They do try this among other things. TCP/IP offload engines on NICs are a ripe field seldom reported. With NICs that process packets it's conceivably possible to exploit the host without the OS giving permission for code to run with a corrupt packet. NICs with offload engines have their own primitive CPU and memory, DMA access to the system memory, and hardware interrupts to key system CPU actions. Pretty cutting edge stuff though. I haven't heard of anybody actually exploiting that end. That would be a bad one.
If you could corrupt a server TCP/IP offload NIC with a bad packet and get it to execute code in the NIC, you could conceivably compromise the server at the hardware level outside of the OS permissions system regardless of the OS. Scary stuff, that. It might work with smart TV's, teleconferencing systems, or similar gear. With Infiniband, frankly, it's done in one - but the bad guys aren't targeting HPC clusters yet I don't think, because that gear doesn't generally have access to do the things bad guys want to do (send mail mostly) and the people who herd those boxes are first rate. If Infiniband comes mainstream that could be an issue, and that's looking more likely with iSCSI over infiniband for SSD storage SANs gaining traction in the market. It's hard to pass up 250k IOPs in a 2U box like Whiptail has got if you need to do VDI on a budget and need end-user buy-in.
This is highly speculative. It would probably involve a disgruntled former high-level engineer, or a nation-state level cyber-intelligence effort. It might make a good fiction device. You can be sure that if I can think of it somebody's working on a dumbed down toolkit for it that your kids could use as a game.
As for malware injecting own-source packets, that's how it's generally done. It's called a "raw socket." Steve Gibson has been campaigning against them being available to user apps for a decade or more - to the point that maybe he's lost his edge. Raw sockets are how you forge the return IP address when you're pulling a Joe Job. Not that I would know anything about that. But malware that's running on the system has generally "got root," downloaded its toolkit and will be interacting with the NIC on a hardware level without the operating sytem's permission anyway. You don't need to have a TCP/IP stack to lay some forged packets on the NIC because your app isn't expecting packets back, doesn't need to handle sessions and whatnot. The OS tells you how to access the NIC at the hardware level and helpfully gets out of the way.
Java's just a language. It's not important. All that code can be ported to a language that doesn't have Larry Ellison attached to it. Most of the work happens in libraries anyway. Projects that are dependent on a specific languages or platforms that are not fully free are castles built on sand. Eventually someone will erode away the beach, and the castle will fall down. Always. Nothing of value was lost because it was a tissue fort - a mockup - a grand illusion the whole time. It was a distraction, and a lot of people didn't fall for it.
And of course Walmart is also selling the ipad and iPod, two of the top three gifts in Santa's email inbox this holiday season. For the other of the top three they offer hundreds of iPhone accessories.
Google is selling this phone because it advances the technology and their phone partners wouldn't sell it. Expect them to sell an Android + Snapdragon slate for the same reasons.... I doubt Google even wants to sell phones - I think they just want to get the new good technologies adopted so that people can get used to Internet everywhere quicker. This serves their bottom line because when most people use the Internet they use Google services, which Google sells ads on. You can't very well sell Internet ads to be viewed by people who aren't close to a browser. [me]
It links to this interesting article where the CEO of Asus was backing away from the Android smartbook they had recently pulled in mid-computex.
"Currently, I still don't see a clear market for smartbooks," said Jerry Shen, CEO of Asustek Computer, during an investors' conference in Taipei.
So he pulls the Linux Snapdragon smartbook and shows up a few days later at an investors conference - just before the W7 launch - flanked by reps from Microsoft and Intel - probably glancing cautiously from one to the other hoping nothing bad happens to his precious W7 netbooks (little does he know...). And he gives a carefully prepared speech about how Intel and Microsoft are going to crush their enemies, see them driven before them, hear the lamentation of their women...
And now world & dog sees Microsoft as a fading power, Apple mobile platforms - and mobile platforms in general - as the next generation of user interface, and suddenly now he sees a future in it again. Intel is driving as hard as they can to be the thing that gives people what they want. Microsoft? Let's just say the KIN didn't work out and WP7 has a steeper hill to climb than it might have. What a difference a year makes.
I love my Samsung Epic Android phone, but obviously I know I would not have any such thing if both Apple and Google had not dared to bring us change, each in their own way.
That article was about Google's Nexus 1 phones. Remember that Google shopped its candybar phone to every phone vendor and they wouldn't take it, so Google made it, sold a grip of them, and ushered in all this sweet tech we enjoy today. If they had not done so when they did, we'd not have seen the first good big-screen Android platforms until after WP7 launched, if ever. And now those phones are selling 20M units a quarter in the US alone, giving 44% market share, driving every phone vendor that builds it into profitability or record profitability, giving US non-AT&T networks a phone to sell that isn't absolutely pathetic, and putting money in the pockets of a vast economy of app developers and advertising buyers (and of course, Google).
The message is pretty clear. If Google gives you a reference platform, Run With It! Refusing is not going to keep them from bringing new tech to market. They don't want the manufacturing and retail money because they want to leave that business to their partners. It's a messy customer service business with low leverage. It's not their strong point. But if their partners won't give us progress, they aren't averse to bringing it directly and reaping a few billion in hardware revenue along the way. Microsoft and Intel used to be able to prevent progress, to prevent "cannabilization" of their established markets. But now those days are done. Vendors used to be able to hold off the releases with "tomorrow, tommorow" and "any day now". Any more? No. That's not going to fly. We'll have progress now whether the established hardware vendors are ready to give it or not. There will be no stalling any more.
Good point. Maybe one day Unicode will win out, or perhaps EBCDIC will have a resurgence. 'twixt now and then it's best to write the text in ascii, perhaps with a well-documented human-readable escape table for symbols that aren't represented - perhaps even a complete Unicode escape table current to the document. Then in 2050 when somebody wants to use the data again in BufTable or DECSHIN or whatever they're using then, they can rewrite the presentation filter and leave the underlying text in its pristine ascii condition for future data archaeologists. Those future data archaeologists, in 2100, will likely be writing those presentation filters in C, in ascii, if current trends hold true. They may not speak English, but if the raw text is in ascii, they'll be able to figure it out. Unicode? How many revisions will Unicode see between now and then? Thousands? The odds are slim we'll even be aligning on 8-bit words for text by then - but for ASCII one single rotting page from the manual from the first IBM PC or any of thousands of texts since then will be all they need for their Rosetta Stone.
For all of the past that we can see, and the foreseeable future, simpler is better. It's easier to rewrite the presentation layer than to go back and validate that your filter reliably mangled the text. Remember that when the text is approved as acceptably mangled, the source is almost always tossed. 40 years later would be a bad time to discover that the site of the holy grail was encoded in some forgotten scheme thrice since remangled.
Forgive me, but I've spent a disproportionate share of my programming time in data archaeology myself. I'm currently shepherding so much data from the '80's that it's quite not possible to validate that it's translating properly in a dozen man years - and this is a hobby for me. That data isn't 100MB. That was a lot when I started, but now my phone holds over 100x as much. Enterprise data shepherds with terabytes of legacy data, you have my sympathy.
Michael decided to use this huge amount of computer time to search the public domain books that were stored in our libraries, and to digitize these books. He also decided to store the electronic texts (eTexts) in the simplest way, using the plain text format called Plain Vanilla ASCII, so they can be read easily by any machine, operating system or software.
Since its humble beginnings in 1971 Project Gutenberg has reproduced and distributed thousands of works to millions of people in - ultimately - billions of copies. They support ePub now and simple HTML, as well as robo-read audio files, but the one format that has been stable this whole time has been ASCII. It's also the format that is likely to survive the longest without change. Project Gutenberg texts can now be read on every e-reader, smartphone, tablet and PC.
If you want to use Rich Text format, or XML, or PostScript or something else then fine - please do. But don't go trying to deprecate ASCII.
There will never be a lack of fail for you to exploit. Right now there are some good corners in deduplicated filesystems, binocular vision processing, and of course somebody's going to need to build the drag-and-drop CMS for HTML5. This is the song that never ends.
Obviously some people are not clear on the concept that some of us are looking to conclude a purchase when we go to Google, and the ads are actually part of the relevant results we're looking for.
Well if we're going to reference sci-fi the relevant issues in Footfall echo the themes given in Tunnel in the Sky given by the Dean of Science Fiction (Robert A. Heinlein) thirty years before, in 1955. Or Methusela's children in 1958. Kids these days.
If it's not legal, don't do it. This is a pretty basic policy that covers pretty much the entirety of your comment, and it's a fairly standard policy in corporations, schools, state, local and federal government IT departments. I do believe that violating this basic policy is grounds for termination just about everywhere - at least everywhere I've ever worked.
If your software vendor wants to make it illegal for you to move forward in a functional fashion there is a solution for that that doesn't involve breaking the law. It also doesn't involve using the next version of their product.
There will be some of that. That's why you don't send out just one generational ship. There's also the issue of a people who have been in the ship for so long that when they arrive at the destination they have agoraphobia and no longer want to get out of the ship. All of these perils and more await.
The setting out is easy. Men have been abandoning hearth and home to see what's beyond the edge of what we know since before we were Men - and taking our families with us too. It's our privilege as parents to make that choice for our progeny.
Men will span the reaches beyond the planets and after that the stars, or we will not. If not then we will one day find our end, and we may as well have never been.
As always, Garmin-Asus seeks the best for our consumers either on Android platform or on Windows platform. However, we see the potential of Android platform devices, so we are focusing on Android platform currently. - Steven Tu
Meanwhile Microsoft's VP, Corporate Communications Frank X. Shaw is over on Twitter right now trying to repair the damage done by today's CNN Money report. In case y'all want to wander over and lend him a hand.
He's got every book, every fact, every song in the world. If I'm shopping for something, he can show me the prices nearby that it's offered at. He can bring me anything that is sold in all the world. He can tell me how to get to anywhere, and sell me tickets too. He'll even show my my place in the Universe. He knows so much about me that when I ask him the news, he only gives me the news I find interesting - and he found this out because I told him.
It's threatening. Though he's not done anything mean or spiteful yet that much power can't be good. It seems every time I ask him for something he learns even more about what I want, and uses that information to serve me better and even faster. It's creepy. Somebody rescue me please!
I could just leave him home but he's so darned useful. What if I needed to know the emergency treatment for jellyfish sting or something, and he wasn't there?
Compare and contrast Ray Ozzie's farewell with that of another recent high-level departure, J. Allard. These men, at the heart of technology for all their adult lives, were in positions of the highest influence at Microsoft. They're obviously both brilliant, and not needing to cash a paycheck. They see a change coming - a huge change - and they want to be a part of it. They don't see that happening while they work in Redmond. So they go. But on the way out they look back at the poor souls they leave behind and they tell them in their farewell: "You too can be a part of this new world. You just have to think different." The door swings shut with a click and the obvious conclusion remains unsaid: "but you won't."
The newest enterprise versions will have deduplication in the onboard controller both to reduce the write count and increase the effective bandwidth. In two more generations this will be consumer level stuff. Internal write smoothing algorithms are already two generations old, as is extra storage to map out failed cells.
The very fact that we're having this discussion implies that there's a business model associated with being a one-stop third-party repository for all platforms, handling software sale (if cost), download, install, payment processing, vetting software vendors, uninstallation, recommendation, rating and reviewing, and that such a business will be profitable. Since it's a business model, at this moment there are businessmen all over the country filing papers for the corporations that are going to get in on the ground level. There will no doubt be several competing businesses offerring you this boon next month, each advertising "exclusives" and vying for your attention.
The obvious thing to do is to patent an API for the arbitrage of trust and control in competing integrated software market distributors and updaters, so that licensees could implement customer-driven subrogation and promotion of their preferred software markets and still perform their search and update duties in a fashion coordinated with each other. Then lock it in a drawer and forget about it for ten years.
Slashdot Mirror
Wow. Just like that it's already in the bargain bin. They didn't even wait a week. Sales must be absolutely dismal.
They do try this among other things. TCP/IP offload engines on NICs are a ripe field seldom reported. With NICs that process packets it's conceivably possible to exploit the host without the OS giving permission for code to run with a corrupt packet. NICs with offload engines have their own primitive CPU and memory, DMA access to the system memory, and hardware interrupts to key system CPU actions. Pretty cutting edge stuff though. I haven't heard of anybody actually exploiting that end. That would be a bad one.
If you could corrupt a server TCP/IP offload NIC with a bad packet and get it to execute code in the NIC, you could conceivably compromise the server at the hardware level outside of the OS permissions system regardless of the OS. Scary stuff, that. It might work with smart TV's, teleconferencing systems, or similar gear. With Infiniband, frankly, it's done in one - but the bad guys aren't targeting HPC clusters yet I don't think, because that gear doesn't generally have access to do the things bad guys want to do (send mail mostly) and the people who herd those boxes are first rate. If Infiniband comes mainstream that could be an issue, and that's looking more likely with iSCSI over infiniband for SSD storage SANs gaining traction in the market. It's hard to pass up 250k IOPs in a 2U box like Whiptail has got if you need to do VDI on a budget and need end-user buy-in.
This is highly speculative. It would probably involve a disgruntled former high-level engineer, or a nation-state level cyber-intelligence effort. It might make a good fiction device. You can be sure that if I can think of it somebody's working on a dumbed down toolkit for it that your kids could use as a game.
As for malware injecting own-source packets, that's how it's generally done. It's called a "raw socket." Steve Gibson has been campaigning against them being available to user apps for a decade or more - to the point that maybe he's lost his edge. Raw sockets are how you forge the return IP address when you're pulling a Joe Job. Not that I would know anything about that. But malware that's running on the system has generally "got root," downloaded its toolkit and will be interacting with the NIC on a hardware level without the operating sytem's permission anyway. You don't need to have a TCP/IP stack to lay some forged packets on the NIC because your app isn't expecting packets back, doesn't need to handle sessions and whatnot. The OS tells you how to access the NIC at the hardware level and helpfully gets out of the way.
Java's just a language. It's not important. All that code can be ported to a language that doesn't have Larry Ellison attached to it. Most of the work happens in libraries anyway. Projects that are dependent on a specific languages or platforms that are not fully free are castles built on sand. Eventually someone will erode away the beach, and the castle will fall down. Always. Nothing of value was lost because it was a tissue fort - a mockup - a grand illusion the whole time. It was a distraction, and a lot of people didn't fall for it.
The epitaph will read "and nothing of value was lost."
They can't do that. Apparently Microsoft has a business process patent on that.
And of course Walmart is also selling the ipad and iPod, two of the top three gifts in Santa's email inbox this holiday season. For the other of the top three they offer hundreds of iPhone accessories.
From January, The comment is here.
Google is selling this phone because it advances the technology and their phone partners wouldn't sell it. Expect them to sell an Android + Snapdragon slate for the same reasons.... I doubt Google even wants to sell phones - I think they just want to get the new good technologies adopted so that people can get used to Internet everywhere quicker. This serves their bottom line because when most people use the Internet they use Google services, which Google sells ads on. You can't very well sell Internet ads to be viewed by people who aren't close to a browser. [me]
It links to this interesting article where the CEO of Asus was backing away from the Android smartbook they had recently pulled in mid-computex.
"Currently, I still don't see a clear market for smartbooks," said Jerry Shen, CEO of Asustek Computer, during an investors' conference in Taipei.
So he pulls the Linux Snapdragon smartbook and shows up a few days later at an investors conference - just before the W7 launch - flanked by reps from Microsoft and Intel - probably glancing cautiously from one to the other hoping nothing bad happens to his precious W7 netbooks (little does he know...). And he gives a carefully prepared speech about how Intel and Microsoft are going to crush their enemies, see them driven before them, hear the lamentation of their women...
And now world & dog sees Microsoft as a fading power, Apple mobile platforms - and mobile platforms in general - as the next generation of user interface, and suddenly now he sees a future in it again. Intel is driving as hard as they can to be the thing that gives people what they want. Microsoft? Let's just say the KIN didn't work out and WP7 has a steeper hill to climb than it might have. What a difference a year makes.
I love my Samsung Epic Android phone, but obviously I know I would not have any such thing if both Apple and Google had not dared to bring us change, each in their own way.
That article was about Google's Nexus 1 phones. Remember that Google shopped its candybar phone to every phone vendor and they wouldn't take it, so Google made it, sold a grip of them, and ushered in all this sweet tech we enjoy today. If they had not done so when they did, we'd not have seen the first good big-screen Android platforms until after WP7 launched, if ever. And now those phones are selling 20M units a quarter in the US alone, giving 44% market share, driving every phone vendor that builds it into profitability or record profitability, giving US non-AT&T networks a phone to sell that isn't absolutely pathetic, and putting money in the pockets of a vast economy of app developers and advertising buyers (and of course, Google).
The message is pretty clear. If Google gives you a reference platform, Run With It! Refusing is not going to keep them from bringing new tech to market. They don't want the manufacturing and retail money because they want to leave that business to their partners. It's a messy customer service business with low leverage. It's not their strong point. But if their partners won't give us progress, they aren't averse to bringing it directly and reaping a few billion in hardware revenue along the way. Microsoft and Intel used to be able to prevent progress, to prevent "cannabilization" of their established markets. But now those days are done. Vendors used to be able to hold off the releases with "tomorrow, tommorow" and "any day now". Any more? No. That's not going to fly. We'll have progress now whether the established hardware vendors are ready to give it or not. There will be no stalling any more.
/this is me agreeing with you.
Good point. Maybe one day Unicode will win out, or perhaps EBCDIC will have a resurgence. 'twixt now and then it's best to write the text in ascii, perhaps with a well-documented human-readable escape table for symbols that aren't represented - perhaps even a complete Unicode escape table current to the document. Then in 2050 when somebody wants to use the data again in BufTable or DECSHIN or whatever they're using then, they can rewrite the presentation filter and leave the underlying text in its pristine ascii condition for future data archaeologists. Those future data archaeologists, in 2100, will likely be writing those presentation filters in C, in ascii, if current trends hold true. They may not speak English, but if the raw text is in ascii, they'll be able to figure it out. Unicode? How many revisions will Unicode see between now and then? Thousands? The odds are slim we'll even be aligning on 8-bit words for text by then - but for ASCII one single rotting page from the manual from the first IBM PC or any of thousands of texts since then will be all they need for their Rosetta Stone.
For all of the past that we can see, and the foreseeable future, simpler is better. It's easier to rewrite the presentation layer than to go back and validate that your filter reliably mangled the text. Remember that when the text is approved as acceptably mangled, the source is almost always tossed. 40 years later would be a bad time to discover that the site of the holy grail was encoded in some forgotten scheme thrice since remangled.
Forgive me, but I've spent a disproportionate share of my programming time in data archaeology myself. I'm currently shepherding so much data from the '80's that it's quite not possible to validate that it's translating properly in a dozen man years - and this is a hobby for me. That data isn't 100MB. That was a lot when I started, but now my phone holds over 100x as much. Enterprise data shepherds with terabytes of legacy data, you have my sympathy.
Michael decided to use this huge amount of computer time to search the public domain books that were stored in our libraries, and to digitize these books. He also decided to store the electronic texts (eTexts) in the simplest way, using the plain text format called Plain Vanilla ASCII, so they can be read easily by any machine, operating system or software.
- Marie Lebert
Since its humble beginnings in 1971 Project Gutenberg has reproduced and distributed thousands of works to millions of people in - ultimately - billions of copies. They support ePub now and simple HTML, as well as robo-read audio files, but the one format that has been stable this whole time has been ASCII. It's also the format that is likely to survive the longest without change. Project Gutenberg texts can now be read on every e-reader, smartphone, tablet and PC.
If you want to use Rich Text format, or XML, or PostScript or something else then fine - please do. But don't go trying to deprecate ASCII.
There will never be a lack of fail for you to exploit. Right now there are some good corners in deduplicated filesystems, binocular vision processing, and of course somebody's going to need to build the drag-and-drop CMS for HTML5. This is the song that never ends.
Why doesn't Microsoft come out with some sort of search engine? I'm sure they've got plenty of sharp coders that could make that work.
Obviously some people are not clear on the concept that some of us are looking to conclude a purchase when we go to Google, and the ads are actually part of the relevant results we're looking for.
Oops. I meant Orphans of the Sky, 1941 - not Tunnel in the Sky, which was a prequel to Starship Troopers.
Well if we're going to reference sci-fi the relevant issues in Footfall echo the themes given in Tunnel in the Sky given by the Dean of Science Fiction (Robert A. Heinlein) thirty years before, in 1955. Or Methusela's children in 1958. Kids these days.
I missed you old friend. I hope all is well with you.
And then you'd have the thin and light performance profile of Windows 7, combined with the rock solid security of Windows XP with IE6. Brilliant!
If it's not legal, don't do it. This is a pretty basic policy that covers pretty much the entirety of your comment, and it's a fairly standard policy in corporations, schools, state, local and federal government IT departments. I do believe that violating this basic policy is grounds for termination just about everywhere - at least everywhere I've ever worked.
If your software vendor wants to make it illegal for you to move forward in a functional fashion there is a solution for that that doesn't involve breaking the law. It also doesn't involve using the next version of their product.
There will be some of that. That's why you don't send out just one generational ship. There's also the issue of a people who have been in the ship for so long that when they arrive at the destination they have agoraphobia and no longer want to get out of the ship. All of these perils and more await.
The setting out is easy. Men have been abandoning hearth and home to see what's beyond the edge of what we know since before we were Men - and taking our families with us too. It's our privilege as parents to make that choice for our progeny.
Men will span the reaches beyond the planets and after that the stars, or we will not. If not then we will one day find our end, and we may as well have never been.
Remember one string that's unique. Download the file. Uncompress it. Grep.
Apparently in completely unrelated news, Asus is deprecating Windows Phone 7. This even though Google totally cleaned Garmin's clock on the free navi thing.
As always, Garmin-Asus seeks the best for our consumers either on Android platform or on Windows platform. However, we see the potential of Android platform devices, so we are focusing on Android platform currently. - Steven Tu
Meanwhile Microsoft's VP, Corporate Communications Frank X. Shaw is over on Twitter right now trying to repair the damage done by today's CNN Money report. In case y'all want to wander over and lend him a hand.
He's got every book, every fact, every song in the world. If I'm shopping for something, he can show me the prices nearby that it's offered at. He can bring me anything that is sold in all the world. He can tell me how to get to anywhere, and sell me tickets too. He'll even show my my place in the Universe. He knows so much about me that when I ask him the news, he only gives me the news I find interesting - and he found this out because I told him.
It's threatening. Though he's not done anything mean or spiteful yet that much power can't be good. It seems every time I ask him for something he learns even more about what I want, and uses that information to serve me better and even faster. It's creepy. Somebody rescue me please!
I could just leave him home but he's so darned useful. What if I needed to know the emergency treatment for jellyfish sting or something, and he wasn't there?
Compare and contrast Ray Ozzie's farewell with that of another recent high-level departure, J. Allard. These men, at the heart of technology for all their adult lives, were in positions of the highest influence at Microsoft. They're obviously both brilliant, and not needing to cash a paycheck. They see a change coming - a huge change - and they want to be a part of it. They don't see that happening while they work in Redmond. So they go. But on the way out they look back at the poor souls they leave behind and they tell them in their farewell: "You too can be a part of this new world. You just have to think different." The door swings shut with a click and the obvious conclusion remains unsaid: "but you won't."
The SSDs in the Apple PCs mentioned in the fine article use a mini-PCIe interface. But yes, they're replaceable and upgradeable.
The problem with secure erase is a legitimate concern. Fortunately, it will blend.
The newest enterprise versions will have deduplication in the onboard controller both to reduce the write count and increase the effective bandwidth. In two more generations this will be consumer level stuff. Internal write smoothing algorithms are already two generations old, as is extra storage to map out failed cells.
The failure thing is pretty much handled.
The very fact that we're having this discussion implies that there's a business model associated with being a one-stop third-party repository for all platforms, handling software sale (if cost), download, install, payment processing, vetting software vendors, uninstallation, recommendation, rating and reviewing, and that such a business will be profitable. Since it's a business model, at this moment there are businessmen all over the country filing papers for the corporations that are going to get in on the ground level. There will no doubt be several competing businesses offerring you this boon next month, each advertising "exclusives" and vying for your attention.
The obvious thing to do is to patent an API for the arbitrage of trust and control in competing integrated software market distributors and updaters, so that licensees could implement customer-driven subrogation and promotion of their preferred software markets and still perform their search and update duties in a fashion coordinated with each other. Then lock it in a drawer and forget about it for ten years.