99 quid and presumably a year long contract - they're more like 250 quid contract-free. And I have a 6600, and know for a fact that it's a lot thicker than any Palm on the market. Flat things fit in jeans pockets better.
IIRC, GSM was first introduced on the European continent on 900Mhz, and 1800 in the UK (because 900 was already in use at the time GSM was being introduced).
900MHz was used first in the UK, and 1800MHz followed, I believe. My first phone was single-band 900MHz, anyway. But, in the UK at least, both bands are thriving and 850MHz is just a(nother) weird thing that Americans do - there's no chance of it happening here.
$80k/per annum/ is in the right ballpark for renting the fibre you need, so I don't see why $80k fro equipment, plus minimal rent, should be too expensive. And companies, like people, can get loans.
I've got a perfectly good green-screen MGA monitor still. It used to be on my router box (complete with full-length MGA card - I also have a Herc one that does 720x384 1-bit graphics but I couldn't find it at the time).
Not bad, but the phosphor persistence is incredibly long, which makes scolling interesting. It's also a bit burned in, so when it's switched off you can see the 80x24 grid of characters...
I can live without being redirected to Kelkoo every time I search on something I'm interested in buying. Especially since its idea of what you want to buy is often far removed from what you ask for, despite a very specific set of search terms.
Ther are no system call with zero-copy domain sockets(it's memmapped). What, you didn't think we talked about this
at the lkml?
I'm curious: how do you indicate to a "domain socket" (I presume you mean one of "socket" or "unix domain socket") that you've written your data in this memmapped are and you now want it to be passed to the receiver? Do you have any design document references?
They're talking about using a funny self-invented communications system, but the method of communication would appear to be either stream- or message-based. If it's either of those, then surely implementing it over TCP instead is trivial, and the fact that their comms system relies (at the moment) on an x86 feature has no bearing, because it's easily replaced.
To be honest, I suspect that whatever they're doing could be replicated easily enough using a non-x86-specific method, but I haven't looked into the details. Given what they're doing seems to be faffing around with a chunk of shared memory, it might well be possible to do it using standard Unix shared memory systems and no hardware-specific code at all; it's always hard in these situations to decide if someone's inventing a new OS for sound reasons or because they just don't understand what existing OSes are capable of...
The point I was trying to make was that it's all very well saying "X is fast because it uses shared memory and zero-copy sockets to communicate to clients" but this is simply not true in all cases. I'm not arguing that clients can communicate with each other via the X server, but that's not relevant here.
You can beat zero-copy networking by having fewer context switches. Which can be done through a shared-memory system, admittedly, but making networking system calls will cause context switches, which are heavyweight on x86.
Of course, not every application uses local connections - networked apps don't have the option of shared memory. It would be nice if they could run smoothly, too...
Since they're using a stream-oriented protocol, channelling it over a network would be easy.
And none of the design is tied to x86 machines. They plan to use an communcations mechanism with an x86-based implementation, but, as they say, it's not that it couldn't be ported. (Since it's a stream-based protocol it would be easy enough to replace it with sockets or anything else anyway.)
And, if you'd bothered to read the article, they compare their communications method with sockets, and demonstrate the speed advantage. But to my mind, the speed of their native communications system is a separate issue from the SVG/XML concepts they're advocating, and the two parts should be considered separately.
I can't see how creating more X alternatives or derivatives are going to help linux become more mainstream on the desktop?
I can't see where anyone said this had anything to do with Linux (since it's running on a different OS) or the desktop mainstream (no mention of that anywhere).
Indeed, Google's searching for "to be" "or" "not" "to be", judging by the underlining it gives the words. Weird, though there's probably an explanation for this...
Great, not only can we have quantum computers that calculate all the possible answers at once, but now that we can superpose mirrors,we can display them *all at the same time*!
It's the storage library that keeps an index of pages. You need a display front-end and a webcrawler to go with it (there's some code around). It's GPL and it has some clever features.
The problem with sticking PuTTY on one of these things is that you have to trust the machine that you use PuTTY on not to be snooping your password. If you have PuTTY + your ssh keys, worse still. (Assuming that your parents/friends/local internet cafe aren't capable of or interested in this sort of thing is a kind of trust. However, you're also assuming they haven't been infected by the latest steal-your-computer-away virus...)
For this sort of thing, you want a widget that will sign data with your ssh keys without ever handing the key over to untrusted hardware, so that you can plug it in and - while your session is still monitorable - at least your authentication tokens aren't up for grabs.
And if anyone knows of such a widget, then please tell me...
You seem to have a very low opinion of my intelligence. Perhaps instead you should have thought about what I was suggesting before replying. I'm not suggesting that you redraw the image on the server with every mouse movement, which seems to be what you're thinking of.
Javascript can't alter images, but it can make it moderately simple to e.g. pick a zoom area (downbutton marks a point, dragging draws a box in another layer so that the user can see what they're doing, upbutton sends the request with the co-ordinates gleaned). The responsive graphical part of this is client-side, and the relatively infrequent graph redraw is server-side.
I'm sure, with a little imagination, you could come up with a range of operations that you could perform in a similar manner - possibly even as far as e.g. changing bar sizes, if you were prepared to put the time into writing the Javascript. Although in my experience more than a minimum amount of JS brings enormous cross-browser maintenance and browser crashiness problems...
There are also more basic applications of the same idea, where you use an image which you occasionally reload and a set of form elements that change the graph parameters. So, when someone changes the graph from a bar to a line, only the image, and not the whole page, is redrawn. If you look around enough share-price-graphing websites, you should find a few examples of this.
Slashdot Mirror
Richer Sounds website moves about as fast as iced treacle. And judging by comments I've found on the web, it's not just me that has that problem.
;-)
Fortunately, there's a local store
99 quid and presumably a year long contract - they're more like 250 quid contract-free. And I have a 6600, and know for a fact that it's a lot thicker than any Palm on the market. Flat things fit in jeans pockets better.
Seriously, i have a nokia 6600, what can the Zire's do that the 6600 cant.
Be bought for less money? Fit in your pocket?
IIRC, GSM was first introduced on the European continent on 900Mhz, and 1800 in the UK (because 900 was already in use at the time GSM was being introduced).
900MHz was used first in the UK, and 1800MHz followed, I believe. My first phone was single-band 900MHz, anyway. But, in the UK at least, both bands are thriving and 850MHz is just a(nother) weird thing that Americans do - there's no chance of it happening here.
$80k /per annum/ is in the right ballpark for renting the fibre you need, so I don't see why $80k fro equipment, plus minimal rent, should be too expensive. And companies, like people, can get loans.
I've got a perfectly good green-screen MGA monitor still. It used to be on my router box (complete with full-length MGA card - I also have a Herc one that does 720x384 1-bit graphics but I couldn't find it at the time).
Not bad, but the phosphor persistence is incredibly long, which makes scolling interesting. It's also a bit burned in, so when it's switched off you can see the 80x24 grid of characters...
Mutter mutter RISCOS mutter mutter...
I can live without being redirected to Kelkoo every time I search on something I'm interested in buying. Especially since its idea of what you want to buy is often far removed from what you ask for, despite a very specific set of search terms.
Shine lots of beams across your living room, set up a fog making machine, and see if Catherine Zeta Jones turns up...
I'm curious: how do you indicate to a "domain socket" (I presume you mean one of "socket" or "unix domain socket") that you've written your data in this memmapped are and you now want it to be passed to the receiver? Do you have any design document references?
They're talking about using a funny self-invented communications system, but the method of communication would appear to be either stream- or message-based. If it's either of those, then surely implementing it over TCP instead is trivial, and the fact that their comms system relies (at the moment) on an x86 feature has no bearing, because it's easily replaced.
To be honest, I suspect that whatever they're doing could be replicated easily enough using a non-x86-specific method, but I haven't looked into the details. Given what they're doing seems to be faffing around with a chunk of shared memory, it might well be possible to do it using standard Unix shared memory systems and no hardware-specific code at all; it's always hard in these situations to decide if someone's inventing a new OS for sound reasons or because they just don't understand what existing OSes are capable of...
The point I was trying to make was that it's all very well saying "X is fast because it uses shared memory and zero-copy sockets to communicate to clients" but this is simply not true in all cases. I'm not arguing that clients can communicate with each other via the X server, but that's not relevant here.
You can beat zero-copy networking by having fewer context switches. Which can be done through a shared-memory system, admittedly, but making networking system calls will cause context switches, which are heavyweight on x86.
Of course, not every application uses local connections - networked apps don't have the option of shared memory. It would be nice if they could run smoothly, too...
Since they're using a stream-oriented protocol, channelling it over a network would be easy.
And none of the design is tied to x86 machines. They plan to use an communcations mechanism with an x86-based implementation, but, as they say, it's not that it couldn't be ported. (Since it's a stream-based protocol it would be easy enough to replace it with sockets or anything else anyway.)
And, if you'd bothered to read the article, they compare their communications method with sockets, and demonstrate the speed advantage. But to my mind, the speed of their native communications system is a separate issue from the SVG/XML concepts they're advocating, and the two parts should be considered separately.
I can't see where anyone said this had anything to do with Linux (since it's running on a different OS) or the desktop mainstream (no mention of that anywhere).
Indeed, Google's searching for "to be" "or" "not" "to be", judging by the underlining it gives the words. Weird, though there's probably an explanation for this...
Great, not only can we have quantum computers that calculate all the possible answers at once, but now that we can superpose mirrors,we can display them *all at the same time*!
1. no-one mentioned money;
2. XML namespaces are on XML tags, not URLs.
And this is insightful? Hmm...
Wh, exactly, has an indoor patio?
And if that didn't help, bollocks is a collquialism for nadgers.
It's the storage library that keeps an index of pages. You need a display front-end and a webcrawler to go with it (there's some code around). It's GPL and it has some clever features.
The problem with sticking PuTTY on one of these things is that you have to trust the machine that you use PuTTY on not to be snooping your password. If you have PuTTY + your ssh keys, worse still. (Assuming that your parents/friends/local internet cafe aren't capable of or interested in this sort of thing is a kind of trust. However, you're also assuming they haven't been infected by the latest steal-your-computer-away virus...)
For this sort of thing, you want a widget that will sign data with your ssh keys without ever handing the key over to untrusted hardware, so that you can plug it in and - while your session is still monitorable - at least your authentication tokens aren't up for grabs.
And if anyone knows of such a widget, then please tell me...
Wow, apologies on Slashdot. The end of the world is nigh. ;-)
Javascript can't alter images, but it can make it moderately simple to e.g. pick a zoom area (downbutton marks a point, dragging draws a box in another layer so that the user can see what they're doing, upbutton sends the request with the co-ordinates gleaned). The responsive graphical part of this is client-side, and the relatively infrequent graph redraw is server-side.
I'm sure, with a little imagination, you could come up with a range of operations that you could perform in a similar manner - possibly even as far as e.g. changing bar sizes, if you were prepared to put the time into writing the Javascript. Although in my experience more than a minimum amount of JS brings enormous cross-browser maintenance and browser crashiness problems...
There are also more basic applications of the same idea, where you use an image which you occasionally reload and a set of form elements that change the graph parameters. So, when someone changes the graph from a bar to a line, only the image, and not the whole page, is redrawn. If you look around enough share-price-graphing websites, you should find a few examples of this.