GNUstep started to implement the OpenStep specification, which was a public spec for portable application development and was implemented by NeXT and Sun (hence the NS prefix on all of the class names). The most popular implementation of OpenStep is called Cocoa, and includes a lot of extensions to the base spec. We try to implement these extensions as well.
The only concern I have is they do seem to be looking at it more as a framework for porting, which is the least important use from my perspective. This is the tool to build the better desktop on linux everyone claims to want.
GNUstep aims to implement the APIs that Cocoa uses. This has a natural use as a porting tool, but the main reason we're implementing the APIs is that we want to use them (which has the unfortunate side effect that ones we don't like tend not to be implemented quickly, even if lots of OS X code uses them). Over in Étoilé (which, no doubt, Slashdot's early-'90s character encoding support will mangle: Etoile with accents on both 'e's) we're building frameworks for building better environments, some of which also work on OS X and some of which don't.
No. If there were a market for a Linux build of Photoshop, then Adobe would find it much easier to port the Windows version with WINE than the OS X version with GNUstep (and I say this as a GNUstep developer). Applications like OmniGraffle, however, would be easier to port. I think we already implement most of what OmniGraffle needs, but there are lots of missing bits of APIs. I have a Summer of Code student who is working on getting the CoreAnimation / CoreGraphics stuff integrated (our current GUI code uses the NeXT DisplayPostscript APIs) which should help with a lot of things.
And on active sites it looks like IIS has been dropping in use since April 2009. Apache has actually picked up a bit since they had a big drop between 2005 and 2007, although it's down a bit since the small peak in 2011. The most interesting part is the growth of the 'other' line from about 6% to about 12% over the past 3 years. It's sad to see Lighttpd has almost died (not much development for years, now just counted as part of 'other'), but it's good to see a few different servers clustered around the 5-10% mark. Having a single server with over 50% of the market makes it quite an attractive target. It would be good to see more diversity.
I don't know the show, so I've neither pirated it nor watched it legally, but add to that:
The pirated version is a download that you can watch on any device any time. The Netflix stream requires Silverlight, so I can't use it on the FreeBSD box connected to my projector nor on one of my tablets. The other tablet runs Android, so there is a Netflix streaming app, but I don't think it lets me download things and I mostly want to watch things on the tablet when I'm on a train or plane (spotty / expensive / unavailable Internet access).
In rural parts of the UK (e.g. where my mother lives), the ADSL connection isn't fast enough for streaming, but it's fine if you start downloading 10-15 minutes before you start watching, so again the pirated version wins because you can just download it and then watch it later.
Give me a service that lets me download DRM-free videos with some reasonable per-month, download-capped pricing, and I'll very happily subscribe (and, no, I'm not moving the goalposts - this is what I've been asking for for the last 10 years). Something like 30-45 hours for £10-15 would be fine. Until then, I'll keep getting the shiny disks through the post.
The difference is, the person hosting the encrypted mail and the person providing the code are usually different in non-Web cases. If a mail client is regularly making connections to something other than your web server, that's something an IDS is likely to spot. It's even something a corporate firewall may spot, if it isn't doing HTTP requests that can go via a corporate proxy. If your web page is making more requests than strictly necessary to the originating web server, that's a lot harder to spot.
The existence of asynchronous HTTP requests from JavaScript makes that a somewhat meaningless level of protection. The server can't decrypt the message, but you can only decrypt the message by running code that you get from the server, which runs in a context where it is allowed to send arbitrary messages back to the server. Unless you're running a JavaScript debugger and ensuring that the decryption code isn't sending the decrypted version (optionally re-encrypted with the server's key) back to the server, you aren't getting any more security than having the server decrypt it.
TorrentFreak (which, as a Virgin Media customer, I can apparently still access without jumping through hoops) appears to be more of a blog / news site these days. From the front page, it's not even obvious that they link to illegal torrents (do they?). They do list this in their about page:
TorrentFreak was featured on mainstream news outlets such as CNN, The Wall Street Journal, New York Times the BBC, the Guardian and the LA Times.
150,000+ RSS subscribers
Top 50 Techmeme leaderboard
Top 100 blogs on Technorati
Which makes it seem like they are not a sensible thing to block. I've not visited the site for a great many years, but if you only object to the sites that you use being blocked then it's very easy for censors to creep in.
Not necessarily. We know that, at some point in the future, the TARDIS becomes The Doctor's grave marker, but we don't know that it was placed there when he died. His body may have remained in The TARDIS for some hundreds of years of subjective time first. Having him preserved in his near-death state in The TARDIS with the ability to provide advice to Susan occasionally might also work nicely as a plot device.
An Internet Security suite from a company that no longer has its root certificate in my trusted list because of their inability to secure their own systems? Why on earth would I want something like that?
the pilots about 20mins out from landing saying, we can't talk to you anymore until after we land.
And they almost certainly could have landed the plane fine if you'd kept chatting. In almost every case, it would be completely fine. But very occasionally, they'd miss checking a dial or mishear ATC instructions and end up with a plane full of dead passengers, and they don't want to take that risk because, unlike many other people in this thread, they were behaving like responsible adults.
In The Five Doctors, The Master is offered another cycle of regenerations in exchange for his help. It's not clear whether he's given them, but the fact that he accepts indicates that he at least believes that this is within the High Council's power to grant. It's also implied in The Brain of Morbius that the limit on regenerations is arbitrary and imposed so that Time Lord society won't suffer the same kind of stagnation as the Sisterhood of Karn.
John Hurt is The Valeyard, who was 'between the 12th and final' incarnations, so a future one not a past one. The Time Lords in Trial of a Time Lord did say 'final' and not 13th though, but it's entirely possible that the Great Time War messed up The Doctor's timeline such that the 13th was not actually his last, or that The Matrix was simply unaware of the later ones as they were unauthorised.
They aren't changing their reproductive system. The Time Lords' reproductive system is the genetic looms - they're sterile. It's just changing their vestigial physical characteristics.
Sarah Jane Smith was very much a strong female role model, and solved several mysteries on her own. Most of the time The Doctor solved things before her during her era, it was because it either required some esoteric knowledge that you don't get from growing up on Earth, or because it required some technology from the TARDIS.
We never saw Susan die on-screen. I'd have thought that an obvious way of handing it over would be for the TARDIS to take the dying Doctor to find her, and for him to hand over the TARDIS to her. Then you'd get a female lead character, without changing any of the mythology. You'd need some reason why Susan was off Galifrey, but the last time we saw her she was on Earth in 2150 (and while we didn't see her use the Chameleon Arch on screen, there's no reason why she couldn't have done, or constructed something similar from all of the post-invasion Dalek technology left lying around), so that's probably not too hard...
RAM cache is useless for speeding up writes. A significant (although workload-dependent) part of the performance problem with spinning disks is that if you issue a write and then need to block until it's on disk (which you need for consistency), it can easily take 5-10ms (or more) and that severely limits the performance. Often, non-server workloads include doing a lot of small synchronous writes and then no writes for a while. An SSD as a write-through cache works well here because it can reorder a lot of writes to turn (some of) them into sequential writes and it can trickle out a lot of writes while the disk is idle. This is also pretty much the best case for flash longevity: you don't need wear levelling, because you just treat the entire flash as a ring buffer and write to one end and write to the disk from the other end. You can keep the translation layer in RAM, and if there's a power failure you just replay the entire flash journal onto the disk.
The 'only reads 8GB' of unique data per day number is meaningless as an indication of how often each thing is used, however. If each day you always access the same 8GB, then an 8GB cache will be perfect for you. If you always access 8GB a day and you only access 7.5GB of it once, then a 512MB cache will be fine and you'll get no benefit from more, but you will get a big benefit from having a faster underlying storage device.
Doesn't the Windows search box include a calculator? The OS X one allows you to type expressions in and then copy the result. command-space, enter expression, command-c, escape, and you're back where you started with the result of the calculation in the pasteboard.
Linux malware is hard to write, although the SCTP vulnerability last year would have allowed a worm that ran in kernelspace and didn't depend on any installed software. Ubuntu malware, or Android malware, however, are quite easy. That 0.1% figure for Android malware in comparison to Windows malware is probably just about true if you're counting all malware written for both platforms since they were introduced, irrespective of whether it works on recent versions, but it's nowhere near close if you're counting new malware. Take a look at this list and tell me that a widely deployed Linux distribution is hard to write malware for. For example, the CURL CVE-2013-2174 allows a remote attacker with a crafted URL to run arbitrary code, and CVE-2013-1697 in Mozilla allows HTML emails displayed with Thunderbird or web pages displayed with FireFox to execute JavaScript with a privilege level that allows it to make calls to native libraries, effectively meaning arbitrary code execution with ambient privilege. CVE-2013-1052 would allow either of these attacks to upgrade privilege and then install a rootkit.
That's not how government procurement is supposed to work! A company that has failed to deliver on multiple contracts in the past should be given priority, because it has significant experience in government contracting work!
I agree. *Of course* copyright is a net loss to society. Because it is meant to strike a balance between the financial success of individuals and the benefit to society.
No it isn't. It's supposed to strike a balance between benefit to society because creative works exist and balance to society because creative works are available to all. It grants individuals a temporary exclusive distribution license because that was seen as a good way of promoting the first objective, which is a prerequisite of the second (if no one is creating new works then that's a problem), but the financial success of individuals is a means that copyright uses, not a goal of copyright.
Slashdot Mirror
GNUstep started to implement the OpenStep specification, which was a public spec for portable application development and was implemented by NeXT and Sun (hence the NS prefix on all of the class names). The most popular implementation of OpenStep is called Cocoa, and includes a lot of extensions to the base spec. We try to implement these extensions as well.
Like I mentioned, you won't get Cocoa framework on GNUStep anytime soon
That makes no sense at all. GNUstep is an open source implementation of the Cocoa framework.
WindowMaker doesn't use GNUstep and is not a GNUstep project. GNUstep has supported full transparency in applications for a few years.
The only concern I have is they do seem to be looking at it more as a framework for porting, which is the least important use from my perspective. This is the tool to build the better desktop on linux everyone claims to want.
GNUstep aims to implement the APIs that Cocoa uses. This has a natural use as a porting tool, but the main reason we're implementing the APIs is that we want to use them (which has the unfortunate side effect that ones we don't like tend not to be implemented quickly, even if lots of OS X code uses them). Over in Étoilé (which, no doubt, Slashdot's early-'90s character encoding support will mangle: Etoile with accents on both 'e's) we're building frameworks for building better environments, some of which also work on OS X and some of which don't.
No. If there were a market for a Linux build of Photoshop, then Adobe would find it much easier to port the Windows version with WINE than the OS X version with GNUstep (and I say this as a GNUstep developer). Applications like OmniGraffle, however, would be easier to port. I think we already implement most of what OmniGraffle needs, but there are lots of missing bits of APIs. I have a Summer of Code student who is working on getting the CoreAnimation / CoreGraphics stuff integrated (our current GUI code uses the NeXT DisplayPostscript APIs) which should help with a lot of things.
And on active sites it looks like IIS has been dropping in use since April 2009. Apache has actually picked up a bit since they had a big drop between 2005 and 2007, although it's down a bit since the small peak in 2011. The most interesting part is the growth of the 'other' line from about 6% to about 12% over the past 3 years. It's sad to see Lighttpd has almost died (not much development for years, now just counted as part of 'other'), but it's good to see a few different servers clustered around the 5-10% mark. Having a single server with over 50% of the market makes it quite an attractive target. It would be good to see more diversity.
And people say Microsoft doesn't innovate. Making something that's more painful to configure than Apache requires an impressive amount of R&D...
The pirated version is a download that you can watch on any device any time. The Netflix stream requires Silverlight, so I can't use it on the FreeBSD box connected to my projector nor on one of my tablets. The other tablet runs Android, so there is a Netflix streaming app, but I don't think it lets me download things and I mostly want to watch things on the tablet when I'm on a train or plane (spotty / expensive / unavailable Internet access).
In rural parts of the UK (e.g. where my mother lives), the ADSL connection isn't fast enough for streaming, but it's fine if you start downloading 10-15 minutes before you start watching, so again the pirated version wins because you can just download it and then watch it later.
Give me a service that lets me download DRM-free videos with some reasonable per-month, download-capped pricing, and I'll very happily subscribe (and, no, I'm not moving the goalposts - this is what I've been asking for for the last 10 years). Something like 30-45 hours for £10-15 would be fine. Until then, I'll keep getting the shiny disks through the post.
The difference is, the person hosting the encrypted mail and the person providing the code are usually different in non-Web cases. If a mail client is regularly making connections to something other than your web server, that's something an IDS is likely to spot. It's even something a corporate firewall may spot, if it isn't doing HTTP requests that can go via a corporate proxy. If your web page is making more requests than strictly necessary to the originating web server, that's a lot harder to spot.
You might be surprised at how many companies want decent Exchange integration from tablets...
The existence of asynchronous HTTP requests from JavaScript makes that a somewhat meaningless level of protection. The server can't decrypt the message, but you can only decrypt the message by running code that you get from the server, which runs in a context where it is allowed to send arbitrary messages back to the server. Unless you're running a JavaScript debugger and ensuring that the decryption code isn't sending the decrypted version (optionally re-encrypted with the server's key) back to the server, you aren't getting any more security than having the server decrypt it.
TorrentFreak was featured on mainstream news outlets such as CNN, The Wall Street Journal, New York Times the BBC, the Guardian and the LA Times.
Which makes it seem like they are not a sensible thing to block. I've not visited the site for a great many years, but if you only object to the sites that you use being blocked then it's very easy for censors to creep in.
Not necessarily. We know that, at some point in the future, the TARDIS becomes The Doctor's grave marker, but we don't know that it was placed there when he died. His body may have remained in The TARDIS for some hundreds of years of subjective time first. Having him preserved in his near-death state in The TARDIS with the ability to provide advice to Susan occasionally might also work nicely as a plot device.
An Internet Security suite from a company that no longer has its root certificate in my trusted list because of their inability to secure their own systems? Why on earth would I want something like that?
the pilots about 20mins out from landing saying, we can't talk to you anymore until after we land.
And they almost certainly could have landed the plane fine if you'd kept chatting. In almost every case, it would be completely fine. But very occasionally, they'd miss checking a dial or mishear ATC instructions and end up with a plane full of dead passengers, and they don't want to take that risk because, unlike many other people in this thread, they were behaving like responsible adults.
In The Five Doctors, The Master is offered another cycle of regenerations in exchange for his help. It's not clear whether he's given them, but the fact that he accepts indicates that he at least believes that this is within the High Council's power to grant. It's also implied in The Brain of Morbius that the limit on regenerations is arbitrary and imposed so that Time Lord society won't suffer the same kind of stagnation as the Sisterhood of Karn.
John Hurt is The Valeyard, who was 'between the 12th and final' incarnations, so a future one not a past one. The Time Lords in Trial of a Time Lord did say 'final' and not 13th though, but it's entirely possible that the Great Time War messed up The Doctor's timeline such that the 13th was not actually his last, or that The Matrix was simply unaware of the later ones as they were unauthorised.
They aren't changing their reproductive system. The Time Lords' reproductive system is the genetic looms - they're sterile. It's just changing their vestigial physical characteristics.
Sarah Jane Smith was very much a strong female role model, and solved several mysteries on her own. Most of the time The Doctor solved things before her during her era, it was because it either required some esoteric knowledge that you don't get from growing up on Earth, or because it required some technology from the TARDIS.
We never saw Susan die on-screen. I'd have thought that an obvious way of handing it over would be for the TARDIS to take the dying Doctor to find her, and for him to hand over the TARDIS to her. Then you'd get a female lead character, without changing any of the mythology. You'd need some reason why Susan was off Galifrey, but the last time we saw her she was on Earth in 2150 (and while we didn't see her use the Chameleon Arch on screen, there's no reason why she couldn't have done, or constructed something similar from all of the post-invasion Dalek technology left lying around), so that's probably not too hard...
RAM cache is useless for speeding up writes. A significant (although workload-dependent) part of the performance problem with spinning disks is that if you issue a write and then need to block until it's on disk (which you need for consistency), it can easily take 5-10ms (or more) and that severely limits the performance. Often, non-server workloads include doing a lot of small synchronous writes and then no writes for a while. An SSD as a write-through cache works well here because it can reorder a lot of writes to turn (some of) them into sequential writes and it can trickle out a lot of writes while the disk is idle. This is also pretty much the best case for flash longevity: you don't need wear levelling, because you just treat the entire flash as a ring buffer and write to one end and write to the disk from the other end. You can keep the translation layer in RAM, and if there's a power failure you just replay the entire flash journal onto the disk.
The 'only reads 8GB' of unique data per day number is meaningless as an indication of how often each thing is used, however. If each day you always access the same 8GB, then an 8GB cache will be perfect for you. If you always access 8GB a day and you only access 7.5GB of it once, then a 512MB cache will be fine and you'll get no benefit from more, but you will get a big benefit from having a faster underlying storage device.
Doesn't the Windows search box include a calculator? The OS X one allows you to type expressions in and then copy the result. command-space, enter expression, command-c, escape, and you're back where you started with the result of the calculation in the pasteboard.
Linux malware is hard to write, although the SCTP vulnerability last year would have allowed a worm that ran in kernelspace and didn't depend on any installed software. Ubuntu malware, or Android malware, however, are quite easy. That 0.1% figure for Android malware in comparison to Windows malware is probably just about true if you're counting all malware written for both platforms since they were introduced, irrespective of whether it works on recent versions, but it's nowhere near close if you're counting new malware. Take a look at this list and tell me that a widely deployed Linux distribution is hard to write malware for. For example, the CURL CVE-2013-2174 allows a remote attacker with a crafted URL to run arbitrary code, and CVE-2013-1697 in Mozilla allows HTML emails displayed with Thunderbird or web pages displayed with FireFox to execute JavaScript with a privilege level that allows it to make calls to native libraries, effectively meaning arbitrary code execution with ambient privilege. CVE-2013-1052 would allow either of these attacks to upgrade privilege and then install a rootkit.
That's not how government procurement is supposed to work! A company that has failed to deliver on multiple contracts in the past should be given priority, because it has significant experience in government contracting work!
I agree. *Of course* copyright is a net loss to society. Because it is meant to strike a balance between the financial success of individuals and the benefit to society.
No it isn't. It's supposed to strike a balance between benefit to society because creative works exist and balance to society because creative works are available to all. It grants individuals a temporary exclusive distribution license because that was seen as a good way of promoting the first objective, which is a prerequisite of the second (if no one is creating new works then that's a problem), but the financial success of individuals is a means that copyright uses, not a goal of copyright.