That was kind of my point, in some ways. People were arguing towards porting it to C, but I think it would be fair to say that the actual real-world benefits would be fairly minimal (desktop apps – the things that would use the framework – really do not benefit from the efficiency of C, and suffer from its shortcomings). Since Qt4 is considered fairly stable, there'd be little benefit, although the benefit of using C#/Java when developing it as new would be to reduce development time, and the potential for memory bugs and problems.
The main point, though, was that porting to C serves little purpose.
I agree. I use GNOME on my Linux machines, and whenever I have to use a KDE-based app I feel as though I'm somehow missing out, like when I'd run GTK-based apps on my KDE (on BSD) install in the past. The community is clearly divided, and although this is, as much as anything, an advantage of the OSS way of doing things, the resources have been split up in a fairly weakening way. Nobody wants to introduce "better integration" of the other toolkit because it's both a concession to the "other side", and a very technically-difficult task. It's a pretty strange situation in general.
I don't think there'd be much point in switching to C at this point, they'd only have to rig an object model on top of it. They'd probably be better off switching to C# or Java, something which would actually bring tangible benefits.
Surely you're looking for XFCE? I'm not convinced that making the software more "lightweight" is a good argument, that's clearly not what they're aiming for. Although if there's actual structural problems, or bugs, causing the OTT memory usage, yes, those should be dealt with.
I believe their main argument was that OpenSSH was just an SSH system, whereas they had integrated tools for upgrading and so on the clients/servers on loads of different systems (running different OSs) at once, rather than any serious security concern. Obviously package managers destroy a lot of their argument, but integrating everything when you're using multiple platforms has at least some validity.
Bravo. The point here is that the applications will be built on top of the technology, not the other way around. The bandwidth isn't used because it isn't commonplace — I'm sure people will think of plenty of legitimate uses once it is.
I finally finished reading the article (damn slashdot effect) and I have to agree. It seems like the guy is spending two pages whining about all the little things he doesn't like about the distro and then gives it 9 of 10 stars? I don't get it either.
It's like how computer games had to become very close to photorealistic before you noticed that all the characters' eyes were pointing in silly directions — many minor flaws become apparent when the system itself gets to a certain level of quality. The article was quite misleading, though, yes, a little weighted towards the negative considering that almost every concern was trivial.
Maybe the next version will ship with a sane list of enabled sources in/etc/apt/sources.list so I won't have to fiddle with Synaptic just to get vim-gtk and whatnot installed.
I don't think they have plans to change the default-enabled repositories, but Breezy comes with an "Add Applications" frontend to Synaptic (and by association APT) which gives a tree view of apps you can install (mirroring the structure of the menu). This includes apps which require other repositories (they are greyed out), and help is given (along with there being a not-as-simple-as-it-should-be-but-quite-simple "Repositories" menu) to add the repository parts that you need to install the other apps.
It's not perfect, but it brings things forward a long way, particularly for novice users.
As a bit of contrary data, I'm yet to have a problem with Breezy, but the development version of Hoary gave me no end of pain (I remember them breaking the dependencies for gnome-panel, which was quite horrifying).
Actually, I tell a lie — the new kernel doesn't work on my laptop. But then, the new kernel doesn't work on my laptop on any distro (without adding the mysterious "irqpoll" option, which makes it work), so I'm loathe to blame that on Ubuntu.
I investigated Ubuntu on the recommendations of Planet Gnome [gnome.org] when it came out. I had problems with Ubuntu detecting my video cards, on 2 out of my 3 computers. Not sure why. It just gets to X, and then reboots. (Lather, rinse, repeat.) I lost interest.
If it's useful for (your own personal) statistical purposes, I've not had a problem on any of my 3 computers:).
I'm not sure if it's a better Linux distro. (I think Knoppix CDs are cool, because they've worked on all of my computers.) But they get points in my book for being friendly and marketing themselves well, like Firefox. Something kind of overlooked in general.
That's the point, really. People keep saying "Linux! On the desktop! It should be now, but the users are too stupid!". That's just nonsense — if the users you want to adopt can't use it, your system is too hard to use. It really is that simple. Ubuntu is making good headway towards actually being usable (Breezy's "Add Applications" menu option is a great example), and should be applauded for that.
The question mark should be placed inside the quotation marks.
That is a rule of english that has always annoyed me. If it isn't the quoted item which contains the question, I don't put the question mark inside it (same thing with full stops and quoted sentences). Regardless of correctness, I think it's clearer.
Vista's proposed feature WinFS was a much better way of organising data, and it's still in development. Apparently the version that gets released is going to have support for labels ("virtual folders"), but not some of the other features that're coming with WinFS.
Is it possibly symbolic links (on *nix) you're thinking of? Still kinda a way around a flaw, rather than a proper solution. Labels are more elegant, in general.
A hierarchical label system is clearly possible (just impose a structure on them), but in a lot of cases where it might seem like an advantage, it's not really. A hierarchical label structure is actually no different than a non-hierarchical one other than its presentation, I suppose. It's not (as I mentioned or meant to mention) possible with GMail's implementation, but I'm not convinced it's useful for most people.
You can. Messages which skip the inbox are still marked as unread, and still highlight as unread in the list of filters. There is absolutely nothing you can do with folders that you can't do with labels* (they are essentially a less powerful version of the same thing), and there is things you can do with labels that you can't (semantically) do with folders (multiple membership, for instance).
Folders are a clumsy metaphor, I find. The same forces that are moving filesystems (or, in particular, their representation to the user) away from the "directories" approach are making this change. There's no reason to have files which can only be in one "folder" — it's an artificially limited metaphor.
* Although obviously you can't, in Google's implementation, have folders-in-folders or the like, but this isn't really a problem with the metaphor (nor a feature that 95% of people would find useful).
To be fair, the problem is that one company controlling an entire market is never a good thing, even if it is Google.
My personal thought is that systems like this are beneficial to society as a whole, and so long as Google don't use the technology in the way you describe, they should be allowed, whether or not that's what the law says. Copyright law needs radical reform.
In this country we have to enter our PIN to purchase (used to be signature) so it's a little less bad, but yeah, convenience can take precedence over safety sometimes, and that's rarely good.
In this case the party responsible is either the credit card company, for accepting such meagre proof of identity. You should be able to get any money stolen back, and the credit card companies would soon change their policies if this was a valid means of attack.
I suspect it's too traceable to be cost-effective. I'm not convinced you'd get enough return on the cost of your IDs and the machine itself before they'd shut the system down, and even if you did you could probably never do the same thing more than one or two times before your description was well-known. The companies who commit fraud online are generally based outside the reach of the law in the countries of the people they trick, or are not tracable. In addition, you need a lot less information to get money from someone's card online than through one of these systems.
That's not to say that a little extra security isn't warranted, though. The standard they are built on apparently has support for embedded biometric data, which could sort it all out were it ready for the market:). I'd certainly be hesitant about using a card one "just waved", if only in case I lost it.
The Mozilla ones weren't patched at all. Users had to download a replacement browser (at least for Firefox). So there wasn't a patch at all. Just a new browser.
Isn't that just a convenient way of shipping a source patch?:)
Slashdot Mirror
Thanks for that, you saved me a bit of explanation. Now I better get back to work.
That was kind of my point, in some ways. People were arguing towards porting it to C, but I think it would be fair to say that the actual real-world benefits would be fairly minimal (desktop apps – the things that would use the framework – really do not benefit from the efficiency of C, and suffer from its shortcomings). Since Qt4 is considered fairly stable, there'd be little benefit, although the benefit of using C#/Java when developing it as new would be to reduce development time, and the potential for memory bugs and problems.
The main point, though, was that porting to C serves little purpose.
I agree. I use GNOME on my Linux machines, and whenever I have to use a KDE-based app I feel as though I'm somehow missing out, like when I'd run GTK-based apps on my KDE (on BSD) install in the past. The community is clearly divided, and although this is, as much as anything, an advantage of the OSS way of doing things, the resources have been split up in a fairly weakening way. Nobody wants to introduce "better integration" of the other toolkit because it's both a concession to the "other side", and a very technically-difficult task. It's a pretty strange situation in general.
I'm aware of this. Did something I say seem to contradict that?
I don't think there'd be much point in switching to C at this point, they'd only have to rig an object model on top of it. They'd probably be better off switching to C# or Java, something which would actually bring tangible benefits.
Surely you're looking for XFCE? I'm not convinced that making the software more "lightweight" is a good argument, that's clearly not what they're aiming for. Although if there's actual structural problems, or bugs, causing the OTT memory usage, yes, those should be dealt with.
I believe their main argument was that OpenSSH was just an SSH system, whereas they had integrated tools for upgrading and so on the clients/servers on loads of different systems (running different OSs) at once, rather than any serious security concern. Obviously package managers destroy a lot of their argument, but integrating everything when you're using multiple platforms has at least some validity.
Bravo. The point here is that the applications will be built on top of the technology, not the other way around. The bandwidth isn't used because it isn't commonplace — I'm sure people will think of plenty of legitimate uses once it is.
There's a Beagle 2 joke here somewhere, but I can't place it.
One simple, good reason is dyslexia.
It's like how computer games had to become very close to photorealistic before you noticed that all the characters' eyes were pointing in silly directions — many minor flaws become apparent when the system itself gets to a certain level of quality. The article was quite misleading, though, yes, a little weighted towards the negative considering that almost every concern was trivial.
I don't think they have plans to change the default-enabled repositories, but Breezy comes with an "Add Applications" frontend to Synaptic (and by association APT) which gives a tree view of apps you can install (mirroring the structure of the menu). This includes apps which require other repositories (they are greyed out), and help is given (along with there being a not-as-simple-as-it-should-be-but-quite-simple "Repositories" menu) to add the repository parts that you need to install the other apps.
It's not perfect, but it brings things forward a long way, particularly for novice users.
As a bit of contrary data, I'm yet to have a problem with Breezy, but the development version of Hoary gave me no end of pain (I remember them breaking the dependencies for gnome-panel, which was quite horrifying).
Actually, I tell a lie — the new kernel doesn't work on my laptop. But then, the new kernel doesn't work on my laptop on any distro (without adding the mysterious "irqpoll" option, which makes it work), so I'm loathe to blame that on Ubuntu.
If it's useful for (your own personal) statistical purposes, I've not had a problem on any of my 3 computers :).
That's the point, really. People keep saying "Linux! On the desktop! It should be now, but the users are too stupid!". That's just nonsense — if the users you want to adopt can't use it, your system is too hard to use. It really is that simple. Ubuntu is making good headway towards actually being usable (Breezy's "Add Applications" menu option is a great example), and should be applauded for that.
Since you were in my English class, I think you know that we were, in fact, taught nothing, Neil. :P
Hmm. I'm Scottish, and semi-remember the rule being taught, but it's possible that I just don't recall correctly :).
That is a rule of english that has always annoyed me. If it isn't the quoted item which contains the question, I don't put the question mark inside it (same thing with full stops and quoted sentences). Regardless of correctness, I think it's clearer.
Vista's proposed feature WinFS was a much better way of organising data, and it's still in development. Apparently the version that gets released is going to have support for labels ("virtual folders"), but not some of the other features that're coming with WinFS.
Is it possibly symbolic links (on *nix) you're thinking of? Still kinda a way around a flaw, rather than a proper solution. Labels are more elegant, in general.
A hierarchical label system is clearly possible (just impose a structure on them), but in a lot of cases where it might seem like an advantage, it's not really. A hierarchical label structure is actually no different than a non-hierarchical one other than its presentation, I suppose. It's not (as I mentioned or meant to mention) possible with GMail's implementation, but I'm not convinced it's useful for most people.
You can. Messages which skip the inbox are still marked as unread, and still highlight as unread in the list of filters. There is absolutely nothing you can do with folders that you can't do with labels* (they are essentially a less powerful version of the same thing), and there is things you can do with labels that you can't (semantically) do with folders (multiple membership, for instance).
Folders are a clumsy metaphor, I find. The same forces that are moving filesystems (or, in particular, their representation to the user) away from the "directories" approach are making this change. There's no reason to have files which can only be in one "folder" — it's an artificially limited metaphor.
* Although obviously you can't, in Google's implementation, have folders-in-folders or the like, but this isn't really a problem with the metaphor (nor a feature that 95% of people would find useful).
The goggles, they do nothing!
To be fair, the problem is that one company controlling an entire market is never a good thing, even if it is Google.
My personal thought is that systems like this are beneficial to society as a whole, and so long as Google don't use the technology in the way you describe, they should be allowed, whether or not that's what the law says. Copyright law needs radical reform.
In this country we have to enter our PIN to purchase (used to be signature) so it's a little less bad, but yeah, convenience can take precedence over safety sometimes, and that's rarely good.
In this case the party responsible is either the credit card company, for accepting such meagre proof of identity. You should be able to get any money stolen back, and the credit card companies would soon change their policies if this was a valid means of attack.
I suspect it's too traceable to be cost-effective. I'm not convinced you'd get enough return on the cost of your IDs and the machine itself before they'd shut the system down, and even if you did you could probably never do the same thing more than one or two times before your description was well-known. The companies who commit fraud online are generally based outside the reach of the law in the countries of the people they trick, or are not tracable. In addition, you need a lot less information to get money from someone's card online than through one of these systems.
That's not to say that a little extra security isn't warranted, though. The standard they are built on apparently has support for embedded biometric data, which could sort it all out were it ready for the market :). I'd certainly be hesitant about using a card one "just waved", if only in case I lost it.
Isn't that just a convenient way of shipping a source patch? :)