Because protein folding is still a long way from being understood. There have been big improvements in recent years, but there are still big holes in our knowledge about exactly how the protein folding process works.
There are ways to predict the structure of a novel protein sequence, but all of the really reliable methods depend on determining similarity to an already known structure. It's also possible to use energy minimisation to predict the effect of a small change to an existing structure.
What doesn't work well yet (and what folding @ home is doing) is to simulate larger scale folding events in proteins. The ultimate aim is to be able to take a protein sequence and predict accurately what it's folded structure will be. We are still a long way from being able to do that (just look how much people spend on crystallography and NMR rather than using computer predictions of structure).
I'm not saying that folding@home isn't a useful and valuable piece of science (it most certainly is), but don't kid yourself about it producing immediate benefits.
In the OPs situation I think another approach which I've found useful is to attach your name to the work you do. It doesn't have to be conspicuous, but just a note at the bottom of his tutorials saying "If you have any problems with these tutorials please contact Joe Bloggs", gets your name noticed. I tend to stick my contact details on the error pages for applications I write ("You appear not to have entered your name - if you're having problems contact XXX"). This gets me an occasional call to help with a simple problem, but also lets people know who did the work.
If a shopper wants to choose between 3 wifi cards, do the Linux developers bother to tell the shopper quickly and concisely which of those will work??
Hardware compatibility is a continually shifting target so a definitive list is difficult to maintain. It has been tried before but never with much success. I know Fedora have projects trying to use the smolt project to get people to report which hardware they have works and which is problematic.
WiFi cards are especially problematic. There have been several cases of manufacturers changing the underlying hardware used on a Wifi dongle whilst keeping exactly the same model number. Suddenly you have two identically labelled pieces of hardware, one of which works under Linux and one of which doesn't, and there's no way to tell without buying it and plugging it in.
unless we're using different definitions of "mile", ADSL should reach that far just fine. 1 1/2 miles is 7920 feet. ADSL has a maximum range of 18,000 feet and a full-service range of 9000 feet. so you shoudln't have any problem with that, barring the phone lines being rusty barbed wire or something.
How about buried wires sealed in waxed paper? Amazingly that's what we get for the last mile down to our house. We're about 2 miles from an exchange and we do get ADSL, but only just. Our usual connection is about 500kb/s but we get regular dropouts where the signal strength falls too low. Not surprisingly we don't have cable, so this is our only broadband choice.
I'm sure we aren't unique in the UK. The theoretical ADSL distances and speeds you see quoted often don't pan out with the quality of infrastructure we have in some rural areas.
I think that the parent's post is probably one of the favorite myths of this site. And, as a matter of law, it is simply not true. When you "accept" GPL software, you "accept" a few very important things
I think this is somewhat misleading. EULAs as discussed here are licences an end user must accept in order to use the software. The GPL is a copyright licence which you must accept in order to redistribute the software. If you just want to use GPL licenced software then you don't have to accept any licence. The wording of the GPL is very explicit about this.
This License explicitly affirms your unlimited permission to run the unmodified Program.
It doesn't get any less restrictive than that. The restrictions only apply if you want to break copyright on the software. I doubt any EULAs on commercial software give you any rights over the copyright of the program they cover.
I'm running around 7 fedora servers performing various functions (web server, databases, svn etc.) and most of them started off with Fedora Core 1 and have done every update (not reinstall) to where they are now (f8).
I've never had a major OS problem with any of them. I lose about an hour of uptime every 6 months to run the update to the latest version (after testing it on a devel server), and I combine this with some normal housekeeping I'd be doing anyway. I find that these small incremental updates are very convenient and painless.
People go on about Fedora being cutting edge and therefore unstable - but this is only true to any significant extent for more advanced desktop features. For server functions it's been rock solid.
Werewolf and Gutsy Gibbon are code names just like Longhorn and Cairo are for Microsoft.
I hate to get picky, but these are not all equivalent. Some groups use code names during development (Gutsy Gibbon, Longhorn etc), but with Fedora the release name (eg Werewolf) doesn't appear until just before the release itself. During development it's always referred to as Fedora8. In fact they even have an open process for suggesting and voting on release names which goes on during the development of that release. The first alpha of F9 was recently released and the name of the release (Sulphur - which just beat Bathysphere!) came out after that.
Having said all of that I agree with the gist of your argument. I also can't ever remember hearing anyone refer to Fedora releases by their release name.
Perl scripts may not be portable because they use modules that are not installed everywhere.
Then simply bundle the modules with your program. Not always a universal solution, but we use it a lot to distribute perl applications around our site.
These aren't really hybrids in the traditional sense of the word. They're animal eggs which have had their nucleus removed and replaced with a human nucleus. The only animal DNA which remains is in the mitochondria. I'd suspect that these are similar enough that they wouldn't impede development.
There's no way these embryos are going to be allowed to develop since the success rate for the gestation of animal clones (Dolly the sheep and friends) is pretty low and I'd hope that no ethics board would approve that kind of study with humans.
Yes, but as has been pointed out to you, you've wandered outside of strictly software. It obviously doesn't matter to you, but consider that the argument is (roughly) equivalent to complaining that some GPLed i386 assembly code wont run on a PPC.
GPLv3 does wander out of software but this is intentional, and in my opinion at least a good thing. I believe that equating the Tivo case to making i386 code run on PPC is misleading though (and this is the crux of the matter).
The whole point of the Tivo argument was that Tivo made changes to the kernel specifically to make it work with the Tivo device. They then released those changes as required by GPLv2. However, at the same time they introduced a code signing check in the hardware which meant that even if you had the code you couldn't change it and run it on the device it was designed to work with.
I suppose the argument boils down to why you want the code to be free. If you just want to be able to see other peoples code to use in other projects then I guess GPLv2 is all you need. If however you want to be able to take a device which runs on GPLd code and be able to make changes to that code to alter the functionality of that device (even just to fix bugs), then only GPLv3 can give you that.
People make a strong argument for open source software arguing that it protects you against the people or company who wrote it going out of business and not being able to support you. If you have the code you can hire someone to fix it and go on your way. This all breaks down if the people who wrote the code have some other mechanism of stopping you from running a modified version. In that case you're no better off than with closed source.
To take this to an extreme, try imagining that Trusted Computing takes off in a big way such that all consumer computers will only run trusted code. Now imagine that getting a code signing licence is REALLY expensive so that only large corporations can afford one. Now let one of those corporations release a linux distribution. Even if they then release the source to everything they change you won't be able to make use of that on any consumer hardware because you won't have the keys required to sign the resulting binaries to allow them to run on the trusted platform.
This is (hopefully) an extreme and unlikely situation, but it's the scenario that the new clauses in GPLv3 are try to protect against.
Is viewing the content on the iPlayer considered to be receiving TV broadcasts?
Nope. The conditions which require a licence are actually fairly specific in that you only need a licence to "to watch or record TV programmes as they are being shown on TV". This was probably intended to differentiate between watching TV and watching videos, but iPlayer wouldn't be covered as your not watching the program at the time it was broadcast, and you didn't record it.
Having said that I reckon it won't take long before there is a change in the charter so that timeshifted viewing is also covered (TV licence required to own a computer with an internet connection anyone?).
Shouldn't they be upset at Sun? Why is Apple getting the flack?
Well one reason I'm a bit miffed is that Apple turned up at ISMB this year and were touting the advantages of running Java on OSX. They were very vocal about the continuing and improved support for Java in 10.6 and I guess that doesn't sit well with 10.6 coming with essentially the same version of Java as 10.5.
OK so it's more of a theme which makes the native Java widgets look platform specific, but frankly for most applications it's good enough. I have a set of Java client application which run on Linux, OSX and Windows and my users would have a heck of a job telling the difference between them and a 'real' native application.
Can you OFFICIALLY update a server (no X11, no physical access) yet?
No. This isn't officially supported yet. Of course many people do this very successfully and there are plenty of guides around detailing work-rounds for any problems.
For Fedora 9, being able to do a live upgrade is one of the targeted features. If you're interested in this then join the SIG and help out with the testing so this can become an official upgrade method.
Picture this. Red Hat and Novell are both sued. Both lose. Microsoft magnanimously steps forward and pays all the costs for Novell. Or Microsoft works out a licensing deal with IP Innovation LLC on behalf of Novell, and leaves Red Hat out to dry. Then Microsoft says, "see what happens if you don't sign our licensing deal?"
If this did happen then Novel are still screwed. Unless whatever settlement they come to allows them to pass on the rights to these patents to whoever they distribute their software to (and allows those people to do the same) then they are in breach of the patent protection parts of the GPL. This means that although the lawsuit may have gone away they could no longer distribute the software which breached it.
If this functionality is to stay in any GPL software then the patent either needs to be struck down or licensed in a universal way (at least for the next 6 months until it runs out!).
Your statement is only true IF Apple didn't sell phones in the UK (to name just one country with such a law). They do.
Not yet they don't (although they have announced that they're going to).
Also, UK law isn't as strict on this as people think. You are allowed to lock a phone to a provider whilst you are under the contract in which the phone was provided. Only after the contract has expired do you have the right to get the phone unlocked.
The interesting part is that you can buy out your contract at any point (which will probably cost you all of the monthly fees for the rest of the contract term as a lump sum). I'd be interested to see what happens if someone buys an iPhone with an O2 contract then immediately buys their way out of the contract and tells O2 they want the phone unlocked. There were various posters who were baiting the Register to try to get them do do this.
A pretty minor bug, me'thinks. There's a perfectly good accepted extension for comma-delimited files, it's called CSV.
Actually I don't think that was the really telling part of the original comment. The thing that bit me was that I had tab delimited text files that I wanted to open in Calc. To do this you have to import them as CSV files (doesn't the C stand for something???). There is no TSV option and the.txt option just opens the file in writer. I'd call this pretty unintuitive (it took me a while to figure it out!)
It can't be that hard to take account of which program was used to open a text file when deciding what to do with it. Heck, just give me an option dialog which says "Which OO.org program do you want to use to open this file" and I'd be happy.
check/tmp isn't mounted executable...that kind of thing.
What does that mean? why would/tmp be a mount. And if it's a directory it has to be executable right?
It's very common practice to put/tmp onto its own partition. This means that you can limit the amount of space it has available (so that a rogue process can't write into/tmp until / fills up) and you can also use the mount options to specify that you can't launch executables from within it. For similar reasons/var is also often separated (to stop rogue logs bringing down the whole system).
Not true. There is a gst-fluendo-mp3-2.7.rpm package on the goldmaster DVD. MP3 playback works out of the box. Eat this, american law system.
The American law system won't lose any sleep over this. Fluendo are a commercial venture who have paid the licensing fees for a number of different multimedia formats. They then give you MP3 support as a freebie in the hope that you will pay for other formats. Kind of the drug pusher mentality applied to codecs.
Slashdot Mirror
Or maybe they know that IQ is normally distributed - therefore the median and the mean would be the same :-)
add:
export DISPLAY=":0.0"
to ~/.profile
You can now start X programs from Terminal.app (as long as X11.app is running).
I remember having a similar problem trying to find where to download laTex...
Because protein folding is still a long way from being understood. There have been big improvements in recent years, but there are still big holes in our knowledge about exactly how the protein folding process works. There are ways to predict the structure of a novel protein sequence, but all of the really reliable methods depend on determining similarity to an already known structure. It's also possible to use energy minimisation to predict the effect of a small change to an existing structure. What doesn't work well yet (and what folding @ home is doing) is to simulate larger scale folding events in proteins. The ultimate aim is to be able to take a protein sequence and predict accurately what it's folded structure will be. We are still a long way from being able to do that (just look how much people spend on crystallography and NMR rather than using computer predictions of structure). I'm not saying that folding@home isn't a useful and valuable piece of science (it most certainly is), but don't kid yourself about it producing immediate benefits.
In the OPs situation I think another approach which I've found useful is to attach your name to the work you do. It doesn't have to be conspicuous, but just a note at the bottom of his tutorials saying "If you have any problems with these tutorials please contact Joe Bloggs", gets your name noticed. I tend to stick my contact details on the error pages for applications I write ("You appear not to have entered your name - if you're having problems contact XXX"). This gets me an occasional call to help with a simple problem, but also lets people know who did the work.
Hardware compatibility is a continually shifting target so a definitive list is difficult to maintain. It has been tried before but never with much success. I know Fedora have projects trying to use the smolt project to get people to report which hardware they have works and which is problematic.
WiFi cards are especially problematic. There have been several cases of manufacturers changing the underlying hardware used on a Wifi dongle whilst keeping exactly the same model number. Suddenly you have two identically labelled pieces of hardware, one of which works under Linux and one of which doesn't, and there's no way to tell without buying it and plugging it in.
How about buried wires sealed in waxed paper? Amazingly that's what we get for the last mile down to our house. We're about 2 miles from an exchange and we do get ADSL, but only just. Our usual connection is about 500kb/s but we get regular dropouts where the signal strength falls too low. Not surprisingly we don't have cable, so this is our only broadband choice.
I'm sure we aren't unique in the UK. The theoretical ADSL distances and speeds you see quoted often don't pan out with the quality of infrastructure we have in some rural areas.
I think this is somewhat misleading. EULAs as discussed here are licences an end user must accept in order to use the software. The GPL is a copyright licence which you must accept in order to redistribute the software. If you just want to use GPL licenced software then you don't have to accept any licence. The wording of the GPL is very explicit about this.
It doesn't get any less restrictive than that. The restrictions only apply if you want to break copyright on the software. I doubt any EULAs on commercial software give you any rights over the copyright of the program they cover.
Someone pointed me to this nice little script which uses the JetDirect language to change the display on most recent HP printers.
Plenty of potential for confusing coworkers...
It depends on what you need.
I'm running around 7 fedora servers performing various functions (web server, databases, svn etc.) and most of them started off with Fedora Core 1 and have done every update (not reinstall) to where they are now (f8).
I've never had a major OS problem with any of them. I lose about an hour of uptime every 6 months to run the update to the latest version (after testing it on a devel server), and I combine this with some normal housekeeping I'd be doing anyway. I find that these small incremental updates are very convenient and painless.
People go on about Fedora being cutting edge and therefore unstable - but this is only true to any significant extent for more advanced desktop features. For server functions it's been rock solid.
I hate to get picky, but these are not all equivalent. Some groups use code names during development (Gutsy Gibbon, Longhorn etc), but with Fedora the release name (eg Werewolf) doesn't appear until just before the release itself. During development it's always referred to as Fedora8. In fact they even have an open process for suggesting and voting on release names which goes on during the development of that release. The first alpha of F9 was recently released and the name of the release (Sulphur - which just beat Bathysphere!) came out after that.
Having said all of that I agree with the gist of your argument. I also can't ever remember hearing anyone refer to Fedora releases by their release name.
Then simply bundle the modules with your program. Not always a universal solution, but we use it a lot to distribute perl applications around our site.
Nope. My MacBook Pro has a Matshita UK-857 drive.
What the heck are you doing kissing his wife!?
These aren't really hybrids in the traditional sense of the word. They're animal eggs which have had their nucleus removed and replaced with a human nucleus. The only animal DNA which remains is in the mitochondria. I'd suspect that these are similar enough that they wouldn't impede development.
There's no way these embryos are going to be allowed to develop since the success rate for the gestation of animal clones (Dolly the sheep and friends) is pretty low and I'd hope that no ethics board would approve that kind of study with humans.
GPLv3 does wander out of software but this is intentional, and in my opinion at least a good thing. I believe that equating the Tivo case to making i386 code run on PPC is misleading though (and this is the crux of the matter).
The whole point of the Tivo argument was that Tivo made changes to the kernel specifically to make it work with the Tivo device. They then released those changes as required by GPLv2. However, at the same time they introduced a code signing check in the hardware which meant that even if you had the code you couldn't change it and run it on the device it was designed to work with.
I suppose the argument boils down to why you want the code to be free. If you just want to be able to see other peoples code to use in other projects then I guess GPLv2 is all you need. If however you want to be able to take a device which runs on GPLd code and be able to make changes to that code to alter the functionality of that device (even just to fix bugs), then only GPLv3 can give you that.
People make a strong argument for open source software arguing that it protects you against the people or company who wrote it going out of business and not being able to support you. If you have the code you can hire someone to fix it and go on your way. This all breaks down if the people who wrote the code have some other mechanism of stopping you from running a modified version. In that case you're no better off than with closed source.
To take this to an extreme, try imagining that Trusted Computing takes off in a big way such that all consumer computers will only run trusted code. Now imagine that getting a code signing licence is REALLY expensive so that only large corporations can afford one. Now let one of those corporations release a linux distribution. Even if they then release the source to everything they change you won't be able to make use of that on any consumer hardware because you won't have the keys required to sign the resulting binaries to allow them to run on the trusted platform.
This is (hopefully) an extreme and unlikely situation, but it's the scenario that the new clauses in GPLv3 are try to protect against.
Nope. The conditions which require a licence are actually fairly specific in that you only need a licence to "to watch or record TV programmes as they are being shown on TV". This was probably intended to differentiate between watching TV and watching videos, but iPlayer wouldn't be covered as your not watching the program at the time it was broadcast, and you didn't record it.
Having said that I reckon it won't take long before there is a change in the charter so that timeshifted viewing is also covered (TV licence required to own a computer with an internet connection anyone?).
Well one reason I'm a bit miffed is that Apple turned up at ISMB this year and were touting the advantages of running Java on OSX. They were very vocal about the continuing and improved support for Java in 10.6 and I guess that doesn't sit well with 10.6 coming with essentially the same version of Java as 10.5.
Kind of like...
OK so it's more of a theme which makes the native Java widgets look platform specific, but frankly for most applications it's good enough. I have a set of Java client application which run on Linux, OSX and Windows and my users would have a heck of a job telling the difference between them and a 'real' native application.
No. This isn't officially supported yet. Of course many people do this very successfully and there are plenty of guides around detailing work-rounds for any problems.
For Fedora 9, being able to do a live upgrade is one of the targeted features. If you're interested in this then join the SIG and help out with the testing so this can become an official upgrade method.
If this did happen then Novel are still screwed. Unless whatever settlement they come to allows them to pass on the rights to these patents to whoever they distribute their software to (and allows those people to do the same) then they are in breach of the patent protection parts of the GPL. This means that although the lawsuit may have gone away they could no longer distribute the software which breached it.
If this functionality is to stay in any GPL software then the patent either needs to be struck down or licensed in a universal way (at least for the next 6 months until it runs out!).
Not yet they don't (although they have announced that they're going to).
Also, UK law isn't as strict on this as people think. You are allowed to lock a phone to a provider whilst you are under the contract in which the phone was provided. Only after the contract has expired do you have the right to get the phone unlocked.
The interesting part is that you can buy out your contract at any point (which will probably cost you all of the monthly fees for the rest of the contract term as a lump sum). I'd be interested to see what happens if someone buys an iPhone with an O2 contract then immediately buys their way out of the contract and tells O2 they want the phone unlocked. There were various posters who were baiting the Register to try to get them do do this.
Actually I don't think that was the really telling part of the original comment. The thing that bit me was that I had tab delimited text files that I wanted to open in Calc. To do this you have to import them as CSV files (doesn't the C stand for something???). There is no TSV option and the .txt option just opens the file in writer. I'd call this pretty unintuitive (it took me a while to figure it out!)
It can't be that hard to take account of which program was used to open a text file when deciding what to do with it. Heck, just give me an option dialog which says "Which OO.org program do you want to use to open this file" and I'd be happy.
It's very common practice to put /tmp onto its own partition. This means that you can limit the amount of space it has available (so that a rogue process can't write into /tmp until / fills up) and you can also use the mount options to specify that you can't launch executables from within it. For similar reasons /var is also often separated (to stop rogue logs bringing down the whole system).
The American law system won't lose any sleep over this. Fluendo are a commercial venture who have paid the licensing fees for a number of different multimedia formats. They then give you MP3 support as a freebie in the hope that you will pay for other formats. Kind of the drug pusher mentality applied to codecs.