I'm wondering if the gradual loss of hearing fidelity with age is one part of the reason why most people at a certain age stop listening to new music. The music they already listen to is adjusted by their minds to sound like it used to, but new music sounds bad because their ears lack the ability to hear it properly.
There are really two competing architectures here. One is to use a powerful db, know it really well, and tune every query and every schema change so load is managed well. Stack exchange runs 100 million page views a month off of a handful of sql server instances with such an approach. The entire stackoverflow.com site basically runs on one well-tended database.
The other one is that of using the db as a dumb data store, hidden behind an api. The api hides the fact that you have dozens of sharded mysql instances behind the scenes. You can care less about query performance and throw code against the wall faster, but growing comes at a bigger hardware cost. Digg manages comparable loads to stack exchange using a few dozen mysql instances this way. I believe facebook is even running on something like a thousand mysql instances.
The second strategy is very much a case of "worse is better". Yes, you need more hardware, but you only need to buy it when you have the users (and revenue) to match, whereas the expertise in the first strategy is an upfront investment. This is why you'll see many startups doing the second strategy. They can't afford to know better.
HTML5 appcache. It's not well known, but you can build web apps that "install" their code locally and keep running when the network is down or slow. IE is the only browser that doesn't support it, but IE10 will rectify that.
I just built an offline mobile web app using appcache. Works just fine. I was even able to show a progress bar to the user while downloading application updates. It's just like the abilities you get from native apps, except that updates are installed automatically (without user intervention or action).
It has to not be able to load books that are purchased from your company. So, you can either sell the books or you can sell the reader, but not both.
Or you could build it as a web app. I've made a proof of concept before, even with offline support. Amazon could ship kindle to iOS as a pure web app, without being in the iOS store.
Which is why I am hoping the kindle app gets pulled. Up to this point the "locked down" aspect of iOS hasn't been really felt by the users, but not having kindle purchases would be noticeable.
On the other hand, apple is clearly building a walled garden. Perhaps their whole intention is to make amazon leave iOS, and they're betting that users won't mind being forced to use iBooks.
Can you claim zero remote code vulnerability in linux, despite it being open source?
Having the source is meaningless when it consists of tens or hundreds of millions of lines of code. Back of the envelope calculations indicate that it would take you about 500 years to review 100 million lines of code, provided 8 hours a day are spent on it, every day. And then there's the bootstrapping issue. How can you be sure that the binary components you use to bootstrap the OS (be they executables or just a compiler) actually are secure?
In short, the only security metric that matters for operating systems is "do i trust my vendor?". Having the source doesn't buy you a single bit of security.
If you don't think microsoft can be trusted, I would have to ask why. Granted, in the 90's they had an awful track record, but if I look at the past decade, I see a business that "gets it" when it comes to security.
It's not just the app support. XP's user interface and security model is far inferior to windows 7's. For most people going back to XP when they're used to 7 is painful. Even if XP ran every single app that 7 runs, 7 would still be a worthy upgrade.
This is what frustrates me most. In my experience, the actual incidence of malware being installed without the user's knowing is close to zero these days. Since Vista, whenever I've heard of someone who got a virus it turned out that they were actually clearly warned that they were doing something very dangerous, sometimes even their antivirus software protested that it was a virus, and still they click continue. Why? Free movies online! Just download this video plugin first!
Users cannot be relied upon to make security decisions. The only way to make a secure OS is to remove the human factor, to take away all decision power for installing malware, which means you have to prevent them from installing anything that's not from a curated app store. In short, although it pains me to admit it, apple's model is the only one that can be truly secure.
Or (5) like in any large organization there's no mastermind that controls all the NSA's actions, and this is a case of the left hand not knowing what the right hand is doing. Someone could be honestly arguing in favor of better security for end users, while another part of the organization is working to undermine that. The question is: which of the two sides sent out this advice?
I find my efficiency increases by sequentializing the tasks into a stream. Research I've read seems to back this up. We're not meant to do a lot of context switching (the brain sucks at context switching). That's also why it's better to group tasks together by subject matter, even if that means you'll do a low priority task before a high priority one.
Wouldn't it be better to use the emulation route? For example, writing a program for the original gameboy, and running it through the emulator. I remember at university we learned assembly on an emulated MIPS. We could focus on the individual instructions, on hardware that was simple and clean, but it all ran on the unix servers (x terminals).
Governments can't hold on to infrastructure that can be exploited commercially. Whether it's buildings or cable networks, eventually it gets sold off to balance that year's budget. The belgian government went on a decade-long selling spree to balance a structurally unbalanced budget, and the consequence is that now there are gigantic budget issues and the government needs to make the deepest cuts in the history of the country.
I'm glad they didn't implement those form elements, because once they implement a part of a standard, their implementation becomes the rule. If they implemented HTML5 form elements now, that essentially means marking the current HTML5 draft as finalized. I don't think that would be good for HTML5.
Only firefox renders that page correctly. All the other browsers fail in some way. Chrome doesn't even support color profiles at all.
IE9 's platform preview supposedly has full support for gamma, and v2 and v4 color profiles, but it has issues rendering that page. I'll report that page in their bugtracking system, so IE9 has fully accurate color rendering by release time.
And all of that for what? Pretty graphics (Aero) that will be turned off right after Autorun?
If they're pretty, why do you turn them off? They run on the graphics hardware, so they don't affect the performance of your apps.
Why is it that a linux system isn't configured well if it doesn't run compiz, but at the same time aero on windows is a bloated monstrosity that needs disabling?
Anyway, the main reason to run windows 7 is responsiveness. XP regularly blocks if you're multi-tasking, because it has lousy CPU scheduling and no IO scheduling. Windows 7 remains responsive in almost all situations (at least in my personal experience).
This isn't limited to IE though. Mac Office has vast incompatibilities with the windows version (most notably a lack of support for macro's, which they'll rectify in the next release).
It's because the mac team at microsoft is completely separate from everyone else.
So any web site which uses Javascript is open to compromise and therefore could become a mal-Javascript distributor.
XSS leverages javascript, sure, but there are many ways of breaking a site's security without using javascript. Javascript itself is not a security problem, since it runs in a sandbox. Security problems, even those manifested through javascript, are always caused by bad design on the back-end (not filtering user input correctly).
The overemphasis on how fast Javascript runs seems to be due to a lack of serious thought as to how to make browsers better at doing what they were designed to do -- which was *not* to run "web-apps". We used the Internet very successfully for over a decade to provide information -- not to run apps -- if it wasn't (isn't) broken why the emphasis on fixing(?) it?
Just like the purpose of land-line communication has changed from voice calls to data, so is the purpose of browsers changing from document viewing to applications. The browser as a rich app platform is a good thing. It takes a lot of worry away from end users (upgrading, security, installation,...). In the long run, we're all going to have better and easier to use apps because of it. We're finally going to be able to get rid of our personal computer as a physical piece of hardware. A PC should be a metaphorical construct that follows you around as needed, regardless of the hardware involved. The web is the only credible way of doing that.
I note this with an aside that the U.S. Government (NIH NCBI) no longer allows complete access to its *public* databases, e.g. PubMed, by browsers which do not have Javascript enabled. (One is compelled to ask *who* for the most part paid for that information but can no longer access it?).
Javascript is as essential to a modern browser as HTML and CSS. Disabling javascript has no point anymore (ever since accessibility products learned to cope with ajax). If you're talking about using tools like curl to extract content, then I agree that ways have to be provided to easily obtain all the content from a site. That doesn't mean that these sites should cater to the lowest common denominator and give everyone a shitty experience to allow a corner case. It just means they need to implement the corner case as a separate solution. Ever used a web app without any javascript? It's always a lousy experience. I don't see why that should be foisted on all users.
The memory use was either misinterpretation of statistics (happens often), or plugins. Firefox itself doesn't have a memory problem, but its plugins and add-ons do have major memory and stability issues. A clean firefox install with a clean profile and no addons or plugins is almost invulnerable. You might argue that it's also near useless, but the reality is that the other browsers are just as vulnerable to plugin and addon issues.
Now that flash is running in a separate process (as of FF 3.6.4) we can finally see that it's the flash plugin that's taking up the majority of RAM (at least, that's what I see on my system).
Slashdot Mirror
I'm wondering if the gradual loss of hearing fidelity with age is one part of the reason why most people at a certain age stop listening to new music. The music they already listen to is adjusted by their minds to sound like it used to, but new music sounds bad because their ears lack the ability to hear it properly.
There are really two competing architectures here. One is to use a powerful db, know it really well, and tune every query and every schema change so load is managed well. Stack exchange runs 100 million page views a month off of a handful of sql server instances with such an approach. The entire stackoverflow.com site basically runs on one well-tended database.
The other one is that of using the db as a dumb data store, hidden behind an api. The api hides the fact that you have dozens of sharded mysql instances behind the scenes. You can care less about query performance and throw code against the wall faster, but growing comes at a bigger hardware cost. Digg manages comparable loads to stack exchange using a few dozen mysql instances this way. I believe facebook is even running on something like a thousand mysql instances.
The second strategy is very much a case of "worse is better". Yes, you need more hardware, but you only need to buy it when you have the users (and revenue) to match, whereas the expertise in the first strategy is an upfront investment. This is why you'll see many startups doing the second strategy. They can't afford to know better.
HTML5 appcache. It's not well known, but you can build web apps that "install" their code locally and keep running when the network is down or slow. IE is the only browser that doesn't support it, but IE10 will rectify that.
I just built an offline mobile web app using appcache. Works just fine. I was even able to show a progress bar to the user while downloading application updates. It's just like the abilities you get from native apps, except that updates are installed automatically (without user intervention or action).
It has to not be able to load books that are purchased from your company. So, you can either sell the books or you can sell the reader, but not both.
Or you could build it as a web app. I've made a proof of concept before, even with offline support. Amazon could ship kindle to iOS as a pure web app, without being in the iOS store.
Which is why I am hoping the kindle app gets pulled. Up to this point the "locked down" aspect of iOS hasn't been really felt by the users, but not having kindle purchases would be noticeable.
On the other hand, apple is clearly building a walled garden. Perhaps their whole intention is to make amazon leave iOS, and they're betting that users won't mind being forced to use iBooks.
Can you claim zero remote code vulnerability in linux, despite it being open source?
Having the source is meaningless when it consists of tens or hundreds of millions of lines of code. Back of the envelope calculations indicate that it would take you about 500 years to review 100 million lines of code, provided 8 hours a day are spent on it, every day. And then there's the bootstrapping issue. How can you be sure that the binary components you use to bootstrap the OS (be they executables or just a compiler) actually are secure?
In short, the only security metric that matters for operating systems is "do i trust my vendor?". Having the source doesn't buy you a single bit of security.
If you don't think microsoft can be trusted, I would have to ask why. Granted, in the 90's they had an awful track record, but if I look at the past decade, I see a business that "gets it" when it comes to security.
It's not just the app support. XP's user interface and security model is far inferior to windows 7's. For most people going back to XP when they're used to 7 is painful. Even if XP ran every single app that 7 runs, 7 would still be a worthy upgrade.
This is what frustrates me most. In my experience, the actual incidence of malware being installed without the user's knowing is close to zero these days. Since Vista, whenever I've heard of someone who got a virus it turned out that they were actually clearly warned that they were doing something very dangerous, sometimes even their antivirus software protested that it was a virus, and still they click continue. Why? Free movies online! Just download this video plugin first!
Users cannot be relied upon to make security decisions. The only way to make a secure OS is to remove the human factor, to take away all decision power for installing malware, which means you have to prevent them from installing anything that's not from a curated app store. In short, although it pains me to admit it, apple's model is the only one that can be truly secure.
Or (5) like in any large organization there's no mastermind that controls all the NSA's actions, and this is a case of the left hand not knowing what the right hand is doing. Someone could be honestly arguing in favor of better security for end users, while another part of the organization is working to undermine that. The question is: which of the two sides sent out this advice?
I propose "Comet Office". Maybe the renaming will incentivize them to finally do something about its performance.
Word is no speed demon, but OOo Writer makes computers sprout physical gears just to be able to grind them.
Maybe they release a scandal now and then to distract us from the real cover-up?
I find my efficiency increases by sequentializing the tasks into a stream. Research I've read seems to back this up. We're not meant to do a lot of context switching (the brain sucks at context switching). That's also why it's better to group tasks together by subject matter, even if that means you'll do a low priority task before a high priority one.
That's why god invented the rest room. If anyone complains, tell them you have irritable bowel syndrome.
It's what I do when I get stuck on a programming problem. Works every time.
Wouldn't it be better to use the emulation route? For example, writing a program for the original gameboy, and running it through the emulator. I remember at university we learned assembly on an emulated MIPS. We could focus on the individual instructions, on hardware that was simple and clean, but it all ran on the unix servers (x terminals).
Governments can't hold on to infrastructure that can be exploited commercially. Whether it's buildings or cable networks, eventually it gets sold off to balance that year's budget. The belgian government went on a decade-long selling spree to balance a structurally unbalanced budget, and the consequence is that now there are gigantic budget issues and the government needs to make the deepest cuts in the history of the country.
In other words, just another typical government.
I'm glad they didn't implement those form elements, because once they implement a part of a standard, their implementation becomes the rule. If they implemented HTML5 form elements now, that essentially means marking the current HTML5 draft as finalized. I don't think that would be good for HTML5.
PNG is also about 14 or 15 years old, but IE still cannot handle its color correction chunks (gAMA, iCCP) properly:
http://www.libpng.org/pub/png/png-gammatest.html
Only firefox renders that page correctly. All the other browsers fail in some way. Chrome doesn't even support color profiles at all.
IE9 's platform preview supposedly has full support for gamma, and v2 and v4 color profiles, but it has issues rendering that page. I'll report that page in their bugtracking system, so IE9 has fully accurate color rendering by release time.
Your problem is not that it is bad. Netscape 4 was just as bad. Your problem is that it is still widely used.
No, netscape 4 was much, much worse. IE6 was to Netscape 4 what Chrome is to IE6.
And all of that for what? Pretty graphics (Aero) that will be turned off right after Autorun?
If they're pretty, why do you turn them off? They run on the graphics hardware, so they don't affect the performance of your apps.
Why is it that a linux system isn't configured well if it doesn't run compiz, but at the same time aero on windows is a bloated monstrosity that needs disabling?
Anyway, the main reason to run windows 7 is responsiveness. XP regularly blocks if you're multi-tasking, because it has lousy CPU scheduling and no IO scheduling. Windows 7 remains responsive in almost all situations (at least in my personal experience).
IE8 does not have an IE6 compatibility mode. It can only emulate IE7.
This isn't limited to IE though. Mac Office has vast incompatibilities with the windows version (most notably a lack of support for macro's, which they'll rectify in the next release).
It's because the mac team at microsoft is completely separate from everyone else.
So any web site which uses Javascript is open to compromise and therefore could become a mal-Javascript distributor.
XSS leverages javascript, sure, but there are many ways of breaking a site's security without using javascript. Javascript itself is not a security problem, since it runs in a sandbox. Security problems, even those manifested through javascript, are always caused by bad design on the back-end (not filtering user input correctly).
The overemphasis on how fast Javascript runs seems to be due to a lack of serious thought as to how to make browsers better at doing what they were designed to do -- which was *not* to run "web-apps". We used the Internet very successfully for over a decade to provide information -- not to run apps -- if it wasn't (isn't) broken why the emphasis on fixing(?) it?
Just like the purpose of land-line communication has changed from voice calls to data, so is the purpose of browsers changing from document viewing to applications. The browser as a rich app platform is a good thing. It takes a lot of worry away from end users (upgrading, security, installation, ...). In the long run, we're all going to have better and easier to use apps because of it. We're finally going to be able to get rid of our personal computer as a physical piece of hardware. A PC should be a metaphorical construct that follows you around as needed, regardless of the hardware involved. The web is the only credible way of doing that.
I note this with an aside that the U.S. Government (NIH NCBI) no longer allows complete access to its *public* databases, e.g. PubMed, by browsers which do not have Javascript enabled. (One is compelled to ask *who* for the most part paid for that information but can no longer access it?).
Javascript is as essential to a modern browser as HTML and CSS. Disabling javascript has no point anymore (ever since accessibility products learned to cope with ajax). If you're talking about using tools like curl to extract content, then I agree that ways have to be provided to easily obtain all the content from a site. That doesn't mean that these sites should cater to the lowest common denominator and give everyone a shitty experience to allow a corner case. It just means they need to implement the corner case as a separate solution. Ever used a web app without any javascript? It's always a lousy experience. I don't see why that should be foisted on all users.
The memory use was either misinterpretation of statistics (happens often), or plugins. Firefox itself doesn't have a memory problem, but its plugins and add-ons do have major memory and stability issues. A clean firefox install with a clean profile and no addons or plugins is almost invulnerable. You might argue that it's also near useless, but the reality is that the other browsers are just as vulnerable to plugin and addon issues.
Now that flash is running in a separate process (as of FF 3.6.4) we can finally see that it's the flash plugin that's taking up the majority of RAM (at least, that's what I see on my system).
Wow, it's almost as if when they ported it to Windows, they *actually* ported it to Windows instead of completely half-assing it!
Chrome was developed on windows first, and then ported to mac and linux.
Stopwatch measurements of page loads? Back in the day, that's how we did it (uphill in the snow).