MS wants to take advantage of UEFI, which has obvious benefits. Chromebooks work the same way, but we don't read any heated/. articles about it because Google is charmed and MS is "evil".
It is up to the device manufacturers to figure out a way to let the end-user ultimately take control of their own PCs. They could do that Chromebooks style -- a hardware switch -- or by distributing the key in a secure manner, such as mailing it to the owner's registered home address. Consumers who care about this issue should look for this feature in whatever device they purchase. What's all the fuss?
Re:Warning: Excessive buzzwords can be fatal
on
Windows 8 Roundup
·
· Score: 1
I feel your pain (especially on Silverlight.) The big question to me now is whether it makes sense to target any OS explicitly any more. It's hard to argue that HTML5 can't handle the user experience aspects; it's more than capable of handling the job on just about any device you care to name. The problem is with device access (camera, sensors, etc.). To get that, I have to use each OS vendor's toolchain, which is expensive and everybody hates.
Win8 seems to be at least a step in the right direction. I get to use all of my HTML5 UI code, and there's a JS way of reaching down to the devices if the user is running Win8. I can reuse much the UI on the other platforms, but I'll still have to maintain different ways of packaging them.
This feature was mentioned in one of the keynotes. Under Win8, explicit saves will "not be required", but no details about how that would work were supplied. (Of course, as saving is the app's job,/any/ app can already implement this feature today; I presume you were referring to some system-wide support for incremental saves.)
I haven't seen anything non-standard yet in the demos. They consolidated all of their WinRT support (WinRT is the new Windows kernel API) to a distinct set of JS libraries, and mapped them all to a single top-level JS object called "Windows". You see all the normal things one would expect in a JS object: functions, objects, arrays, and properties. Looked pretty clean and well thought-out to me.
A good clarification. "WinRT" as a whole refers to a new API to the Windows kernel. What I was calling the "JS bridge" is the trimmed-down version of that API that gets exposed as JS objects and functions; perhaps "JS binding" might be a more accurate term. Also, several demos highlighted the fact that complied C++ and C# programs can be packaged with JS and referenced at runtime.
Re:Warning: Excessive buzzwords can be fatal
on
Windows 8 Roundup
·
· Score: 1
Actually there are some important concepts in those terms; they're not just buzz words for show. A "continuous service" is a resource that is fault tolerant and scales horizontally. (And if you think fault tolerance and horizontal scaling are themselves buzz words, then I herewith banish you from/. until you read at least the Wikipedia articles that cover these concepts.) At the keynote, we saw a demo of a SQL query running against a clustered file share of the data. The presenter killed one the NICs in the cluster, and Windows handed off the query to the other share in the cluster with only a momentary interruption. The caller was not dropped, just delayed a short period. That's "continuous service".
Re:This is cool ... rush to judgment?
on
Windows 8 Roundup
·
· Score: 1
You can have any number of apps open; a simple swipe from top edge reveals them in a floating bar -- touch the one you want. Honestly, aren't you over-reacting? This is not fundamentally different from how Windows launches apps today, except that the UI now bleeds to the full edge of the screen -- no chrome -- so you need one finger-swipe to show the tiles at the start. Big deal. It's very touch-friendly, and the edge-to-edge UI makes everything look like a magazine, beautiful and functional.
Win8 will be competitive
on
Windows 8 Roundup
·
· Score: 3, Interesting
I'm at the BUILD conference, and my impression is that Windows 8 will be very competitive. The re-imagining effort is sweeping, and touches everything from the back end to Consumer devices. The big news is that HTML 5 is now a "native" programming platform for the client UI. There are two JS libraries, a pure JS library that implements the new Metro look-and-feel (WinJS), and a Windows/JS bridge library that exposes the Windows API (and hence the Windows-controlled hardware such as the camera) in Javascript (WinRT). Tooling improvements include terrific new debugging scenarios and a major upgrade to Expression Blend to be able to edit HTML/CSS as well as XAML.
Basically, MS has taken the best ideas of the web development world, and leveraged them to massively improve the development experience for their next OS. If I wanted to write Windows-specific apps, Win8 is a huge improvement. It's an open question, however, whether people want to write Windows-specific apps as opposed to web-centric apps. Even then, Win8 will definitely shorten and simplify the transition from a web app to a Windows app.
Not a lot of details yet, only one short (but tantalizing) talk at the general keynote today. Active Directory will now be paired up with some kind of security rules engine that can inspect claims (user attributes) and the contents of files and implement enterprise-wide access policies based on the values it finds. The example shown was someone putting a file with sensitive data in a public share. When an unauthorized user inspected that share, the sensitive file wasn't even shown.
MS has put a lot of effort into developer productivity in Windows 8, especially when crossing boundaries between the client, server, and cloud. Visual Studio is now buttressed with a new version of Expression Blend that edits both HTML5 and XAML. The editor has a live preview that lets you navigate a site with an inboard browser, then dive into the code on a single rendered element, even if it is dynamically generated. Likewise, there are new remote debugging tools that lets the developer go from a running web app to the output of the server to the code that generates the output, all in one view. There are single-click deployment tools for publishing to server, app store, and cloud.
This all goes/way/ beyond what Eclipse can do. Very impressive; tooling has always been a strength for MS.
ME: I'm thinking of a number. CB: Between 1 and 20. ME: Would you like to know what it is? CB: Yes, I would. ME: Guess. CB: I live in a bubble from which I will never escape.
Maybe they've decided that no US company can win a market where manufacturing is outsourced, which means unless they win the software race, they're out of business eventually anyway.
It might work even so. The predictions are done in real time, using sensing data gathered by an iphone camera mounted in the car. As long as the triggering wasn't hugely irregular, the predictions would probably still be accurate enough to be useful. The paper did mention that they tried it with both fixed-timing and predictive-timing systems, and it worked equally well on both.
The way to think about this is that G+ is part of a very long list of things that people can do together using Google tools and services. It's the part of the system that identifies a person as a person ('identity service'), really the linchpin of the whole system of person-to-person networking. The "social" uses of this are but one application of the identity service.
As for what to do next, I recommend you simply look for something in everyday life that bothers you, doesn't have a simple/good solution yet, and give 'er a shot. Worked pretty good last time...
Thanks for what has been/by far/ the most interesting site on the Internet.
You might be looking at this too narrowly. HTML is just one client platform among many. You use it when it makes sense; in particular, when you don't want the bother of installing a native app in the first place. Like all other client platforms, HTML has limitations, and only fools with lots of time and money to throw away fail to respect them. In its sweet spot, HTML offers some truly unparalleled features that no other stack can touch (such as hypertext linking, for example.)
Pretty silly to complain about HTML 5 "standards" when the real problem isn't with the standards, it's with the implementation of those standards. That's why we have tests such as Acid, of course.
I will say, however, that the implementations of the browser standards for HTML 5 and CSS3 are SO much better than earlier rounds of the browser wars. At least it's not a complete nightmare as before. Where you find problems are in edge-cases such as websocket and threads for which there is really no workaround possible, you just have to wait for the browser vendor.
We all know the long-term cost of a laptop is driven by the software and its upkeep. On one platform, you've got tens of thousands of apps, all free, and -- no small thing -- all centrally distributed, installed and automatically patched. On the other you've got a mere handful of useful free apps out of the box. Everything else you have to go hunting for on websites you pray to God aren't infested, install yourself, and patch yourself.
Really, there's no comparison when looked at from a my-time-is-valuable-and-no-I-can't-help-you-speed-up-your-windows-box point of view.
arguments to uniqueness normally have a religious foundation, because there is no evidence whatsoever that they are correct
What? This is exactly backwards. There is NO evidence yet that life exists outside of Earth. Paradoxically, however, we/believe/ that life must exist elsewhere because here we are. This is such a common paradox, it even has a name, the Fermi Paradox.
In a service like Google+, where the user controls the network participants, then anonymity doesn't really make sense, right? I mean, who would accept an invite from someone they didn't know? Ditto for FB.
In another context such as a comment on Slashdot, where somebody else decides what gets seen by whom, then anonymity would be handy for all the usual reasons people give.
- Do you trust your bank to keep your records? It's a black box, you have no idea what they do. - Do you trust your email app to save your emails. That's a black box, too. - Do you trust your best friend to keep your secrets?
You can't smell, touch, feel, see your data or your secrets. However, you do trust them because they've been reliable in the past.
The same is true of voting machines. They run sample votes before they put the machines into service. The sample votes are double-blind. (In my city, a machine has NEVER failed one of these test votes, by the way.) That's how you establish trust.
The people who handle the machines put them under lock and key while not in use, and apply physical seals to them when they are moved. The people who open the machines verify that the seals are unbroken, and record the serial numbers, which are cross-checked at the end of the day. That's the kind of process that also establishes trust.
At the end of the day, you're ultimate trust is in the people who run the system, not in the machines.
NO, paper is a horrible medium for tallying a vote. Paper can be lost, destroyed, ripped, etc. and the error rate can run as high as 1-2%. Electronic voting by contrast is highly reliable, and removes the complexity and mess of a physical ballot. There are lots of cross-checks and controls built into the voting process that make it extremely unlikely that a voting machine could be compromised. Please volunteer to be a poll worker, so you can see what I'm talking about. I think you'll be pleasantly surprised.
Many people are so focused on absolute accuracy for a recount, but recounts are actually very rare events - something like only 0.5% of the races ever need a recount. It is MUCH wiser to optimize for the 99.5% of the races that don't involve recounts. Forcing 99.5% of the races to use a paper-based system with an error rate of 1-2% would be taking a huge step backwards in the trustworthiness of the voting process.
I wish more technical people would volunteer to work the polls, and could spread the word about the controls built into our voting process.
The first thing they'd learn is that votes are counted at the PRECINCT level. There's no "master server" in the sky where votes can be manipulated. The real votes are counted machine-by-machine, under the eye of volunteers who swear under oath that it has been operated properly. The machines print out a paper receipt of the tally, and that gets backed-up on hard disk and flash. The paper tape total is called into the Registrar. The paper records of the vote are certified by a local Board of Election, the machines are sealed, and the paper and flash media is typically also sealed and sequestered under a local Court.
The servers used at the state levels are merely there to REPORT the results of the counts made at each precinct. They are not the actual vote tally. If the database is wrong, the Board goes back to the paper trail and updates it with the correct tally.
Paper receipts at the voting machines are actually NOT a good idea, IMHO. Paper is a horrible medium for conducting an election: it can get lost, smeared, ripped, crumpled, folded, etc. There's a reason we don't run our accounting systems using ledger-books anymore, but instead use a computer. Those reasons apply double for voting. A computer-based tally is a dream to manage compared to the nightmare that is paper.
I would like to see better use of paper for making spot-certifications that a machine is operating properly, but I would never want to run a whole election using paper. The error rate of paper can run as high as 1-2%. The error rate of a computer tally is minuscule by comparison.
Reading these hyper-ventilating comments on Slashdot is so depressing. Can't we just let facts speak for themselves? Do we have to politicize everything into terms of "victory" and "defeat"? If the Science is correct, it doesn't need us to defend it. Truth carries within itself its own proof.
May I suggest that some simple humility and a more open-minded attitude is more in order? Human beings have been around -- if the Science is correct --- for a mere blink of an eye in geologic time. We must know almost NOTHING yet. We don't even know what what 95% of matter in the universe is, we just call it "dark" matter, as if that explains anything.
I'm not a scientist, but I have a brain. The universe is bound to be a whole lot bigger than our piddly little theory of evolution can yet explain.
Slashdot Mirror
MS wants to take advantage of UEFI, which has obvious benefits. Chromebooks work the same way, but we don't read any heated /. articles about it because Google is charmed and MS is "evil".
It is up to the device manufacturers to figure out a way to let the end-user ultimately take control of their own PCs. They could do that Chromebooks style -- a hardware switch -- or by distributing the key in a secure manner, such as mailing it to the owner's registered home address. Consumers who care about this issue should look for this feature in whatever device they purchase. What's all the fuss?
I feel your pain (especially on Silverlight.) The big question to me now is whether it makes sense to target any OS explicitly any more. It's hard to argue that HTML5 can't handle the user experience aspects; it's more than capable of handling the job on just about any device you care to name. The problem is with device access (camera, sensors, etc.). To get that, I have to use each OS vendor's toolchain, which is expensive and everybody hates.
Win8 seems to be at least a step in the right direction. I get to use all of my HTML5 UI code, and there's a JS way of reaching down to the devices if the user is running Win8. I can reuse much the UI on the other platforms, but I'll still have to maintain different ways of packaging them.
This feature was mentioned in one of the keynotes. Under Win8, explicit saves will "not be required", but no details about how that would work were supplied. (Of course, as saving is the app's job, /any/ app can already implement this feature today; I presume you were referring to some system-wide support for incremental saves.)
I haven't seen anything non-standard yet in the demos. They consolidated all of their WinRT support (WinRT is the new Windows kernel API) to a distinct set of JS libraries, and mapped them all to a single top-level JS object called "Windows". You see all the normal things one would expect in a JS object: functions, objects, arrays, and properties. Looked pretty clean and well thought-out to me.
A good clarification. "WinRT" as a whole refers to a new API to the Windows kernel. What I was calling the "JS bridge" is the trimmed-down version of that API that gets exposed as JS objects and functions; perhaps "JS binding" might be a more accurate term. Also, several demos highlighted the fact that complied C++ and C# programs can be packaged with JS and referenced at runtime.
Actually there are some important concepts in those terms; they're not just buzz words for show. A "continuous service" is a resource that is fault tolerant and scales horizontally. (And if you think fault tolerance and horizontal scaling are themselves buzz words, then I herewith banish you from /. until you read at least the Wikipedia articles that cover these concepts.) At the keynote, we saw a demo of a SQL query running against a clustered file share of the data. The presenter killed one the NICs in the cluster, and Windows handed off the query to the other share in the cluster with only a momentary interruption. The caller was not dropped, just delayed a short period. That's "continuous service".
You can have any number of apps open; a simple swipe from top edge reveals them in a floating bar -- touch the one you want. Honestly, aren't you over-reacting? This is not fundamentally different from how Windows launches apps today, except that the UI now bleeds to the full edge of the screen -- no chrome -- so you need one finger-swipe to show the tiles at the start. Big deal. It's very touch-friendly, and the edge-to-edge UI makes everything look like a magazine, beautiful and functional.
I'm at the BUILD conference, and my impression is that Windows 8 will be very competitive. The re-imagining effort is sweeping, and touches everything from the back end to Consumer devices. The big news is that HTML 5 is now a "native" programming platform for the client UI. There are two JS libraries, a pure JS library that implements the new Metro look-and-feel (WinJS), and a Windows/JS bridge library that exposes the Windows API (and hence the Windows-controlled hardware such as the camera) in Javascript (WinRT). Tooling improvements include terrific new debugging scenarios and a major upgrade to Expression Blend to be able to edit HTML/CSS as well as XAML.
Basically, MS has taken the best ideas of the web development world, and leveraged them to massively improve the development experience for their next OS. If I wanted to write Windows-specific apps, Win8 is a huge improvement. It's an open question, however, whether people want to write Windows-specific apps as opposed to web-centric apps. Even then, Win8 will definitely shorten and simplify the transition from a web app to a Windows app.
Not a lot of details yet, only one short (but tantalizing) talk at the general keynote today. Active Directory will now be paired up with some kind of security rules engine that can inspect claims (user attributes) and the contents of files and implement enterprise-wide access policies based on the values it finds. The example shown was someone putting a file with sensitive data in a public share. When an unauthorized user inspected that share, the sensitive file wasn't even shown.
MS has put a lot of effort into developer productivity in Windows 8, especially when crossing boundaries between the client, server, and cloud. Visual Studio is now buttressed with a new version of Expression Blend that edits both HTML5 and XAML. The editor has a live preview that lets you navigate a site with an inboard browser, then dive into the code on a single rendered element, even if it is dynamically generated. Likewise, there are new remote debugging tools that lets the developer go from a running web app to the output of the server to the code that generates the output, all in one view. There are single-click deployment tools for publishing to server, app store, and cloud.
This all goes /way/ beyond what Eclipse can do. Very impressive; tooling has always been a strength for MS.
ME: I'm thinking of a number.
CB: Between 1 and 20.
ME: Would you like to know what it is?
CB: Yes, I would.
ME: Guess.
CB: I live in a bubble from which I will never escape.
Maybe they've decided that no US company can win a market where manufacturing is outsourced, which means unless they win the software race, they're out of business eventually anyway.
It might work even so. The predictions are done in real time, using sensing data gathered by an iphone camera mounted in the car. As long as the triggering wasn't hugely irregular, the predictions would probably still be accurate enough to be useful. The paper did mention that they tried it with both fixed-timing and predictive-timing systems, and it worked equally well on both.
The way to think about this is that G+ is part of a very long list of things that people can do together using Google tools and services. It's the part of the system that identifies a person as a person ('identity service'), really the linchpin of the whole system of person-to-person networking. The "social" uses of this are but one application of the identity service.
As for what to do next, I recommend you simply look for something in everyday life that bothers you, doesn't have a simple/good solution yet, and give 'er a shot. Worked pretty good last time ...
Thanks for what has been /by far/ the most interesting site on the Internet.
Some very good analysis and thinking. I second the post about Google's Native-Client technology. The API is here.
You might be looking at this too narrowly. HTML is just one client platform among many. You use it when it makes sense; in particular, when you don't want the bother of installing a native app in the first place. Like all other client platforms, HTML has limitations, and only fools with lots of time and money to throw away fail to respect them. In its sweet spot, HTML offers some truly unparalleled features that no other stack can touch (such as hypertext linking, for example.)
Pretty silly to complain about HTML 5 "standards" when the real problem isn't with the standards, it's with the implementation of those standards. That's why we have tests such as Acid, of course.
I will say, however, that the implementations of the browser standards for HTML 5 and CSS3 are SO much better than earlier rounds of the browser wars. At least it's not a complete nightmare as before. Where you find problems are in edge-cases such as websocket and threads for which there is really no workaround possible, you just have to wait for the browser vendor.
We all know the long-term cost of a laptop is driven by the software and its upkeep. On one platform, you've got tens of thousands of apps, all free, and -- no small thing -- all centrally distributed, installed and automatically patched. On the other you've got a mere handful of useful free apps out of the box. Everything else you have to go hunting for on websites you pray to God aren't infested, install yourself, and patch yourself.
Really, there's no comparison when looked at from a my-time-is-valuable-and-no-I-can't-help-you-speed-up-your-windows-box point of view.
What? This is exactly backwards. There is NO evidence yet that life exists outside of Earth. Paradoxically, however, we /believe/ that life must exist elsewhere because here we are. This is such a common paradox, it even has a name, the Fermi Paradox.
In a service like Google+, where the user controls the network participants, then anonymity doesn't really make sense, right? I mean, who would accept an invite from someone they didn't know? Ditto for FB.
In another context such as a comment on Slashdot, where somebody else decides what gets seen by whom, then anonymity would be handy for all the usual reasons people give.
OK, I'll bite:
- Do you trust your bank to keep your records? It's a black box, you have no idea what they do.
- Do you trust your email app to save your emails. That's a black box, too.
- Do you trust your best friend to keep your secrets?
You can't smell, touch, feel, see your data or your secrets. However, you do trust them because they've been reliable in the past.
The same is true of voting machines. They run sample votes before they put the machines into service. The sample votes are double-blind. (In my city, a machine has NEVER failed one of these test votes, by the way.) That's how you establish trust.
The people who handle the machines put them under lock and key while not in use, and apply physical seals to them when they are moved. The people who open the machines verify that the seals are unbroken, and record the serial numbers, which are cross-checked at the end of the day. That's the kind of process that also establishes trust.
At the end of the day, you're ultimate trust is in the people who run the system, not in the machines.
NO, paper is a horrible medium for tallying a vote. Paper can be lost, destroyed, ripped, etc. and the error rate can run as high as 1-2%. Electronic voting by contrast is highly reliable, and removes the complexity and mess of a physical ballot. There are lots of cross-checks and controls built into the voting process that make it extremely unlikely that a voting machine could be compromised. Please volunteer to be a poll worker, so you can see what I'm talking about. I think you'll be pleasantly surprised.
Many people are so focused on absolute accuracy for a recount, but recounts are actually very rare events - something like only 0.5% of the races ever need a recount. It is MUCH wiser to optimize for the 99.5% of the races that don't involve recounts. Forcing 99.5% of the races to use a paper-based system with an error rate of 1-2% would be taking a huge step backwards in the trustworthiness of the voting process.
I wish more technical people would volunteer to work the polls, and could spread the word about the controls built into our voting process.
The first thing they'd learn is that votes are counted at the PRECINCT level. There's no "master server" in the sky where votes can be manipulated. The real votes are counted machine-by-machine, under the eye of volunteers who swear under oath that it has been operated properly. The machines print out a paper receipt of the tally, and that gets backed-up on hard disk and flash. The paper tape total is called into the Registrar. The paper records of the vote are certified by a local Board of Election, the machines are sealed, and the paper and flash media is typically also sealed and sequestered under a local Court.
The servers used at the state levels are merely there to REPORT the results of the counts made at each precinct. They are not the actual vote tally. If the database is wrong, the Board goes back to the paper trail and updates it with the correct tally.
Paper receipts at the voting machines are actually NOT a good idea, IMHO. Paper is a horrible medium for conducting an election: it can get lost, smeared, ripped, crumpled, folded, etc. There's a reason we don't run our accounting systems using ledger-books anymore, but instead use a computer. Those reasons apply double for voting. A computer-based tally is a dream to manage compared to the nightmare that is paper.
I would like to see better use of paper for making spot-certifications that a machine is operating properly, but I would never want to run a whole election using paper. The error rate of paper can run as high as 1-2%. The error rate of a computer tally is minuscule by comparison.
Reading these hyper-ventilating comments on Slashdot is so depressing. Can't we just let facts speak for themselves? Do we have to politicize everything into terms of "victory" and "defeat"? If the Science is correct, it doesn't need us to defend it. Truth carries within itself its own proof.
May I suggest that some simple humility and a more open-minded attitude is more in order? Human beings have been around -- if the Science is correct --- for a mere blink of an eye in geologic time. We must know almost NOTHING yet. We don't even know what what 95% of matter in the universe is, we just call it "dark" matter, as if that explains anything.
I'm not a scientist, but I have a brain. The universe is bound to be a whole lot bigger than our piddly little theory of evolution can yet explain.