Read the article, that's not what this does. This is a project to automatically generate optimising compilers for custom architectures. The summary is a little unclear:-(
It reduces time to market because you don't have to spend ages making an optimising compiler for your custom chip.
Yes folks, those lovely roads you cycle on are paid from the taxation of motorists
No they are not, please check your figures. The money raised by fuel duty, road tax and VAT on vehicles does NOT cover the cost of the UK road network. It has to be subsidised by general taxation.
Cyclists are (usually, heh) tax payers and have as much right to use the road as you do.
Mono is just like Java: it's a compiled, statically-typed, managed language. So you get near to the speed of C/C++, but memory errors are mostly impossible. The downside (compared to C) is typically a doubling in memory requirements, though it depends a lot on the program.
Eliminating memory errors produces a surprisingly large increase in productivity. The first Java program I wrote was about 5,000 lines of code, and once I got it through the compiler, it pretty much worked first time, except for some tuning. A very different experience from my C programs, heh.
Having said that, I still prefer a mix of C/C++/Python myself.
I think you're missing the key point of a repository: it's a (large) set of software packages that are tested to work together. If you let devs update their packages willy-nilly, you're going to get horrible breakage very quickly.
Apple's app-store works because there are almost never (as far as I know) dependencies between apps. Updating an app might break the user's savefiles I guess, but it won't stop another app working. To make an analogy: an app store which devs can update can only ever contain leaves, you can't put any twigs/branches/trunk in there.
I suppose you can imagine an app store built on top of apt. An extra repository which is guaranteed to only contain packages upon which nothing else depends, and which has much looser restrictions on updates.
Re:Windows has more and more Unix features
on
Unix Turns 40
·
· Score: 1
In fact X pre-dates Windows, by a small amount at least. Windows 1.0 launched in November 1985, X version 6 (the first public one, I think) was released in January 1985. X has had hardware acceleration since the beginning, Windows picked that up much later.
I just missed X10. My first workstation ran X11, a tricky 10 MB source code download by acoustic modem (!!!) as I recall.
Two issues are being confused there. First, do you use a cross-platform toolkit, or do you write a true native GUI for every platform and just keep the backend in common? Google have decided to write a new GUI for every platform, and I think they are probably correct to do this. Qt (and GTK+) are cross-platform, but they are not quite native (though arguably Qt is better at this).
Once that choice is made, all you are doing is picking a toolkit for Linux. GTK+ has the advantages of being familiar to the chrome devs, matching the existing ff dependency, being the most widely-used toolkit (and therefore appearing native for the largest number of users), and being "good enough".
A power factor of 0.5 does NOT mean that the device uses twice the power (where power means joules per second). It means it knocks the supply out of phase slightly and as a result, the grid has to use capacitors to correct the phase, or generate more VA (but not more power).
Agreed, what a good post. I've worked on many EU-funded projects and the admin is really not too bad.
If you want to criticise EU research projects there are much more obvious fish to shoot at in that barrel. Six-monthly (or often more frequent) project meetings of international partners are dazzlingly expensive and wasteful, and many of your partners will be absolutely useless, but you have to have them to meet various quotas.
In one project I was in, the Greek partner received his initial payment, ran to the bank, cashed the cheque, and fled to Brazil. This was comically described as a "minor project startup issue" by the lead. Fortunately no one had counted on them doing anything useful anyway, so the project was unaffected.
Despite the problems I still have a positive view of these programmes. They really do encourage EU-wide collaboration; they sometimes, even indirectly, produce useful results, and are actually much less wasteful than many things about the EU.
MS have said this is for netbooks in developed markets.
Brad Brooks, corporate VP for Windows Consumer Product Marketing said "These engineering investments allow small notebook PCs to run any version of Windows 7, and allow customers complete flexibility to purchase a system which meets their needs. For OEMs that build lower-cost small notebook PCs, Windows 7 Starter will now be available in developed markets," he added. Windows 7 Starter is a light-weight version of the OS without the fancy interface, and is also limited to having three applications open at once (perhaps in an effort to prevent the system from getting bogged down).
I don't know why you picked that one out. It will add a feature that visibly improves the quality of all image shrinks, past what Photoshop can do out of the box. It's a really useful, basic improvement.
By default, Vista users aren't running as root and the only way to become root is either a UAC dialog or a privilege escalation exploit.
That's not quite accurate. UAC is not a security barrier in the way that sudo is, and there are a great many unfixable (thanks for history) privilege escalations.
Some time later, when someone starts an elevated command prompt, you'll see an elevated regedit pop up too. UAC provides almost no protection against privilege escalation and it's not intended to.
That does sound slow. On my Ubuntu 8.10 machine (2.7GHz Opteron) FileChooser shows/usr/bin (2,500 files) for the first time in about 4s. This time is mostly (I think) to load the icons for all the files. Subsequent visits are very quick (under a second?). Logging out and in again makes it slower for the first visit again.
A lot of work went into filechooser performance, mostly by the wonderful Federico. He has an interesting technical blog here:
If we're comparing default installs, then vista sp1 is a lot less secure than ubuntu. The default login on vista is an administrator and UAC offers very little protection. There are multiple, trivial, unfixable ways to get around it due to Window's legacy. As MS themselves say, UAC is NOT a security barrier.
Ubuntu on the other hand has an unprivileged default user who has to sudo to do anything, with a proper security barrier between the two states. It's also had ASLR, stack protection, heap protection and various other features on by default for a while.
Do not make the user press an OK or Apply button to make the changes happen, unless either:
the change will take more than about one second to apply, in which case applying the change immediately could make the system feel slow or unresponsive, or
the changes in the window have to be applied simultaneously to prevent the system entering a potentially unstable state. For example, the hostname and proxy fields in a network properties window.
No, this is using an application "registry" already. The debate is about how the registry keeps its data on disc.
It currently stores each set of name/value pairs as a chunk of XML in a hierarchy of files. Updating this large set of files and being safe in case of crashes is difficult.
Take any linux binary compiled 10 years ago and run it today on a shipping kernel. Oh wait... you can't. Do the same under Macintosh. Oh wait... can't.
Of course you can, don't be silly.
And why do you dismiss source-code compatibility so quickly? It's great to not be tied to a supplier. It's fantastic to be able to move your stuff to another platform easily.
MS are dominant in business principally because of Office and the ecosystem that's built up around it.
My firefox has been running for 8 days, has adblock, flashblock and firebug running, and it using 230 MB of RSS at the moment.
Have you tried removing all your extensions and seeing if it improves?
I expect you saw this article last year:
They are measuring ice thickness, not extent. You can get a rough idea of thickness from satellite radar, but this should give much more accurate numbers (though only for a relatively small part of the ice, of course).
Sure CO2 only makes a small difference, a few percent. But a few percent change in the atmosphere's warming effect is a degree C. Exactly what the IPCC are warning about.
Sadly macports and fink are pretty poor:( They don't have enough people and most of the packages are broken or out of date. I have simple patches for projects which I run which have been sitting in the macports tracker for more than six months and still have not been approved.
Debian/Ubuntu/etc. still have by far the best package repository and that's enough to make my mac almost useless and my linux laptop the place where I do most of my work. Plus OS X is rather slow, argh.
True, but not the same. Those are chunks of the core OS, they are not common desktop programs. On the current win7 beta, by default calc.exe (!!!!!) effectively runs elevated.
It's good they've responded, but this change does not fix the fundamental problems with win7's UAC whitelist.
The problem is that 70 applications are on the whitelist and are allowed to silently elevate without the user's knowledge. You just have to inject code into one of these 70 applications and you have admin rights. There are multiple ways of doing this. You can use the debug API, you can get them to load a DLL, use your imagination.
Here's a page with a sample exploit and a lot more information:
Slashdot Mirror
Read the article, that's not what this does. This is a project to automatically generate optimising compilers for custom architectures. The summary is a little unclear :-(
It reduces time to market because you don't have to spend ages making an optimising compiler for your custom chip.
Yes folks, those lovely roads you cycle on are paid from the taxation of motorists
No they are not, please check your figures. The money raised by fuel duty, road tax and VAT on vehicles does NOT cover the cost of the UK road network. It has to be subsidised by general taxation.
Cyclists are (usually, heh) tax payers and have as much right to use the road as you do.
Mono is just like Java: it's a compiled, statically-typed, managed language. So you get near to the speed of C/C++, but memory errors are mostly impossible. The downside (compared to C) is typically a doubling in memory requirements, though it depends a lot on the program.
Eliminating memory errors produces a surprisingly large increase in productivity. The first Java program I wrote was about 5,000 lines of code, and once I got it through the compiler, it pretty much worked first time, except for some tuning. A very different experience from my C programs, heh.
Having said that, I still prefer a mix of C/C++/Python myself.
You can buy an unlocked phone from most of the manufactures now. Just get a SIM card for the network you want to use and pop it in.
For example, the N85 is $399 unlocked on the Nokia store.
I think you're missing the key point of a repository: it's a (large) set of software packages that are tested to work together. If you let devs update their packages willy-nilly, you're going to get horrible breakage very quickly.
Apple's app-store works because there are almost never (as far as I know) dependencies between apps. Updating an app might break the user's savefiles I guess, but it won't stop another app working. To make an analogy: an app store which devs can update can only ever contain leaves, you can't put any twigs/branches/trunk in there.
I suppose you can imagine an app store built on top of apt. An extra repository which is guaranteed to only contain packages upon which nothing else depends, and which has much looser restrictions on updates.
In fact X pre-dates Windows, by a small amount at least. Windows 1.0 launched in November 1985, X version 6 (the first public one, I think) was released in January 1985. X has had hardware acceleration since the beginning, Windows picked that up much later.
I just missed X10. My first workstation ran X11, a tricky 10 MB source code download by acoustic modem (!!!) as I recall.
Two issues are being confused there. First, do you use a cross-platform toolkit, or do you write a true native GUI for every platform and just keep the backend in common? Google have decided to write a new GUI for every platform, and I think they are probably correct to do this. Qt (and GTK+) are cross-platform, but they are not quite native (though arguably Qt is better at this).
Once that choice is made, all you are doing is picking a toolkit for Linux. GTK+ has the advantages of being familiar to the chrome devs, matching the existing ff dependency, being the most widely-used toolkit (and therefore appearing native for the largest number of users), and being "good enough".
You're wrong about power factor.
A power factor of 0.5 does NOT mean that the device uses twice the power (where power means joules per second). It means it knocks the supply out of phase slightly and as a result, the grid has to use capacitors to correct the phase, or generate more VA (but not more power).
There's a good discussion of it here:
http://www.homepower.com/article/?file=HP96_pg128_Letters_1
Windows is only free if your time is worth nothing.
Agreed, what a good post. I've worked on many EU-funded projects and the admin is really not too bad.
If you want to criticise EU research projects there are much more obvious fish to shoot at in that barrel. Six-monthly (or often more frequent) project meetings of international partners are dazzlingly expensive and wasteful, and many of your partners will be absolutely useless, but you have to have them to meet various quotas.
In one project I was in, the Greek partner received his initial payment, ran to the bank, cashed the cheque, and fled to Brazil. This was comically described as a "minor project startup issue" by the lead. Fortunately no one had counted on them doing anything useful anyway, so the project was unaffected.
Despite the problems I still have a positive view of these programmes. They really do encourage EU-wide collaboration; they sometimes, even indirectly, produce useful results, and are actually much less wasteful than many things about the EU.
MS have said this is for netbooks in developed markets.
Brad Brooks, corporate VP for Windows Consumer Product Marketing said "These engineering investments allow small notebook PCs to run any version of Windows 7, and allow customers complete flexibility to purchase a system which meets their needs. For OEMs that build lower-cost small notebook PCs, Windows 7 Starter will now be available in developed markets," he added. Windows 7 Starter is a light-weight version of the OS without the fancy interface, and is also limited to having three applications open at once (perhaps in an effort to prevent the system from getting bogged down).
http://www.tomshardware.com/news/windows-7-netbooks-xp-vista,6974.html
I don't know why you picked that one out. It will add a feature that visibly improves the quality of all image shrinks, past what Photoshop can do out of the box. It's a really useful, basic improvement.
Read about it here if you're curious:
http://wiki.gimp.org/gimp/SummerOfCode2009ideas#head-ee0a4959625baa7bff3da72ec494b0f5f10859dd
That's not quite accurate. UAC is not a security barrier in the way that sudo is, and there are a great many unfixable (thanks for history) privilege escalations.
For example, as a regular user, try entering:
reg add "HKCU\Software\Microsoft\Command Processor" /v AutoRun /d "regedit" /f
Some time later, when someone starts an elevated command prompt, you'll see an elevated regedit pop up too. UAC provides almost no protection against privilege escalation and it's not intended to.
That does sound slow. On my Ubuntu 8.10 machine (2.7GHz Opteron) FileChooser shows /usr/bin (2,500 files) for the first time in about 4s. This time is mostly (I think) to load the icons for all the files. Subsequent visits are very quick (under a second?). Logging out and in again makes it slower for the first visit again.
A lot of work went into filechooser performance, mostly by the wonderful Federico. He has an interesting technical blog here:
http://www.gnome.org/~federico/news.html
If we're comparing default installs, then vista sp1 is a lot less secure than ubuntu. The default login on vista is an administrator and UAC offers very little protection. There are multiple, trivial, unfixable ways to get around it due to Window's legacy. As MS themselves say, UAC is NOT a security barrier.
Ubuntu on the other hand has an unprivileged default user who has to sudo to do anything, with a proper security barrier between the two states. It's also had ASLR, stack protection, heap protection and various other features on by default for a while.
Instant-apply is only used for quick, non-destructive options. You can read the guidelines here:
http://library.gnome.org/devel/hig-book/stable/windows-utility.html.en
It has nice scripting and batch capabilities, if you're dorky enough to like script-fu or Python. Ars had a nice review of 2.6.4 recently:
http://arstechnica.com/media/news/2009/01/gimp-2-6-review.ars
They compare it to photoshop and list a few areas where they think it is ahead. The last page has a set of good/bad/ugly bullet points.
No, this is using an application "registry" already. The debate is about how the registry keeps its data on disc.
It currently stores each set of name/value pairs as a chunk of XML in a hierarchy of files. Updating this large set of files and being safe in case of crashes is difficult.
Of course you can, don't be silly.
And why do you dismiss source-code compatibility so quickly? It's great to not be tied to a supplier. It's fantastic to be able to move your stuff to another platform easily.
MS are dominant in business principally because of Office and the ecosystem that's built up around it.
My firefox has been running for 8 days, has adblock, flashblock and firebug running, and it using 230 MB of RSS at the moment. Have you tried removing all your extensions and seeing if it improves? I expect you saw this article last year:
http://arstechnica.com/open-source/news/2008/03/firefox-3-goes-on-a-diet-eats-less-memory-than-ie-and-opera.ars
They are measuring ice thickness, not extent. You can get a rough idea of thickness from satellite radar, but this should give much more accurate numbers (though only for a relatively small part of the ice, of course).
Sure CO2 only makes a small difference, a few percent. But a few percent change in the atmosphere's warming effect is a degree C. Exactly what the IPCC are warning about.
You can read up on some of the science here:
http://www.realclimate.org/index.php/archives/2005/04/water-vapour-feedback-or-forcing
It's not difficult and you might find it interesting.
Sadly macports and fink are pretty poor :( They don't have enough people and most of the packages are broken or out of date. I have simple patches for projects which I run which have been sitting in the macports tracker for more than six months and still have not been approved.
Debian/Ubuntu/etc. still have by far the best package repository and that's enough to make my mac almost useless and my linux laptop the place where I do most of my work. Plus OS X is rather slow, argh.
True, but not the same. Those are chunks of the core OS, they are not common desktop programs. On the current win7 beta, by default calc.exe (!!!!!) effectively runs elevated.
It's good they've responded, but this change does not fix the fundamental problems with win7's UAC whitelist.
The problem is that 70 applications are on the whitelist and are allowed to silently elevate without the user's knowledge. You just have to inject code into one of these 70 applications and you have admin rights. There are multiple ways of doing this. You can use the debug API, you can get them to load a DLL, use your imagination.
Here's a page with a sample exploit and a lot more information:
http://www.pretentiousname.com/misc/win7_uac_whitelist2.html