The final nail in the desktop coffin is the news that the Arm version of Windows will not support desktop applications. So if you want to write an app and you want it to work on any version of Windows then it has to be a Metro app.
Slowly we have to transition from Win32,.NET and Silverlight to HTML/XAML built on top of WinRT.
But I think they are now saying that win8's desktop tile will be on ARM tablets. But who knows! I think we'll have to wait and see.
I agree on gnome, I'm enjoying gnome-shell a lot, so far.
True, but would win32 programs be very useful on an ARM tablet? I imagine the reason MS originally said "no win32 on ARM" was to try to force devs to redesign for touch, not because there were significant technical problems.
The bytecode can run, but the APIs are different on win8 metro. Some stuff added, some stuff removed, you can imagine.
It might run unmodified in win8 desktop (the win32 compatibility mode) but it's still a little unclear whether this will be supported on ARM tablets. MS said it wouldn't be supported, then someone said it would, and maybe it's changed again since I last heard.
In any case, using a program designed for mouse+kb on a small touchscreen will not be a great experience, so it might not be very useful anyway.
They have a new add-on update thing for exactly this reason.
Previously, all add-ons were marked "incompatible" by default on a major version change and authors needed to test their add-on and explicitly mark it as OK for it to work. This caused the painful loss of add-ons during major updates that you mention.
The new system scans the source code for every add-on automatically and flags as compatible all those which don't touch parts of firefox which have changed. As a result, as long as the author of the add-on didn't mess with the internals of the browser too much, add-ons will automatically move between versions with no work from anyone. I have 9 add-ons in my ff and they all moved to ff7 painlessly.
Add-ons which depend on some changed or removed feature will need updating, but there's not much you can do about that. Jetpack aimed to make add-ons less closely tied to the browser, but that seems to have gone rather quiet.
Another point is maintenance: the package installed via Synaptic will be updated automatically for you. Moreover, packages are updated in synchrony, so if a security fix needs changes to several libraries, all those libraries are updated for you in one go.
On Windows you have to do all that work yourself. Not just checking for and installing updates, but ensuring that the stack of installed packages you are using function together correctly.
I'm hoping that CoApp will bring something like this fantastic system to Windows. Fingers crossed!
The elevations people usually cite involve cases where things can be written by an unprivileged process which are then used by an elevated process. For example, there are various registry keys which a low-priv process can write which are executed by an elevating command prompt. The Ubuntu equivalent would be appending "alias sudo/my/sneaky/attack" to someone's.bashrc. Though this Windows once is a little worse since you can't (as far as I know) as a user inject things into gtk-sudo, which would be the main elevation route for most people.
I read a very long and interesting thread with Mark and others debating the details of this a few years ago, but of course now I can't find the link:( sorry.
The whitelist trick is just one of many mostly unfixable holes in Windows that make win7 UAC in default mode trivial to bypass. As you say, pushing the slider to maximum gets you Vista-level security: better but still not secure. You need a separate admin account to get something close to sudo.
As vendors make their software more UAC friendly, MS will eventually be able to have a non-admin default account without it being too annoying. But we're still a few years from that, sadly.
UAC is not much like sudo since it is not a security feature. It is not supposed to stop bad software doing bad things (since it can't, it's trivial to bypass), it's supposed to let users know that good software is doing system-level things.
I know Cambridge gets everything with an ISBN, and from your post it sounds like Wales and Scotland do too. Things like PhD thesis only go to the BL though.
Java and c# are rather close on that list, arent they? And both are slower than C++, sometimes significantly, on all but one test.
Plus those tests are mostly very small numerical benchmarks which will not stress the GC much and which are not very representative of typical applications. As the google paper showed, gc tuning is an important part of application performance for larger projects.
"Every bit as capable" for CPU, but the iPad2 was two to five times faster than the Tab 10 at graphics, and that's not changed with the 10.1. Strangely, the performance comparison linked at the top misses out any iPad/Tab graphics comparisons:
You need to reboot win7 for a lot more than just kernel changes. Though they can swap graphics drivers very cleanly now, which is great.
Package managers have a list of repositories they watch for updates and users can add any they please (if they have admin rights), or even make their own (handy for corporate IT). Everything is signed and encrypted, so it's pretty safe.
Google, for example, run a repository for Chrome. You could have a repository for any software source and it could manage any sort of installation for you. Hopefully the win8 software store and the OS X appstore will bring something like this to Windows and Mac as well, though I'm rather doubtful. We'll see!
MacPorts and Fink are more trouble than they are worth, usually. They just don't have enough package maintainers, sadly, and their policy of constant updates also means constant breakages. I've switched to jhbuild for making.app bundles.
Debian and Ubuntu are in a totally different class. Their repositories are just wonderful -- huge choice, expertly packaged, thoroughly tested -- and have saved me days and days of thankless tinkering.
Here's hoping OS X and Win get something as good in less than 5 years. You never know.
I think you're confusing freedom for developers and freedom for users of software.
The GPL takes freedom away from developers. If you incorporate software with a GPL license in your project, you will have a long list of:you musts" and "you mustn'ts" to comply with. Many developers hate this, of course.
The GPL guarantees freedoms for users of software. If you select a GPL package to solve a problem you have, you can be certain you will always be able to build and modify the latest version yourself.
Selecting the GPL for a project is a promise to your users to stay free.
Well, that's true. I work in medical research so I've seen plenty of papers overturning a result from decades earlier.
When accepted results are junked they are (generally) either details, or results which were details and which have become important. Other researchers try to build on the earlier result, discover that their experiments are failing, and start methodically checking their assumptions.
Peer review is not perfect of course, but it does work reasonably well. The overwhelming success of science. which rests in part on peer-review, is evidence of this.
I'm not sure you put enough emphasis on the competitiveness of research. Competition between scientists for grants is intense and getting grants depends on getting good papers out and (optionally) taking apart the work of your rivals. It just isn't possible for a good idea to be suppressed for very long or for a false idea to gain wide acceptance without being tested.
Slashdot Mirror
I do this too. Ubuntu works well on my 6,1 macbook, though the wifi can take a while to reconnect after coming out of suspend.
They were saying "no win32" originally:
But I think they are now saying that win8's desktop tile will be on ARM tablets. But who knows! I think we'll have to wait and see.
I agree on gnome, I'm enjoying gnome-shell a lot, so far.
True, but would win32 programs be very useful on an ARM tablet? I imagine the reason MS originally said "no win32 on ARM" was to try to force devs to redesign for touch, not because there were significant technical problems.
The bytecode can run, but the APIs are different on win8 metro. Some stuff added, some stuff removed, you can imagine.
It might run unmodified in win8 desktop (the win32 compatibility mode) but it's still a little unclear whether this will be supported on ARM tablets. MS said it wouldn't be supported, then someone said it would, and maybe it's changed again since I last heard.
In any case, using a program designed for mouse+kb on a small touchscreen will not be a great experience, so it might not be very useful anyway.
They have a new add-on update thing for exactly this reason.
Previously, all add-ons were marked "incompatible" by default on a major version change and authors needed to test their add-on and explicitly mark it as OK for it to work. This caused the painful loss of add-ons during major updates that you mention.
The new system scans the source code for every add-on automatically and flags as compatible all those which don't touch parts of firefox which have changed. As a result, as long as the author of the add-on didn't mess with the internals of the browser too much, add-ons will automatically move between versions with no work from anyone. I have 9 add-ons in my ff and they all moved to ff7 painlessly.
Add-ons which depend on some changed or removed feature will need updating, but there's not much you can do about that. Jetpack aimed to make add-ons less closely tied to the browser, but that seems to have gone rather quiet.
That is K&R style, they have newlines for the braces that enclose function bodies.
They do omit brace newlines for struct, if, else, for and so on. I suppose the idea was to get more useful lines per page in the book.
Another point is maintenance: the package installed via Synaptic will be updated automatically for you. Moreover, packages are updated in synchrony, so if a security fix needs changes to several libraries, all those libraries are updated for you in one go.
On Windows you have to do all that work yourself. Not just checking for and installing updates, but ensuring that the stack of installed packages you are using function together correctly.
I'm hoping that CoApp will bring something like this fantastic system to Windows. Fingers crossed!
http://coapp.org/
The elevations people usually cite involve cases where things can be written by an unprivileged process which are then used by an elevated process. For example, there are various registry keys which a low-priv process can write which are executed by an elevating command prompt. The Ubuntu equivalent would be appending "alias sudo /my/sneaky/attack" to someone's .bashrc. Though this Windows once is a little worse since you can't (as far as I know) as a user inject things into gtk-sudo, which would be the main elevation route for most people.
I read a very long and interesting thread with Mark and others debating the details of this a few years ago, but of course now I can't find the link :( sorry.
Mark Russinovich says UAC is not a security feature:
http://www.networkworld.com/news/2007/021407-microsoft-uac-not-a-security.html
The whitelist trick is just one of many mostly unfixable holes in Windows that make win7 UAC in default mode trivial to bypass. As you say, pushing the slider to maximum gets you Vista-level security: better but still not secure. You need a separate admin account to get something close to sudo.
As vendors make their software more UAC friendly, MS will eventually be able to have a non-admin default account without it being too annoying. But we're still a few years from that, sadly.
UAC is not much like sudo since it is not a security feature. It is not supposed to stop bad software doing bad things (since it can't, it's trivial to bypass), it's supposed to let users know that good software is doing system-level things.
http://www.pretentiousname.com/misc/win7_uac_whitelist2.html
If you have a separate admin account UAC does work more like sudo. But that's not the default, sadly.
Strange, mine just went to the BL. Perhaps it depends upon the examining institution.
Actually the BL really is the only one to automatically get all publications. Five other libraries are entitled to a free copy upon request.
http://en.wikipedia.org/wiki/Legal_deposit#United_Kingdom
I know Cambridge gets everything with an ISBN, and from your post it sounds like Wales and Scotland do too. Things like PhD thesis only go to the BL though.
UK taxes are not that high. They are lower than in Canada, for example, according to Tax Freedom Day.
Java and c# are rather close on that list, arent they? And both are slower than C++, sometimes significantly, on all but one test.
Plus those tests are mostly very small numerical benchmarks which will not stress the GC much and which are not very representative of typical applications. As the google paper showed, gc tuning is an important part of application performance for larger projects.
"Every bit as capable" for CPU, but the iPad2 was two to five times faster than the Tab 10 at graphics, and that's not changed with the 10.1. Strangely, the performance comparison linked at the top misses out any iPad/Tab graphics comparisons:
http://www.anandtech.com/show/4216/apple-ipad-2-gpu-performance-explored-powervr-sgx543mp2-benchmarked/
Few consumers care about GPU specs, of course, but they are likely not notice if games on iOS look better.
You need to reboot win7 for a lot more than just kernel changes. Though they can swap graphics drivers very cleanly now, which is great.
Package managers have a list of repositories they watch for updates and users can add any they please (if they have admin rights), or even make their own (handy for corporate IT). Everything is signed and encrypted, so it's pretty safe.
Google, for example, run a repository for Chrome. You could have a repository for any software source and it could manage any sort of installation for you. Hopefully the win8 software store and the OS X appstore will bring something like this to Windows and Mac as well, though I'm rather doubtful. We'll see!
MacPorts and Fink are more trouble than they are worth, usually. They just don't have enough package maintainers, sadly, and their policy of constant updates also means constant breakages. I've switched to jhbuild for making .app bundles.
Debian and Ubuntu are in a totally different class. Their repositories are just wonderful -- huge choice, expertly packaged, thoroughly tested -- and have saved me days and days of thankless tinkering.
Here's hoping OS X and Win get something as good in less than 5 years. You never know.
XP SP2 is supported, isn't it?
http://gimp-win.sourceforge.net/stable.html
GIMP requires Windows XP SP2 or newer to run.
I think you're confusing freedom for developers and freedom for users of software.
The GPL takes freedom away from developers. If you incorporate software with a GPL license in your project, you will have a long list of :you musts" and "you mustn'ts" to comply with. Many developers hate this, of course.
The GPL guarantees freedoms for users of software. If you select a GPL package to solve a problem you have, you can be certain you will always be able to build and modify the latest version yourself.
Selecting the GPL for a project is a promise to your users to stay free.
But this contest is about exploiting via a browser (and perhaps email? I forget if they allow that).
Holes in GNOME aren't really relevant. Once you get some code running in the firefox process you co do whatever you like to the user's account.
Ubuntu was in Pwn2Own in 2008 and was not hacked in the three-day contest:
http://www.theregister.co.uk/2008/03/29/ubuntu_left_standing/
(though it sounds like they might have been able to break out of flash given a bit more time, who knows)
Well, that's true. I work in medical research so I've seen plenty of papers overturning a result from decades earlier.
When accepted results are junked they are (generally) either details, or results which were details and which have become important. Other researchers try to build on the earlier result, discover that their experiments are failing, and start methodically checking their assumptions.
Peer review is not perfect of course, but it does work reasonably well. The overwhelming success of science. which rests in part on peer-review, is evidence of this.
I'm not sure you put enough emphasis on the competitiveness of research. Competition between scientists for grants is intense and getting grants depends on getting good papers out and (optionally) taking apart the work of your rivals. It just isn't possible for a good idea to be suppressed for very long or for a false idea to gain wide acceptance without being tested.
(disclosure: I'm a working scientist)
I've ordered one of these for my kids for xmas:
http://www.amazon.com/Eyeclops-03548-Bionic-Eye-SE/dp/B0026G8SE6
It would certainly have appealed to me, anyway. We'll see! I'm anticipating a lot of DADDY DADDY LOOK AT MY SNOT
It's not stereo, actually. It gets depth by measuring dot spacing from the single IR camera.