I'm suggesting that 90% of the ones that are currently granted should be rejected. The Australian system is actually a good idea - it means that inventors can have their rubber stamp cheaply, and that it doesn't arm the patent trolls. But seriously, when more than half the patent suits are brought by non-practising entities (the balance tipped last year), and when patent thickets are so severe that innovators have to just ignore the patents and hope not to be sued... the system is broken. We should just scrap the whole thing. Patents help lawyers, and sometimes as a paper-trail for VC-funding. But in reaility, the whole patent system is now parasitic upon inventors and manufacturers.
No. In the (exceptionally rare) case where a patent is genuinely a good and valid one, the owner can get to prove it is valid when he goes to court. In the common case (trivial and bogus patents), the failure of the patent-office to deny them would not hurt the defendant.
In general, however, I think that the entire patent system is worse than useless, and we should simply abolish "intellectual monopolies".
To address your point, many people have the mis-perception that patents help small inventors. This isn't true. Small inventors are far likely to be harmed by the predatory actions of a large company (which uses its own patents to crush the small guy) than they are likely to be protected. To put it another way, If I have an idea, I want the right to use it. Patents give me the right (which I don't want) to destroy your business; they also cause me the risk (which I greatly fear) that someone else wilI come along and destroy mine.
A large amount of the trouble with patents comes from the fact that:
* The patent office doesn't have the resources to properly validate platent claims. They basically grant anything, and assume that validity will be litigated in the courts. * The courts tend to assume that anything granted must be valid.
So, why not change it to:
* The patent office merely registers the patent filing. It acknowledges the inventor's name, and publishes the details. but, at this stage, the patent is not deemed valid.. * When there is an actual patent suit, this is the time when the patent is carefully examined, and the question of validity can be debated in court.
We just deployed 3x Pi in a warehouse. I have to say, I'm really impressed with them. They are small, robust, and best of all, fanless (our last Mini-Itx died from dust-inhalation). System upgrades are easy - just swap over the SD card.
Just a couple of gotchas: * Overclocking isn't just about heat (I added a heatsink and the CPU runs cool). The jump from 950MHz to 1Ghz is a very steep one (it suddenly bumps up all the other system clocks by a large amount) and this can make it unstable, corrupting the filesystem. 950 seems to be reliable.
* Power for USB (especially WiFi) is dodgy. Hotplugging a dongle will make the Pi reboot from brownout. It seems to be worse because the "5V" supplies aren't actually 5V. I tested several; surprisingly, the branded Nokia/HTC ones put out about 4.7V, whereas the unbranded ones are nearer 4.9. I suspect that in a USB supply that is really designed to charge a 3.7V LiPo cell, the more energy efficient ones may aim to come in slightly under 5V to reduce waste. Even with the newest model B rev 2, there is still one polyfuse on the input: I shorted this to gain another 10mV.
Anyway, I really want a Model C, perhaps with a 1.5GHz CPU, 2GB of RAM, 4 USB ports, embedded Wifi/bluetooth, and a better power supply.
Well, how we decide is fairly simple: look at a selection of open source code, and count the function calls, then filter out anything too complex. I said C, because many other languages are actually written in C, or use similar primitives. Also, I would suspect that if atoi() were implemented in hardware, strtol() could use that as a part. Printf() is indeed horridly complex - it can take a significant fraction of 1ms to run. But what would happen if we devoted, say 5M transistors to implementing a decent chunk of libc and friends. Yes it would be very hard to do (and vulnerable to bugs that make the "Pentium" bug look trivial) - but it could allow 100x improvement in effective clock speed, OR dramatically drop the power consumption.
How about implementing just a few of the most common C-library functions in dedicated hardware. For example, atoi(), strlen(), or printf(). Although the software routines are highly optimised, they still take hundreds to thousands of cycles. Dedicated libc functions would require a significant amount of chip die space, BUT, they would be really power-efficient - powered off most of the time, and simply used when needed. Imagine being able to use these functions as single-cycle commands... even if the core ran at 100MHz, the performance would be amazing. Essentially it lets us trade a few hundred thousand transistors (now very cheap) for a few mW (still rather valuable).
When South Africa did this (to black people, rather than women), under Apartheid, the civilised world rightly condemned it, and imposed trade sanctions. Where are the trade embargoes on Saudi Arabia? They're in contravention of the UN declaration of Human Rights.
What I mean was... how is an ordinary developer meant to find out what the current state of play is? I want to keep my site HTML5 compliant and fast. I don't want any backward compatibility bloat in either CSS or JS to hack cosmetic support into old browsers...as long as the site remains functional. Yet many instances of documentation either only work in the very latest browser, or they have lots and lots of legacy cruft from 10 years ago just to keep IE6 (or even NS4) happy.
But it's very difficult for the developers to know when a "de-facto" standard can be used and then droppped. For example, I use -moz-border-radius, -webkit-border-radius. Given that this is purely cosmetic, I'd like to know at what point I can move to just border-radius and have it work in most browsers (by which I mean anything released in the last 2 years; cosmetic support in older ones doesn't matter). For example, there was a time when firefox supported "-moz-border-radius" but didn't support the unprefixed border-radius. But with both CSS and Javascript, it's very hard to know when we can drop support for old stuff. Essentially, we need a way to exipre old documentation.
There is a difference here... while it might be reasonable to object to offensive calls that are directed TO someone specific, this one isn't.
We should celebrate the freedom to burn poppies, just as we celebrate the right to wear them. We should honour those men and women who died for our country - but nor should we forget those whom our soldiers killed. And just maybe a spot of anti-establishment protest is necessary on Remembrance Sunday... to remember WW1 in particular, it's important to remember the dead, and that we do not forgive and forget the politicians who got us there.
Well... consider the recent case in the UK where (good) Costa Coffee pays its "fair share" of Corporation Tax, but (naughty) Starbucks doesn't. [Starbucks UK license their "IP" from the corporate parent for a large cost, and thus ship the money out of the country]. Both companies compete in essentially the same market, and substantially on price.
If VAT went up, and corp. tax went down, then Costa could afford to maintain their current price, and their current after-tax profit margin. Starbucks, on the other hand, would either have to put their price up (and lose customers), or they'd have to drop their costs (and have less proffits to send offshore).
That's not true... many essentials (eg non-restaurant food) are zero-rated, or less-rated for VAT. (Though I would agree that if VAT were made to replace corp. tax, then we might want to look at increasing the list of essentials, or have a half-rate for certain things).
Why not simply re-balance the tax-take. Over 3 years, ramp corporation tax down to 0, and ramp VAT up to 30% (while perhaps increasing the scope of some of the exemptions). Problem solved.
I'm with you here... I went from KDE3.5 -> Gnome2 -> XFCE4 -> KDE 4.6+ But I recently resurrected an old machine with a KDE 3.5 install. What shocked me was how much better looking it is than KDE 4.x. We can configure 4.x to do everything right, and at least it doesn't crash any more... but it still has this obsession with featureless large areas of gray, and hiding all the obvious borders between UI elements. And my favourite clock (the 7-segment one) still hasn't come back!
One other thing i really miss from Gnome was gconf: it let us *scriptably* configure the environment. KDE does have kwiteconfig, which does the same thing.... but only after you magically infer which key in which file needs to be edited.
P.S. Any chance of KDE allowing Gtk-keybindings to take effect? Ctrl-A should be Emacs-like (go to start of line), not Windows-like (Select All). QT apps work under KDE, GTK apps work under Gnome, but there is no way to change the default keybindings in Firefox/Thunderbird when in KDE!
If I were Moto, I'd offer Apple a choice: $20 per device, OR a promise that neither party would ever sue the other for any patent (and Google would have to cover Android too) except if countersuing.
I always thought it was the backlight that used the power? My phone can be on standby for 3 days, yet 1 hour of ebook reading in flight mode kills the battery.
Exactly. I've got 2048x1535 on my 15" T60p (upgraded with a new-old-stock panel that was manufactured in 2002!) and this is pretty amazing. Part of the problem is the "widescreen" misnomer: "shortscreen" would be more accurate (given that an x-inch widescreen has less area than an x-inch 4:3).
Interestingly, different people perceive these things differently. I personally find that the flicker caused by LEDs (especially at 50Hz, without capacitors) is really annoying.
But the real problem is that one can't really use "colour temperature" to get a good understanding of lighting quality. You have to look at the distribution too.
Colour temperature is essentially a "weighted average" of the emission spectrum. But white LEDs have a strongly bi-modal spectrum (basically blue + yellow), which looks nothing like the smooth blackbody spectrum from tungsten/tungsten-halogen. So you can match colour temperature, but have a widely different spectrum.
Also, while the human eye has only 3 separate types of sensor (roughly "R,G,B"), with their own response curves, the resonances of the dyes in clothing, paint, or skin pigments have their own resonances too. The result is that, even if you can't distinguish LED and tungsten light from each other (when looking straight at the bulb), you can tell the difference in the scene that is illuminated.
*Shielded* wiring: especially the Cat6. We just networked our house, and we installed:
- lots of Cat5E
- a "lutron" lighting control system
- electronic curtains.
The lutron generates a lot of RFI, especially at mid-level dimming. The interference can sometimes make the curtains jitter, as if we had a ghost! Our eventual solution was to change the curtain switch so that it powered down the system when it wasn't supposed to move.
Build a comms-rack, expect that area to be noisy, try to do something sensible with the resulting heat.
[Lastly, don't use LED lighting; use halogen. It's just better to live with. ]
You already can revoke permissions (in cyanogen at least), but it usually breaks the app. What we actually need is to be able to sandbox the app, and grant permissions only to "fake" data. Eg the app can have my phone number (but not the true one), or my position (but be hardcoded where I put it), or access the internet (but always get faked 404s), etc.
Yes... but Mandriva did it *really* badly. You could pay, but only $60 (nothing less). And for that sum, one got a DVD and printed packaging, shipped internationally 3 weeks after the release. So I was spending $60, knowing that $30 was going completely to waste on costs, for a useless DVD and shipping. I'd have been much happier to donate $20 each time, with the aim of "we'll support you in order to keep the service free for those who need it".
For that matter, let them learn with a real PS/2 keyboard - just not connected to anything. That would still be quite useful for the look and feel part.
[Otherwise, PS2 keyboards are easy to interface and power - you could build something cheap ($10) out of a PIC and a 1 line LCD display]
Slashdot Mirror
I'm suggesting that 90% of the ones that are currently granted should be rejected.
The Australian system is actually a good idea - it means that inventors can have their rubber stamp cheaply, and that it doesn't arm the patent trolls.
But seriously, when more than half the patent suits are brought by non-practising entities (the balance tipped last year), and when patent thickets are so severe that innovators have to just ignore the patents and hope not to be sued... the system is broken.
We should just scrap the whole thing. Patents help lawyers, and sometimes as a paper-trail for VC-funding. But in reaility, the whole patent system is now parasitic upon inventors and manufacturers.
Well, my assumption of invalidity would be the same as innocent until proven guilty.
No.
In the (exceptionally rare) case where a patent is genuinely a good and valid one, the owner can get to prove it is valid when he goes to court.
In the common case (trivial and bogus patents), the failure of the patent-office to deny them would not hurt the defendant.
In general, however, I think that the entire patent system is worse than useless, and we should simply abolish "intellectual monopolies".
To address your point, many people have the mis-perception that patents help small inventors. This isn't true. Small inventors are far likely to be harmed by the predatory actions of a large company (which uses its own patents to crush the small guy) than they are likely to be protected. To put it another way, If I have an idea, I want the right to use it. Patents give me the right (which I don't want) to destroy your business; they also cause me the risk (which I greatly fear) that someone else wilI come along and destroy mine.
A large amount of the trouble with patents comes from the fact that:
* The patent office doesn't have the resources to properly validate platent claims. They basically grant anything, and assume that validity will be litigated in the courts.
* The courts tend to assume that anything granted must be valid.
So, why not change it to:
* The patent office merely registers the patent filing. It acknowledges the inventor's name, and publishes the details. but, at this stage, the patent is not deemed valid..
* When there is an actual patent suit, this is the time when the patent is carefully examined, and the question of validity can be debated in court.
Think of this as "lazy-evaluation" for patents.
We just deployed 3x Pi in a warehouse. I have to say, I'm really impressed with them. They are small, robust, and best of all, fanless (our last Mini-Itx died from dust-inhalation). System upgrades are easy - just swap over the SD card.
Just a couple of gotchas:
* Overclocking isn't just about heat (I added a heatsink and the CPU runs cool). The jump from 950MHz to 1Ghz is a very steep one (it suddenly bumps up all the other system clocks by a large amount) and this can make it unstable, corrupting the filesystem. 950 seems to be reliable.
* Power for USB (especially WiFi) is dodgy. Hotplugging a dongle will make the Pi reboot from brownout. It seems to be worse because the "5V" supplies aren't actually 5V. I tested several; surprisingly, the branded Nokia/HTC ones put out about 4.7V, whereas the unbranded ones are nearer 4.9. I suspect that in a USB supply that is really designed to charge a 3.7V LiPo cell, the more energy efficient ones may aim to come in slightly under 5V to reduce waste. Even with the newest model B rev 2, there is still one polyfuse on the input: I shorted this to gain another 10mV.
Anyway, I really want a Model C, perhaps with a 1.5GHz CPU, 2GB of RAM, 4 USB ports, embedded Wifi/bluetooth, and a better power supply.
Well, how we decide is fairly simple: look at a selection of open source code, and count the function calls, then filter out anything too complex. I said C, because many other languages are actually written in C, or use similar primitives. Also, I would suspect that if atoi() were implemented in hardware, strtol() could use that as a part. Printf() is indeed horridly complex - it can take a significant fraction of 1ms to run.
But what would happen if we devoted, say 5M transistors to implementing a decent chunk of libc and friends. Yes it would be very hard to do (and vulnerable to bugs that make the "Pentium" bug look trivial) - but it could allow 100x improvement in effective clock speed, OR dramatically drop the power consumption.
How about implementing just a few of the most common C-library functions in dedicated hardware. For example, atoi(), strlen(), or printf(). Although the software routines are highly optimised, they still take hundreds to thousands of cycles. Dedicated libc functions would require a significant amount of chip die space, BUT, they would be really power-efficient - powered off most of the time, and simply used when needed. Imagine being able to use these functions as single-cycle commands... even if the core ran at 100MHz, the performance would be amazing. Essentially it lets us trade a few hundred thousand transistors (now very cheap) for a few mW (still rather valuable).
When South Africa did this (to black people, rather than women), under Apartheid, the civilised world rightly condemned it, and imposed trade sanctions.
Where are the trade embargoes on Saudi Arabia? They're in contravention of the UN declaration of Human Rights.
What I mean was... how is an ordinary developer meant to find out what the current state of play is? I want to keep my site HTML5 compliant and fast. I don't want any backward compatibility bloat in either CSS or JS to hack cosmetic support into old browsers...as long as the site remains functional. Yet many instances of documentation either only work in the very latest browser, or they have lots and lots of legacy cruft from 10 years ago just to keep IE6 (or even NS4) happy.
But it's very difficult for the developers to know when a "de-facto" standard can be used and then droppped.
For example, I use -moz-border-radius, -webkit-border-radius. Given that this is purely cosmetic, I'd like to know at what point I can move to just border-radius and have it work in most browsers (by which I mean anything released in the last 2 years; cosmetic support in older ones doesn't matter). For example, there was a time when firefox supported "-moz-border-radius" but didn't support the unprefixed border-radius.
But with both CSS and Javascript, it's very hard to know when we can drop support for old stuff. Essentially, we need a way to exipre old documentation.
There is a difference here... while it might be reasonable to object to offensive calls that are directed TO someone specific, this one isn't.
We should celebrate the freedom to burn poppies, just as we celebrate the right to wear them. We should honour those men and women who died for our country - but nor should we forget those whom our soldiers killed. And just maybe a spot of anti-establishment protest is necessary on Remembrance Sunday... to remember WW1 in particular, it's important to remember the dead, and that we do not forgive and forget the politicians who got us there.
Well... consider the recent case in the UK where (good) Costa Coffee pays its "fair share" of Corporation Tax, but (naughty) Starbucks doesn't.
[Starbucks UK license their "IP" from the corporate parent for a large cost, and thus ship the money out of the country].
Both companies compete in essentially the same market, and substantially on price.
If VAT went up, and corp. tax went down, then Costa could afford to maintain their current price, and their current after-tax profit margin. Starbucks, on the other hand, would either have to put their price up (and lose customers), or they'd have to drop their costs (and have less proffits to send offshore).
That's not true... many essentials (eg non-restaurant food) are zero-rated, or less-rated for VAT.
(Though I would agree that if VAT were made to replace corp. tax, then we might want to look at increasing the list of essentials, or have a half-rate for certain things).
true... but if corporation tax were abolished, companies could keep their profit margins constant, while cutting pre-tax prices.
Why not simply re-balance the tax-take. Over 3 years, ramp corporation tax down to 0, and ramp VAT up to 30% (while perhaps increasing the scope of some of the exemptions). Problem solved.
I'm with you here... I went from KDE3.5 -> Gnome2 -> XFCE4 -> KDE 4.6+
But I recently resurrected an old machine with a KDE 3.5 install. What shocked me was how much better looking it is than KDE 4.x. We can configure 4.x to do everything right, and at least it doesn't crash any more... but it still has this obsession with featureless large areas of gray, and hiding all the obvious borders between UI elements. And my favourite clock (the 7-segment one) still hasn't come back!
One other thing i really miss from Gnome was gconf: it let us *scriptably* configure the environment. KDE does have kwiteconfig, which does the same thing.... but only after you magically infer which key in which file needs to be edited.
P.S. Any chance of KDE allowing Gtk-keybindings to take effect? Ctrl-A should be Emacs-like (go to start of line), not Windows-like (Select All). QT apps work under KDE, GTK apps work under Gnome, but there is no way to change the default keybindings in Firefox/Thunderbird when in KDE!
I'm actually suggesting that Google (which, in this one instance, has the upper hand) could use it to force Apple into the Open Invention Network.
If I were Moto, I'd offer Apple a choice: $20 per device, OR a promise that neither party would ever sue the other for any patent (and Google would have to cover Android too) except if countersuing.
I always thought it was the backlight that used the power? My phone can be on standby for 3 days, yet 1 hour of ebook reading in flight mode kills the battery.
Exactly. I've got 2048x1535 on my 15" T60p (upgraded with a new-old-stock panel that was manufactured in 2002!) and this is pretty amazing.
Part of the problem is the "widescreen" misnomer: "shortscreen" would be more accurate (given that an x-inch widescreen has less area than an x-inch 4:3).
Interestingly, different people perceive these things differently. I personally find that the flicker caused by LEDs (especially at 50Hz, without capacitors) is really annoying.
But the real problem is that one can't really use "colour temperature" to get a good understanding of lighting quality. You have to look at the distribution too.
Colour temperature is essentially a "weighted average" of the emission spectrum. But white LEDs have a strongly bi-modal spectrum (basically blue + yellow), which looks nothing like the smooth blackbody spectrum from tungsten/tungsten-halogen. So you can match colour temperature, but have a widely different spectrum.
Also, while the human eye has only 3 separate types of sensor (roughly "R,G,B"), with their own response curves, the resonances of the dyes in clothing, paint, or skin pigments have their own resonances too. The result is that, even if you can't distinguish LED and tungsten light from each other (when looking straight at the bulb), you can tell the difference in the scene that is illuminated.
*Shielded* wiring: especially the Cat6. We just networked our house, and we installed:
- lots of Cat5E
- a "lutron" lighting control system
- electronic curtains.
The lutron generates a lot of RFI, especially at mid-level dimming. The interference can sometimes make the curtains jitter, as if we had a ghost!
Our eventual solution was to change the curtain switch so that it powered down the system when it wasn't supposed to move.
Build a comms-rack, expect that area to be noisy, try to do something sensible with the resulting heat.
[Lastly, don't use LED lighting; use halogen. It's just better to live with. ]
You already can revoke permissions (in cyanogen at least), but it usually breaks the app. What we actually need is to be able to sandbox the app, and grant permissions only to "fake" data. Eg the app can have my phone number (but not the true one), or my position (but be hardcoded where I put it), or access the internet (but always get faked 404s), etc.
Yes... but Mandriva did it *really* badly. You could pay, but only $60 (nothing less). And for that sum, one got a DVD and printed packaging, shipped internationally 3 weeks after the release. So I was spending $60, knowing that $30 was going completely to waste on costs, for a useless DVD and shipping. I'd have been much happier to donate $20 each time, with the aim of "we'll support you in order to keep the service free for those who need it".
For that matter, let them learn with a real PS/2 keyboard - just not connected to anything. That would still be quite useful for the look and feel part.
[Otherwise, PS2 keyboards are easy to interface and power - you could build something cheap ($10) out of a PIC and a 1 line LCD display]