One panelist stated that "the [patent] system discourages you from looking very hard [at patent disclosures] because
... simply by virtue of poking around to find out what patents exist you expose yourself to willfulness claims which can triple the amount of damages and exposure to attorney's fees."
- The panelist summed up the problem with the statement "there's too much information and it is no longer meaningful."
Computer scientists do not go to patent databases to look up information on how to do something. I guess that's why these same giant corporations are arguing that they need software patents, because it's supposedly so easy to reverse engineer and reimplement their precious techniques. They just want it all: very strong protection for abstract things which do not require such guarantees in order for them to be made public.How come? Because you publish software, you don't manufacture it (just like you don't publish an industrial valve or a chemical reaction, you manufacture or conduct those). Publishing by definition is "making public". It's the same with business methods: either you use a business method and it automatically becomes public, or you don't use it.
Software patents are a giant corporation's wet dream, because they pretty much only benefit the patent holders.
In case of "methods of technology that use these algorithms", you do not get a patent on the algorithm that is implemented in the technical apparatus, but on the way you implemented it technically. In software, you get copyright protection for the way you implement an algorithm (because it's more like writing a story based on an idea, than on constructing a valve based on some physical properties of the materials it's made of).
Software patents however, give you patent protection for an algorithm, which is a mixture (from hell) of both principles. As such, they lead to enormously broad (you get exclusive usage of some principle) and conflicting (copyright grants me the right to publish my program, but a patent owner forbids it because it uses a patented technique) protections.
In the European pro-swpat camp, the reasoning is that a computer executing a computer program is no longer a computer program as such and thus patentable, at least if the computer program allows you to make more efficient use of the computer than without it (or with other known programs). It's really stupid of course... What else can you do with a computer program than run it on a computer? Use it as wall paper?
If you make a processing device coupled to a memory, input device and screen guess the word you want to type, you violate that patent. The more specific claims simply add different kinds of keyboards you can use, that you can also use the frequency of chosen words to guide suggestions, allowing people to turn on/off this guiding algorithm, showing a list of possible words if there's ambiguity and let the user choose from those etc... Not really things that narrow the scope of the claims very much.
They don't take the right in their own hands, i.e. they don't launch attacks on ISP's, don't hack spammer websites, don't try to crash open proxies, ... They simply publish a list of IP-address of what they claim are spam-friendly ISP's.
- dslreports.com has address 209.123.109.175. That address only appears in a level 2 listing. Very few people use level 2 listings, the "real" SPEWS are the level 1 addresses. What level 2 really means, is explained in their FAQ (Q22).
- SPEWS did not add dslreports.com to their blacklist (search the linked page for dslreports, it's not mentioned). This does not make it less annoying for the owners of dslreports.com obviously, but there are differences. E.g., if a spammers moves, the blacklisting will be moved too, for dslreports.com it obviously wouldn't (no, that doesn't mean I think dslreports should simply move and shut up, I know things like that cost money).
- The blacklist that SPEWS publishes is an *opinion*. Everyone is free to follow their opinion or not and use it to (over-)protect their property or not. If an ISP uses it (or any other blacklist) and doesn't clearly inform its customers about that fact, then this ISP is at fault.
Nevertheless, I completely agree it's sad that the spammer situation has gotten so much out of hand that people resort to this kind of carpet-blacklisting to try to force ISP's to stop their spam support (as larger ip-blocks are only added when an ISP refuses to remove its spammers, or starts moving them around to non-blacklisted IP-addresses).It's however pretty much the last resort that other people have to do anything about it. If an ISP does not experience any significant harm from hosting spammers (and in facts profits largely from it) and does not want to remove them because it's the right thing to do, what else can you do to tell the ISP to FOAD if you don't want to become a vigilante?
(putting on asbestos suit)
I think it's fairly widespread, as Apple's CFO mentioned the unusually high number of warranty expenses for Powerbooks (due to initial white spots problem with the 15" AlBooks, now completely resolved according to them) and iBooks (not further specified) as the reason that their margins were lower than expected during the previous fiscal quarter.
Or just hold ctrl-alt from the start :)
Quote from Phil Schiller, VP worldwide product marketing Apple
There's also an article somewhere where Steve Jobs says more or less the same (and says he told the labels so), but I can't find it currently.Maybe it is indeed a manufacturing instead of a design defect. However, I mainly took issue with your assertion that "If people with failed logic boards did have AppleCare they'd be up and running again in under a week." People should not need to buy AppleCare to cover for design or more or less systematic manufacturing defects.
The standard 1 year warranty is meant to cover those and it's true that if you have bad luck, your machine may die 1 day after your warranty expires. However, if it turns out that an unusually large amount of machines show these failures after the standard warranty has expired, the manufacturer should take up responsibility for that. All imho, of course.
Well, the warranty runs out in 3 months actually, because the last repair is only a week ago and you get 90 days warranty after a repair. But you can't buy AppleCare anymore in those extra months after the first year, afaik...
FWIW, I think the iPod lawsuit is completely bogus, but the people suing over the iBook most certainly do have a case imho. I guess the main problem for Apple is all those schools that bought iBooks...
That's what they do already: the version of gcc that ships with Mac OS X contains several optimizations (30+ last time I heard) which are not in the main gcc. They try to keep that at a minimum though, since every time they sync they have to verify whether nothing broke their code.
In your scheme, the low priority process will never get a chance to run, because the high priority process is running all the time. So the high priority process keeps running forever doing nothing useful.
So if your email client was checking mail in the background, its connection will time out, file sharing will completely die, background printing fails due to communication time-outs,That is incorrect. Only the G5 has software controlled fans, and if no software controls them, they run at full speed all the time.
I have found nothing about this in my warranty document. Are you certain of this? Apple doesn't support Linux (in the sense that if you have some kind of problem which only occurs under Linux, they won't give you support for it), but running Linux shouldn't void your warranty.
How long till that bubble bursts as well? Unless you're a company like Eolas, you don't build a business on software patents. They also don't really help, except if you are attacked by someone else for infringement on their patents (which again is not possible if there are no swpats in the first place). For a business to be succeed, it needs actual products, customers, etc, not some pieces of paper that let it appropriate some general principles when implemented on a computer.
There are several problems with this argument when applied to software patents, as nicely highlighted in the recent FTC study (the link is a summary to the swpat-related stuff, a link to the full study is available at the top of that page as well).First of all, a lot of software patents are simply business methods implemented in a computer program. By their very nature, if such business methods are put to use, they become public (since you use them to interact with the public).
Secondly, most companies forbid their employees from reading any patents, for the fear that it will result in them being sued for willful infringement (making them risk tripple damages and attorneys fees)! Besides, in general the developers can put very little of the patents to use because of all the legalese and the absence of source code.
As someone in the FTC study says "there's too much information and it's no longer meaningful".
That doesn't matter, because this is about patents and not about copyright. The patents actually are on specific features, so if you take them away, you don't infringe the patents anymore.