Sounds like a classic Unix architecture to me! Before threads, there were processes and signals, and it still works pretty well... I wonder how it will affect memory usage though.
The comic addresses this.
There is an upfront memory overhead in starting a new process. However, there are long term memory usage gains. The comic suggests that fragmentation and memory leakage occurs in the browser's memory space.
Sure, you could work to improve the browser's memory management - but by moving to process-per-tab, you make it a non-problem. When you close a tab, the OS reclaims that process's memory. You'd hope that the OS has rock solid memory management code.
This is all common sense, except that sometimes there's a high cost to finding out whether you want some framework.
Spring is a perfect example. Even to understand what Spring is for, you're going to need a couple of days of fairly solid reading. "Inversion of Control", "Dependency Injection", "Aspect Oriented Programming" - these are all big new concepts to most people.
Put yourself in the position of someone who doesn't know Spring. How much study are they going to need to do before deciding "yes, I can apply this stuff to my project"? Time that, if the answer turns out to be "Oh no, that framework's not for my kind of work", would be wasted.
So, one needs guidance. If someone you trust were to say "Hey, that feature of Spring would really help you with that part of your project", that would be a great help.
Not picking on Spring, by the way. The same applies to most useful frameworks. If they didn't involve big new concepts, they probably wouldn't be all that useful anywhere.
Will IBM? I would posit that they will, and that as people become less stuck on "MY PROJECT" and think a little more globally ("MY TEAM'S PROJECTS"), people will become more efficient and thus more happy. For every person who is "torn away" from their project, there will be others who aren't stuck working 80 hour weeks because some awful manager didn't predict the work load accurately.
I'd be more concerned about being "torn away" from a team, than from a project. Of course, the models *could* factor in the fact that it takes time to build a working relationship, and build that cost into its heuristics.
All true. But parts of IBM (and I'm sure the rest of the corporate world) have already forgotten that.
I used to work there, and there was a big effort on to have employees maintain a 'skills' database. It was clear that despite running top class courses on teams, there were influential people in the corporation who saw staff as being nothing more than a set of D&D type stats who could be deployed like pawns.
And hey, the losses in morale, effectiveness and customer satisfaction might be offset by the cost reductions. Who knows. I'm just glad I don't work there any more. (Which is quite lucky, because I didn't resign - my business unit was sold).
Just like a typical panorama stitcher, it identifies similar points, then runs an optimisation algorithm in order to line those points up.
Whereas a panorama stitcher warps the images to match a particular projection, and optimises the points in 2D, PhotoSynth optimises the points in 3D.
The viewer application then lets you view the collection of photographs, as if they were hanging in 3D space -- in the right part of space -- and fade in and out of view as you stand in the right place to see them.
For quite a small number of photos, you get a BIG cloud of control points, and the application lets you view that cloud and hides the photos. Often the result is quite a good 3D model - it's clear that if you were to draw vertices between them you'd get a decent wireframe of the subject.
However, the application does not attempt to turn your photos into a convincing fully rendered 3D model. Rather it provides a spacial model for navigating between photos. It's always explicit that you are looking at one photo, with some other photos, dimmer, around it.
I think that's quite nice - that it doesn't pretend to be more than it is.
The slideshow option is rather neat. It simply steps through all the photos, but the transition between them shows you how they are spacially related.
But can we honestly respect climate predictions ranging in the 10+ years when the same scientists can't predict whether it's going to rain tomorrow with more than a 10% success?
Actually I can tell you with almost 100% certainty that it will rain tomorrow. If you want me to be more specific about location, well, we could draw a 1000km grid, and I reckon I could make pretty good predictions about each million square kilometer square.
Forecasting gets harder with granularity, and harder with time. 10 years is a long time, but "the whole damn planet" is not very granular at all.
A battery that will hold a couple of days' worth of charge with minimal loss? Please give an example.
If you're getting enough cheap renewable energy, even quite high losses are acceptable.
Power companies quite often have hydroelectric schemes, in order to smooth out discrepancies in supply and demand of power. When supply exceeds demand, the surplus is used to pump water uphill into a reservoir. When demand exceeds supply, sluices are opened, and water turbines generate electricity.
Yes, there are losses - thanks a bunch Newton - but the alternative is to throw energy away, or to run power stations at suboptimal rates.
Traditionally, this is used when the power supply wants to run at a constant rate (e.g. a coal fired power station running at its optimum rate) and the consumers demand varies. But it's totally adaptable to unpredictable sources such as wind.
OTOH when we get serious about wind power, we'll have kite driven generators at elevations where it's *always* windy.
No need for the attitude. But other than that, you're right.
I walk to work. Previous to that I took the bus, and for a while before that I rode a moped.
I like to collaborate, and not being able to leap up and whiteboard stuff on the spur of the moment would be a real barrier for me. Netmeeting etc. is OK (and I do this with transatlantic colleagues) but it's not as good.
I keep diaries of my holidays. I'm certain that without the diaries, there's stuff I'd forget that I wanted to remember. I know this because if I write the diary a day late, I already struggle to remember details and have to ask people ("Where was it we had lunch yesterday?").
So what would happen if I put a minor untruth in there - a distortion perhaps? Maybe I wouldn't read it until 10 years later, when I'll forget that I lied, combine my lie with my hazy real memories, and end up remembering the lie as truth.
I've never heard the term "ski mask" before, I don't think it's used here (we don't have many good places for skiing). Everyone knows what a balaclava is though.
You know how nowadays, a cartoon of a terrorist will have a headscarf? Well, in the 70s and 80s, a cartoon IRA terrorist always wore a balaclava.
I attended an Akamai presentation a while back (some of my colleagues were reselling Akamai services). At the time, and I see no reason why it would change, they said each cache location deal was negotiated individually.
As you said, there are gains on both sides. Akamai gain performance for their customers by having a cache on the ISP's network. The ISP saves on traffic leaving their own network, and their customers see improved performance from Akamai-enabled sites.
Who wins most, depends on the network. Hence, in some cases, Akamai pay to have their server at your site. In other cases, no money changes hands. In some cases, the ISP actually pays Akamai.
Brackup is a Perl script for doing incremental backups. It's a bit like JungleDisk (see other posts here), except the client is freeware.
While JungleDisk uses S3 for storage, Brackup has the option to store on S3 or on a filesystem. You basically end up with (optionally) encrypted blocks, referenced by keys.
I've taken a peek at the code. You wouldn't need to be much of a Perl guru to hack in FTP support if that's what you want.
This isn't an endorsement for Brackup - I don't use it myself, although I've given it a trial run, been impressed, and may well take it further when I get around to it.
I was going to say someone completely different than Google. I do trust Google to do a good for a whole lot of things. But handing them all the applications on the internet would make them too powerful.
Issuing certificates doesn't give you that much power. The CA never gets to know the certificate private key. The CA can't masquerade as the legitimate certificate owner (although of course, the CA can create whatever certificates it likes).
An SSL certificate is for encryption of a link. Absolutely not trust and absolutely not identity.
No. A certificate is ALL ABOUT identity. A certificate is something you process before you begin sending/receiving encrypted data. You might end up using the key that was in the certificate for encryption, but not necessarily.
If that were the case why in the world did the major certificate vendors come up with an "EV Cert" which claims to do what you say? As opposed to the certs they already offer which obviously don't offer that trust. Otherwise they wouldn't need the new EV certs.
This is because trust is not binary, and it's never 100%.
They are notifying you because your house previously had a premises license. However as it's now domestic (I presume) then simply notifying them that it no longer has a premises license and that it is a domestic dwelling should be enough.
Well, obviously. Ignoring them is fine. The question is how much fun can be had by entering into a dialogue.
So you'd either have to play a record which is out of copyright (good luck finding such an old record and hooking up a suitable record player to a modern amp - few modern decks will play at 78RPM) or find 100 year old sheet of music and pay a performer to play that alone.
In other words, there are exceptions but they're sufficiently esoteric that it's vanishingly unlikely that anyone will take advantage of them.
Slashdot Mirror
Sounds like a classic Unix architecture to me! Before threads, there were processes and signals, and it still works pretty well... I wonder how it will affect memory usage though.
The comic addresses this.
There is an upfront memory overhead in starting a new process. However, there are long term memory usage gains. The comic suggests that fragmentation and memory leakage occurs in the browser's memory space.
Sure, you could work to improve the browser's memory management - but by moving to process-per-tab, you make it a non-problem. When you close a tab, the OS reclaims that process's memory. You'd hope that the OS has rock solid memory management code.
If you look at the bigger picture Java, C and C++ are all just languages
Except that Java is a couple of things other than being a language.
Java means different things to different people. I think the original questioner would like to know what Java should mean to him.
This is all common sense, except that sometimes there's a high cost to finding out whether you want some framework.
Spring is a perfect example. Even to understand what Spring is for, you're going to need a couple of days of fairly solid reading. "Inversion of Control", "Dependency Injection", "Aspect Oriented Programming" - these are all big new concepts to most people.
Put yourself in the position of someone who doesn't know Spring. How much study are they going to need to do before deciding "yes, I can apply this stuff to my project"? Time that, if the answer turns out to be "Oh no, that framework's not for my kind of work", would be wasted.
So, one needs guidance. If someone you trust were to say "Hey, that feature of Spring would really help you with that part of your project", that would be a great help.
Not picking on Spring, by the way. The same applies to most useful frameworks. If they didn't involve big new concepts, they probably wouldn't be all that useful anywhere.
Will IBM? I would posit that they will, and that as people become less stuck on "MY PROJECT" and think a little more globally ("MY TEAM'S PROJECTS"), people will become more efficient and thus more happy. For every person who is "torn away" from their project, there will be others who aren't stuck working 80 hour weeks because some awful manager didn't predict the work load accurately.
I'd be more concerned about being "torn away" from a team, than from a project. Of course, the models *could* factor in the fact that it takes time to build a working relationship, and build that cost into its heuristics.
All true. But parts of IBM (and I'm sure the rest of the corporate world) have already forgotten that.
I used to work there, and there was a big effort on to have employees maintain a 'skills' database. It was clear that despite running top class courses on teams, there were influential people in the corporation who saw staff as being nothing more than a set of D&D type stats who could be deployed like pawns.
And hey, the losses in morale, effectiveness and customer satisfaction might be offset by the cost reductions. Who knows. I'm just glad I don't work there any more. (Which is quite lucky, because I didn't resign - my business unit was sold).
Then again, presumably you still have to launch a mouse and keyboard and VDU for this thing, so it's not quite as great a saving as it sounds...
Depends on its job. It could be a headless device for logging/collating/forwarding sensor data, controlling a robot, whatever.
Hold down CTRL.
Alternatively, hit P to cycle through view modes (cloud, photos, both).
Quicktime VR
As seen on many an estate agent's web site.
The GP is correct - aside from platform dependence problems, Photosynth would be a great way to present "virtual tours".
Here's what it does:
Just like a typical panorama stitcher, it identifies similar points, then runs an optimisation algorithm in order to line those points up.
Whereas a panorama stitcher warps the images to match a particular projection, and optimises the points in 2D, PhotoSynth optimises the points in 3D.
The viewer application then lets you view the collection of photographs, as if they were hanging in 3D space -- in the right part of space -- and fade in and out of view as you stand in the right place to see them.
For quite a small number of photos, you get a BIG cloud of control points, and the application lets you view that cloud and hides the photos. Often the result is quite a good 3D model - it's clear that if you were to draw vertices between them you'd get a decent wireframe of the subject.
However, the application does not attempt to turn your photos into a convincing fully rendered 3D model. Rather it provides a spacial model for navigating between photos. It's always explicit that you are looking at one photo, with some other photos, dimmer, around it.
I think that's quite nice - that it doesn't pretend to be more than it is.
The slideshow option is rather neat. It simply steps through all the photos, but the transition between them shows you how they are spacially related.
But can we honestly respect climate predictions ranging in the 10+ years when the same scientists can't predict whether it's going to rain tomorrow with more than a 10% success?
Actually I can tell you with almost 100% certainty that it will rain tomorrow. If you want me to be more specific about location, well, we could draw a 1000km grid, and I reckon I could make pretty good predictions about each million square kilometer square.
Forecasting gets harder with granularity, and harder with time. 10 years is a long time, but "the whole damn planet" is not very granular at all.
A battery that will hold a couple of days' worth of charge with minimal loss? Please give an example.
If you're getting enough cheap renewable energy, even quite high losses are acceptable.
Power companies quite often have hydroelectric schemes, in order to smooth out discrepancies in supply and demand of power. When supply exceeds demand, the surplus is used to pump water uphill into a reservoir. When demand exceeds supply, sluices are opened, and water turbines generate electricity.
Yes, there are losses - thanks a bunch Newton - but the alternative is to throw energy away, or to run power stations at suboptimal rates.
Traditionally, this is used when the power supply wants to run at a constant rate (e.g. a coal fired power station running at its optimum rate) and the consumers demand varies. But it's totally adaptable to unpredictable sources such as wind.
OTOH when we get serious about wind power, we'll have kite driven generators at elevations where it's *always* windy.
The original CERN WWW browser didn't keep cookies, didn't maintain a history, and didn't cache pages. Is that therefore prior art?
Then again, my coffee cup does none of those things either - it doesn't even browse Web pages. Now *that*'s privacy...
How can you patent the absence of a feature (or more accurately, disabling a feature)?
Move closer to work and ride a bicycle, asshole.
No need for the attitude. But other than that, you're right.
I walk to work. Previous to that I took the bus, and for a while before that I rode a moped.
I like to collaborate, and not being able to leap up and whiteboard stuff on the spur of the moment would be a real barrier for me. Netmeeting etc. is OK (and I do this with transatlantic colleagues) but it's not as good.
That's a complete non-sequitur.
It has always been clear that Gmail was vulnerable to eavesdropping.
SSL (which you seem to accept is secure) - is built on top of TCP (which is insecure) - which is built on top of IP (which is not even reliable).
Likewise, if you encrypt stuff right, you can securely store it on an anonymous FTP server.
Likewise, if you encrypt messages, you can broadcast them over short wave radio, and they're only of use to the intended recipient.
I keep diaries of my holidays. I'm certain that without the diaries, there's stuff I'd forget that I wanted to remember. I know this because if I write the diary a day late, I already struggle to remember details and have to ask people ("Where was it we had lunch yesterday?").
So what would happen if I put a minor untruth in there - a distortion perhaps? Maybe I wouldn't read it until 10 years later, when I'll forget that I lied, combine my lie with my hazy real memories, and end up remembering the lie as truth.
Answer: there is NO SECURE STORAGE using a NON_SECURE PROTOCOL. Next.
Just as a trivial counterexample - encrypt a message with S/MIME and you could securely store it as a MySpace profile.
Building trustworthy layer atop untrustworthy layers is what cryptographic security is all about.
I've never heard the term "ski mask" before, I don't think it's used here (we don't have many good places for skiing). Everyone knows what a balaclava is though.
You know how nowadays, a cartoon of a terrorist will have a headscarf? Well, in the 70s and 80s, a cartoon IRA terrorist always wore a balaclava.
I attended an Akamai presentation a while back (some of my colleagues were reselling Akamai services). At the time, and I see no reason why it would change, they said each cache location deal was negotiated individually.
As you said, there are gains on both sides. Akamai gain performance for their customers by having a cache on the ISP's network. The ISP saves on traffic leaving their own network, and their customers see improved performance from Akamai-enabled sites.
Who wins most, depends on the network. Hence, in some cases, Akamai pay to have their server at your site. In other cases, no money changes hands. In some cases, the ISP actually pays Akamai.
LimeLight is probably no different.
Brackup is a Perl script for doing incremental backups. It's a bit like JungleDisk (see other posts here), except the client is freeware.
While JungleDisk uses S3 for storage, Brackup has the option to store on S3 or on a filesystem. You basically end up with (optionally) encrypted blocks, referenced by keys.
I've taken a peek at the code. You wouldn't need to be much of a Perl guru to hack in FTP support if that's what you want.
This isn't an endorsement for Brackup - I don't use it myself, although I've given it a trial run, been impressed, and may well take it further when I get around to it.
I was going to say someone completely different than Google. I do trust Google to do a good for a whole lot of things. But handing them all the applications on the internet would make them too powerful.
Issuing certificates doesn't give you that much power. The CA never gets to know the certificate private key. The CA can't masquerade as the legitimate certificate owner (although of course, the CA can create whatever certificates it likes).
An SSL certificate is for encryption of a link. Absolutely not trust and absolutely not identity.
No. A certificate is ALL ABOUT identity. A certificate is something you process before you begin sending/receiving encrypted data. You might end up using the key that was in the certificate for encryption, but not necessarily.
If that were the case why in the world did the major certificate vendors come up with an "EV Cert" which claims to do what you say? As opposed to the certs they already offer which obviously don't offer that trust. Otherwise they wouldn't need the new EV certs.
This is because trust is not binary, and it's never 100%.
i still don't see how any certificate guarantees that mybank.com hasn't been pwned
I still don't see how any certificate guarantees that there's not a burglar downstairs stealing my cash while I'm upstairs using mybank.com's website.
Oh, you say that's not part of X.509's scope?
True, but by restricting themselves to such music, a cafe could save some £400 per year.
They are notifying you because your house previously had a premises license. However as it's now domestic (I presume) then simply notifying them that it no longer has a premises license and that it is a domestic dwelling should be enough.
Well, obviously. Ignoring them is fine. The question is how much fun can be had by entering into a dialogue.
So you'd either have to play a record which is out of copyright (good luck finding such an old record and hooking up a suitable record player to a modern amp - few modern decks will play at 78RPM) or find 100 year old sheet of music and pay a performer to play that alone.
In other words, there are exceptions but they're sufficiently esoteric that it's vanishingly unlikely that anyone will take advantage of them.
Or music published by the author under a Creative Commons licence. Or my own compositions played by myself. Or folks songs performed by myself without reference to sheet music.