Modern reactor designs will improve things over old-style designs a good deal, but the extremely high energy density in the core makes it pretty much impossible to avoid entirely. And even the best old-style designs (CANDUs) have taught us that small errors in engineering result in billions of extra costs down the road--appropos of the article, the problem with CANDUs was excessive corrosion brought on by unexpectedly high radiation damage to the fuel channels.
Thats the beautiful thing about PBNRs. They DON'T have a "extremely high energy density" which makes them suspectible to unforseen failures causing extremely expensive diasters. It's unfortunate for the US (and fortunate for China/South Africa) that it has decided to sit out and "observe" how PBNRs will work out. I think it will be an excellent long term design.
I agree. KDE's stack is much better then gnome or enlightenment. Maybe because it's using C++ (allowing for cleaner OOD) or maybe the kde devs are being more careful about interface design. I know that the kde APIs tend to be excellent and well thought out. And they seem to be documented too!
DBus is an evolution of dcop. Basically, all of the guys at freedesktop.org saw dcop and said "We want to use this for our entire platform!" (Aka a common messaging protocol for the UI system) So they generalized dcop a bit and came up with dbus.
dbus works very similar to dcop, has an EXTREMELY clean method for defining message parameters and looks extremely well designed. Qt liked it so much, they've made it part of qt4.2+ which means that it'll eventually work on MacOS and windows if anyone ever bothers to write a dbus server for those platforms. (Maybe we'll see a COM+ dbus proxy as well) The interesting bit is that gnome is also using dbus so there's the glimmering possibility that the backend services (aka CUPS) will expose a common dbus interface and gaim/kopete will have common dbus interfaces so manipulating them will be identical. This assumes that the developers can hammer out a common spec.
Solaris sounds like an impressive piece of engineering. I'm sure it's difficult/impossible to replicate the flexibility with linux currently but we'll have to see how the technology evolves. I like the fact that it seems like it's getting easier to mix and match systems for best performance/TOC at least on the UNIX side. Lets you use the best tool for the job. Just try to fit windows in there somewhere and you might be up s**t creek.
Yes, *IF* a secure channel has been set up by the protected media, and only *IF* that is the case.
That's the problem, it's impossible to tell when Vista is going to degradate the media channels because the logic is deeply ingrained inside the Vista kernel. And there's NO way to disable it, period! So would you want to trust it not to f**k you at the wrong time?
I wouldn't and that's why I don't use Windows for my main OS anymore...
Of course, after the M$ marketeers are done with them, CIOs in mid - large sized companies will be dying to lockdown/channelize their operations into the DRM monstrasty that M$ wants to sell them.
Sounds about right for the current american business environment....
not to mention it's 3 times faster on some workloads,
the test was done using a specialized memory allocation library for OpenSolaris! I suspect the other test was done using the same memory allocation library.... Not a really fair comparision...
However, the idea of using an allocation lib optimized for multithreading is extremely clever. It would be better if the application could dynamically choose whether to use the single threaded version or multithreaded version depending on the number of CPUs in the machine.
Linux is known as a server OS in the corporate world, period,
Not really, it's also an embedded OS. at least at our company. I doubt you'll find any company willing to spend $6/unit license for WinCE or Symboian any more. Linux does the job fine. And there's a lot more embedded CPUs then desktop ones.
Use gentoo. It's revdep-rebuild tool is amazing and works 100%. The only problem is identifying when functionality is missing because of a missing kernel driver and making sure that you're not too much on the bleeding edge. (like using dbus 0.62 "stable" vs 1.0.1 "testing")
This requires you to be picky about which packages you "absolutely" need.
Otherwise, I find gentoo's bleeding-edgeness and ability to navigate "dependency hell" unbeatable. Hats off to the maintainers. And I find the time for rebuilds not too bad. (amd x64 3200+ 1GB Ram)
I really hate to break the dream-bubble here guys, but we need a "Standard Linux Desktop" specification that fully defines the available libraries and their versions all the way from libc to gnome.
The problem is that everyone wants to use their own stack: kernel (2.4vs2.6) -> libc (5vs6) -> gcc (2.95 vs 3.4 vs 4.2) -> WM (gnome vs kde), etc, etc. Everyone wants everything customizable but still 100% functional and bug free. It's inevitiable unless one really big player (ala redhat) imposes their stack on everyone else and bug fixes it. I think that's why RHE9 is always 3+ years behind the bleeding edge. Bug fixes and stabilization. And as soon as you define a "Standard" spec. it;ll be out of date and people will complain why there isn't a "Standard" spec for all of the stuff which isn't unified. I'm quite surprised that DBus was so uniformly excepted by everyone. Now I think there'll be namespace fragmentation when people can agree on which dbus channels should be common or not... Its a neverending story...
Funny thing is: Open source software is extremely easy to modify and keep up to date. In fact, it's easier then doing it with closed source because you don't have to start from scratch every time.
If the librarians are smart, they'll continue to insist on contractors developing the modules in an open source manner so they can avoid the lock-in effect and be able to pick and choose vendors for lower prices. Keeping this stuff open is more an economic issue then a political/moral one. I'm sure whoever is running the show at GPLS is probably aware of this.
The very sad thing is that, it seems that whether we know about it or not, there is nothing we the people can do to stop this administration from going to war against Iran if they are determined to do so. It shouldn't be this way.
I really, really hope you are wrong. Letting Bush drag us into a war with Iran would be even more idiotic then the iraqi f**kup. Besides which, I don't think bush can get the backing he had in 2001-2. He's a lame-duck president and the congress is currently democratic. If he tried to bomb without congressional approvable, he'd be impeached so fast it would make your head spin.
Javascript is currently evolving towards Python, which I consider much more fit to what JS currently is and its past history.
Please, please don't tell me that JS will lose it's braces and go for whitespace scoping... not going to happen because of the need to put js code into event handlers. otherwise, I hope it doesn't get OOP tacked on the way python has it. (explicitly having to declare self as a param in every method? WTF?) I see it more like ruby because of the open/closed nature of ruby classes, which is closely mirrored in JS's prototype lookup system. Compare how ruby finds the correct method to execute vs JS. Its the same bloody thing. (i haven't read how python does it though...)
The initial function on which you use new is merely a constructor, the only thing that would be gotten from a more regular class syntax in that point would be... well... nothing at all. Grrrr, this is both freeing and extremely scarely. I really live js for quick and dirty app development but without serious considerations for coding standards, I doubt its possible to build large multi-person projects in it. And I think that the language itself should enforce 99% of the coding standard. So by allowing arbitrary structure into your class declaration/prototype function, you really need to be careful. I guess js is just one of those languages which you can't live with and can't live without.
Recently, I have been looking harder and harder at Linux. Linux offers a much more stable platform, and I can customize the installation to make it much more difficult to corrupt. The issue is that such a high software investment has been placed in specialized Windows solutions, that it is difficult to port everything to another operating system overnight.
If someone was smart, they'd start advertisting the fact that they support a distro with high availability and with standardized APIs and file locations for medical software manufactors. Then get more medical software manufactors to build to their distro. Like the LSB but with the whole system spec'ed out.
Nearly everything is targetted toward Monta Vista, these days.
Sigh... i guess it's easier to go with a prebuild toolchain then roll your own... However, I found it pretty easy just to roll a custom toolchain ala crosstool then to use Montavista's toolchain. Plus then I could use the latest version.
The upshot is that we get a gnu toolchain royalty free which can be customized to different platforms but use the same gcc+binutils combo. The downside is that there is no support... sigh... (of course, the major goto guys in the arm-linux list now are montavista employees...)
That said, with Perl I can apply a regexp with =~ whereas with JavaScript you need to go off creating a new RegExp object with more verbose and hence hard-to-remember syntax...
So, what does =~ mean? I mean, unless you memorize Perl's 1001 operators, you're pretty much up shit cheek. Where as OO languages only have a small number of operators and a very large library.
Perl is the worst were it has way too many operators and it also has a huge library. So you have to memorize both to make sure you understand the language. Plus the compiler and interperter must be complex and inefficent to handle all of the extraneous operators.
There's also M$ jscript inside IIS. On a project awhile ago, I had the choice of working with VBscript or JScript. Of course I went with JScript and was Incrediably happy that I had done so. There's some weird problems in IIS's implementation but in comparsion to VBScript, it was heavenly....
Additionality, it allowed me to share cope between the backend (in JScript) and the flash frontend (in ActionScript). Maybe a bit bizarre but it worked!
I would consider Ruby to be the correct direction that Javascript should evolve towards.
The way it it allows you to easily add properties and it's reflexiviness is extremely similar to javascript. It has a bunch of other freatures I'd love to see in JS one day, like hiding instance variables, class+module concepts instead of having to using function() to initialize members, and continuations.
Otherwise, I think that JS is a perfect lowlevel scripting language.
Its too bad the original poster can't/won't submit his huge speadsheet to OO devs as a test case. I'm sure you could do this with M$ because they can sign an NDA.
Doing the same with Unix extra/user/group/other permissions... Well -- even I have problems doing that. Allowing all users, but denying a specific user? I actually have to think very hard on the best way to-do that.
Thats because Unix e/u/g/o premissions are used to give access, not deny access. There is no concept of deny access. The best answer is to group all of the users you want to have access to a resource into a group and then make the resource group rwx. Of course, normal users can't create groups and add users to it so this isn't very flexible...
Slashdot Mirror
Modern reactor designs will improve things over old-style designs a good deal, but the extremely high energy density in the core makes it pretty much impossible to avoid entirely. And even the best old-style designs (CANDUs) have taught us that small errors in engineering result in billions of extra costs down the road--appropos of the article, the problem with CANDUs was excessive corrosion brought on by unexpectedly high radiation damage to the fuel channels.
Thats the beautiful thing about PBNRs. They DON'T have a "extremely high energy density" which makes them suspectible to unforseen failures causing extremely expensive diasters. It's unfortunate for the US (and fortunate for China/South Africa) that it has decided to sit out and "observe" how PBNRs will work out. I think it will be an excellent long term design.
Cheers
Ben
Hi,
things like using OSX keyboard shortcuts
There is Control Center->Region&Accessibility->Keyboard Shortcuts. You can select the "Mac Scheme"
not having a Start menu
There is KSmoothDock and its forks. Or you can try kbfx for a twist on kicker.
Cheers
Ben
I agree. KDE's stack is much better then gnome or enlightenment. Maybe because it's using C++ (allowing for cleaner OOD) or maybe the kde devs are being more careful about interface design. I know that the kde APIs tend to be excellent and well thought out. And they seem to be documented too!
Cheers,
Ben
DBus is an evolution of dcop. Basically, all of the guys at freedesktop.org saw dcop and said "We want to use this for our entire platform!" (Aka a common messaging protocol for the UI system) So they generalized dcop a bit and came up with dbus.
dbus works very similar to dcop, has an EXTREMELY clean method for defining message parameters and looks extremely well designed. Qt liked it so much, they've made it part of qt4.2+ which means that it'll eventually work on MacOS and windows if anyone ever bothers to write a dbus server for those platforms. (Maybe we'll see a COM+ dbus proxy as well) The interesting bit is that gnome is also using dbus so there's the glimmering possibility that the backend services (aka CUPS) will expose a common dbus interface and gaim/kopete will have common dbus interfaces so manipulating them will be identical. This assumes that the developers can hammer out a common spec.
Cheers
Ben
Solaris sounds like an impressive piece of engineering.
I'm sure it's difficult/impossible to replicate the flexibility with linux currently but we'll have to see how the technology evolves.
I like the fact that it seems like it's getting easier to mix and match systems for best performance/TOC at least on the UNIX side.
Lets you use the best tool for the job. Just try to fit windows in there somewhere and you might be up s**t creek.
Cheers
Ben
Yes, *IF* a secure channel has been set up by the protected media, and only *IF* that is the case.
That's the problem, it's impossible to tell when Vista is going to degradate the media channels because the logic is deeply ingrained inside the Vista kernel. And there's NO way to disable it, period! So would you want to trust it not to f**k you at the wrong time?
I wouldn't and that's why I don't use Windows for my main OS anymore...
Cheers
Ben
Of course, after the M$ marketeers are done with them, CIOs in mid - large sized companies will be dying to lockdown/channelize their operations into the DRM monstrasty that M$ wants to sell them.
Sounds about right for the current american business environment....
Cheers
Ben
not to mention it's 3 times faster on some workloads,
the test was done using a specialized memory allocation library for OpenSolaris! I suspect the other test was done using the same memory allocation library....
Not a really fair comparision...
However, the idea of using an allocation lib optimized for multithreading is extremely clever.
It would be better if the application could dynamically choose whether to use the single threaded version or multithreaded version depending on the number of CPUs in the machine.
Cheers
Ben
Photoshop works fine for me under WINE.
:(
However, directshow still isn't implemented in Crossover office so no video apps are going to work correctly.
Maybe in another year or so....
Ben
Linux is known as a server OS in the corporate world, period,
Not really, it's also an embedded OS. at least at our company.
I doubt you'll find any company willing to spend $6/unit license for WinCE or Symboian any more.
Linux does the job fine. And there's a lot more embedded CPUs then desktop ones.
Sic
Ben
Use gentoo. It's revdep-rebuild tool is amazing and works 100%.
The only problem is identifying when functionality is missing because of a missing kernel driver and
making sure that you're not too much on the bleeding edge. (like using dbus 0.62 "stable" vs 1.0.1 "testing")
This requires you to be picky about which packages you "absolutely" need.
Otherwise, I find gentoo's bleeding-edgeness and ability to navigate "dependency hell" unbeatable. Hats off to the maintainers. And I find the time for rebuilds not too bad. (amd x64 3200+ 1GB Ram)
Cheers
Ben
I really hate to break the dream-bubble here guys, but we need a "Standard Linux Desktop" specification that fully defines the available libraries and their versions all the way from libc to gnome.
The problem is that everyone wants to use their own stack: kernel (2.4vs2.6) -> libc (5vs6) -> gcc (2.95 vs 3.4 vs 4.2) -> WM (gnome vs kde), etc, etc. Everyone wants everything customizable but still 100% functional and bug free. It's inevitiable unless one really big player (ala redhat) imposes their stack on everyone else and bug fixes it. I think that's why RHE9 is always 3+ years behind the bleeding edge. Bug fixes and stabilization. And as soon as you define a "Standard" spec. it;ll be out of date and people will complain why there isn't a "Standard" spec for all of the stuff which isn't unified. I'm quite surprised that DBus was so uniformly excepted by everyone. Now I think there'll be namespace fragmentation when people can agree on which dbus channels should be common or not... Its a neverending story...
Cheers
Ben
Funny thing is: Open source software is extremely easy to modify and keep up to date.
In fact, it's easier then doing it with closed source because you don't have to start from scratch every time.
If the librarians are smart, they'll continue to insist on contractors developing the modules in an open source manner so they can avoid the lock-in effect and be able to pick and choose vendors for lower prices. Keeping this stuff open is more an economic issue then a political/moral one.
I'm sure whoever is running the show at GPLS is probably aware of this.
Cheers
Ben
The very sad thing is that, it seems that whether we know about it or not, there is nothing we the people can do to stop this administration from going to war against Iran if they are determined to do so. It shouldn't be this way.
I really, really hope you are wrong. Letting Bush drag us into a war with Iran would be even more idiotic then the iraqi f**kup.
Besides which, I don't think bush can get the backing he had in 2001-2. He's a lame-duck president and the congress is currently democratic.
If he tried to bomb without congressional approvable, he'd be impeached so fast it would make your head spin.
Just my $0.02
Ben
Javascript is currently evolving towards Python, which I consider much more fit to what JS currently is and its past history.
Please, please don't tell me that JS will lose it's braces and go for whitespace scoping... not going to happen because of the need to put js code into event handlers. otherwise, I hope it doesn't get OOP tacked on the way python has it. (explicitly having to declare self as a param in every method? WTF?)
I see it more like ruby because of the open/closed nature of ruby classes, which is closely mirrored in JS's prototype lookup system. Compare how ruby finds the correct method to execute vs JS. Its the same bloody thing. (i haven't read how python does it though...)
The initial function on which you use new is merely a constructor, the only thing that would be gotten from a more regular class syntax in that point would be... well... nothing at all.
Grrrr, this is both freeing and extremely scarely. I really live js for quick and dirty app development but without serious considerations for coding standards, I doubt its possible to build large multi-person projects in it. And I think that the language itself should enforce 99% of the coding standard. So by allowing arbitrary structure into your class declaration/prototype function, you really need to be careful.
I guess js is just one of those languages which you can't live with and can't live without.
because he who owns factories owns the business, ultimately.
Maybe not. If the service industry can supply jobs and steady salaries to people, maybe it's a enough replacement.
I don't know of course
Cheers
Ben
Recently, I have been looking harder and harder at Linux. Linux offers a much more stable platform, and I can customize the installation to make it much more difficult to corrupt. The issue is that such a high software investment has been placed in specialized Windows solutions, that it is difficult to port everything to another operating system overnight.
If someone was smart, they'd start advertisting the fact that they support a distro with high availability and with standardized APIs and file locations for medical software manufactors. Then get more medical software manufactors to build to their distro. Like the LSB but with the whole system spec'ed out.
Cheers
Ben
Nearly everything is targetted toward Monta Vista, these days.
Sigh... i guess it's easier to go with a prebuild toolchain then roll your own...
However, I found it pretty easy just to roll a custom toolchain ala crosstool then to use Montavista's toolchain. Plus then I could use the latest version.
The upshot is that we get a gnu toolchain royalty free which can be customized to different platforms but use the same gcc+binutils combo.
The downside is that there is no support... sigh... (of course, the major goto guys in the arm-linux list now are montavista employees...)
Cheers
Ben
That said, with Perl I can apply a regexp with =~ whereas with JavaScript you need to go off creating a new RegExp object with more verbose and hence hard-to-remember syntax...
So, what does =~ mean? I mean, unless you memorize Perl's 1001 operators, you're pretty much up shit cheek. Where as OO languages only have a small number of operators and a very large library.
Perl is the worst were it has way too many operators and it also has a huge library. So you have to memorize both to make sure you understand the language. Plus the compiler and interperter must be complex and inefficent to handle all of the extraneous operators.
Give me new RegExp() anyday...
Cheers
Ben
There's also M$ jscript inside IIS.
On a project awhile ago, I had the choice of working with VBscript or JScript.
Of course I went with JScript and was Incrediably happy that I had done so. There's some weird problems in IIS's implementation but in comparsion to VBScript, it was heavenly....
Additionality, it allowed me to share cope between the backend (in JScript) and the flash frontend (in ActionScript). Maybe a bit bizarre but it worked!
Ben
I would consider Ruby to be the correct direction that Javascript should evolve towards.
The way it it allows you to easily add properties and it's reflexiviness is extremely similar to javascript.
It has a bunch of other freatures I'd love to see in JS one day, like hiding instance variables, class+module concepts instead of having to using function() to initialize members, and continuations.
Otherwise, I think that JS is a perfect lowlevel scripting language.
Cheers
Ben
Its too bad the original poster can't/won't submit his huge speadsheet to OO devs as a test case.
I'm sure you could do this with M$ because they can sign an NDA.
Cheers
Ben
Doing the same with Unix extra/user/group/other permissions... Well -- even I have problems doing that. Allowing all users, but denying a specific user? I actually have to think very hard on the best way to-do that.
Thats because Unix e/u/g/o premissions are used to give access, not deny access. There is no concept of deny access.
The best answer is to group all of the users you want to have access to a resource into a group and then make the resource group rwx.
Of course, normal users can't create groups and add users to it so this isn't very flexible...
Cheers
Ben
I dont think Explorer is crashing because of DEP.
;) Ben
The moment it has to build thumbnails of your porn, it probably shuts down and quits...
You Sicko!!!
To an optimist, the glass is half full.
To a pessimist, the glass is half empty.
Maybe I should say that evolutions would like to understand how and why the human body evolved instead of saying they want to "improve" it.
Maybe it's a way to determine how we can survive when natural selection becomes a hell of a lot more picky. (difficult to survive)
Ben