The principle modification being the removal of the "any later version" clause.
Just checked a fresh download of the MySQL 5.1.40 source. This statement is correct – the random sampling of source files I checked do not contain the "or later" clause in the licence notification.
Really? A set of hashes (the only actual data derived from the original work, as opposed to just metadata about the file) is a derivative work? That sets a terrible precedent...
There were a small number of non-PPC applications and a number of plugins that were broken or slightly-buggy after the upgrade. Many more plugins than applications, due to unsupported plugins always getting knocked around by OS upgrades (sometimes even point releases) – the major cause for plugins were the InputManagers removal which I noted. In every case I can think of these plugins simply didn't work (often didn't even try to load) in the new version – not that they broken the system.
Applications were mostly just a bit buggy (and only a handful that I use), and often a fix was released either on 10.6 release date or a few days later. As two examples, Adium has a minor bug in it's release version that causes some preference windows to behave oddly (fixed in the beta, released before 10.6 release date), and Cyberduck wouldn't run in Snow Leopard (also fixed in the beta, which was released exactly on the 10.6 release date).
Turns out the one kernel panic since upgrading was due to faulty hardware. One of my new RAM chips was faulty and is in the process of getting replaced.
As far as the 'average' kernel panics go, one every six weeks isn't quite my normal – but it's not entirely out of the realm of possibility to have one pop up. I did have a significant number (not high, but not extremely low ~1 every month or two) related to the 'safe sleep' (hibernate + suspend) feature in Leopard, and I experienced one of those recently on Snow Leopard (though it was after my initial post, but before I realised the RAM was faulty).
As far as the application crashes go, the Safari crashes were split between probably-bad-RAM-related and not-RAM-related-at-all, but that's been my experience overall – occasional application crashes do happen.
As I linked to another person in this thread, PhotoRec works fine on OS X as long as you aren't deathly afraid of the command line (and have a spare drive for writing out all the files it finds to).
Sure, it's a bit messy with the files (as are most undelete programs – though PhotoRec doesn't even make a cursory attempt, beyond file names), but it's pretty good at getting everything not-written-over in my experience.
PhotoRec (don't let the name fool you) works for most user data files. It doesn't organise them (just fairly-opaque names like "f12948529.png" or "b29458923.zip"), and picks up a fair amount of garbage (OS files, for example), but it's quite good at getting everything that hasn't been overwritten.
It's explicitly noted that it doesn't happen every time. It's very likely they did test it, and just missed it. It's not necessarily an excuse, but bugs do happen, and this has not been reported during the beta – meaning it's either exceptionally rare or a very recent bug. I'd bet on the former.
On a different note, the CNET article takes a very sensationalist approach with using the phrase "plagued with bugs". There's a few bugs, reported by a vocal minority of users (one of which they list – incompatibilities – isn't really a bug, just a consequence of being a new OS version with new features, changed features, and a few removed features*). I've been using Snow Leopard for the past month-and-a-half, and have experienced only a tiny handful of non-damaging crashes. One kernel panic, about three or four Safari crashes. It's around the average number of problems I've experienced on most OS/version combinations.
* One such removal is a relatively undocumented 'hack' called "InputManagers" which loads code into every Cocoa application that starts up. These no longer work in 64-bit applications, and such plugin functionality has to be re-implemented using either an application-specific plugin format (where available) or as a mach_inject background process.
What made this acceptable, even enjoyable, is how this was played with and referenced. Even early on in the series he had a tendency to get killed and revived (initially using the sarcophagus) -- and the references made by the whole team, particularly O'Neill, made the entire thing into a running gag, as opposed to a running annoyance for me.
That's part of what made SG-1 and Atlantis so enjoyable for me -- the humorous, somewhat self-aware edge to the characters and the series as a whole.
The same is done in Canada. My personal domain name has hidden information (just shows registrar and DNS servers), whereas domains registered as belonging to a corporation have full information available. I believe this was changed sometime in the past five years -- as I clearly recall personal domain names having fully visible information back in 2004, and maybe even as late as 2007 or 2008.
Sure, most users would probably immediately delete it, but it would be the users CHOICE too.
You haven't really worked with many users, have you? Their program stops working and gives them an option to make it work again, do you really think they'll read/follow why they shouldn't make it work?
Or back up copies encrypted in another way. Lose one set? No problem, a different key, program, algorithm, or some combination of those unlocks another copy. If it's important enough to encrypt and important enough to backup, it's important enough to backup right.
Flashblock (or equivalent - something is available for most browsers) is even better for this - it blocks legitimate (non-ad) Flash until you want it playing. Great if you have a few tabs open to something like YouTube.
Likely it's identical to the device that comes with/works with some treadmills. It detects BPM (beats per minute) and that's pretty much it. That's about all the data that's useful for pure exercise monitoring anyway. If this is a public middle school and they're just asking you to buy the strap and not the device, then that's likely the most sophisticated they could afford, even if there was 'evil' motivations behind it. Seen physical education budgets lately?
Haiku has an edge over the old BeOS - from what I understand it's mostly POSIX compliant, meaning Linux and other applications won't be anywhere near as difficult to port over in a least a somewhat useful (though maybe not BeOS in philosophy) fashion.
You're seriously suggesting that 4GiB of RAM is functionally necessary for Win7? Most laptops didn't come with more than 1GiB standard until about two years ago. What's it doing that it's chewing up more than 2GiB for active data, and not just spare drive cache?
From what I've seen, Apple definitely got a good deal out of the WebKit community interaction - there's a huge number of people doing bug reporting, snapshot testing, and submitting patches as a result.
How about we go with the library being a developer tool, but the results of what the library does being of benefit to everyone? It both makes utilizing threads easier for developers, as well as making the OS and applications (when most/all of them use this library) significantly 'smarter' about how much it can be doing at once, generally keeping responsiveness at an acceptable level. It doesn't just improve the development experience - it also improves the end-user experience.
Slashdot Mirror
How does this fix the apps they ported being mostly IO bound in a lot of cases and 99% of the cores will still just be eating out of their noses?
Loads and loads of RAM/cache, possibly?
You aren't. You're supposed to pay up for each and every item you ever view, hear, watch, or play.
The principle modification being the removal of the "any later version" clause.
Just checked a fresh download of the MySQL 5.1.40 source. This statement is correct – the random sampling of source files I checked do not contain the "or later" clause in the licence notification.
Really? A set of hashes (the only actual data derived from the original work, as opposed to just metadata about the file) is a derivative work? That sets a terrible precedent...
Was that a stealth pun?
There were a small number of non-PPC applications and a number of plugins that were broken or slightly-buggy after the upgrade. Many more plugins than applications, due to unsupported plugins always getting knocked around by OS upgrades (sometimes even point releases) – the major cause for plugins were the InputManagers removal which I noted. In every case I can think of these plugins simply didn't work (often didn't even try to load) in the new version – not that they broken the system.
Applications were mostly just a bit buggy (and only a handful that I use), and often a fix was released either on 10.6 release date or a few days later. As two examples, Adium has a minor bug in it's release version that causes some preference windows to behave oddly (fixed in the beta, released before 10.6 release date), and Cyberduck wouldn't run in Snow Leopard (also fixed in the beta, which was released exactly on the 10.6 release date).
Turns out the one kernel panic since upgrading was due to faulty hardware. One of my new RAM chips was faulty and is in the process of getting replaced.
As far as the 'average' kernel panics go, one every six weeks isn't quite my normal – but it's not entirely out of the realm of possibility to have one pop up. I did have a significant number (not high, but not extremely low ~1 every month or two) related to the 'safe sleep' (hibernate + suspend) feature in Leopard, and I experienced one of those recently on Snow Leopard (though it was after my initial post, but before I realised the RAM was faulty).
As far as the application crashes go, the Safari crashes were split between probably-bad-RAM-related and not-RAM-related-at-all, but that's been my experience overall – occasional application crashes do happen.
As I linked to another person in this thread, PhotoRec works fine on OS X as long as you aren't deathly afraid of the command line (and have a spare drive for writing out all the files it finds to).
Sure, it's a bit messy with the files (as are most undelete programs – though PhotoRec doesn't even make a cursory attempt, beyond file names), but it's pretty good at getting everything not-written-over in my experience.
PhotoRec (don't let the name fool you) works for most user data files. It doesn't organise them (just fairly-opaque names like "f12948529.png" or "b29458923.zip"), and picks up a fair amount of garbage (OS files, for example), but it's quite good at getting everything that hasn't been overwritten.
It's explicitly noted that it doesn't happen every time. It's very likely they did test it, and just missed it. It's not necessarily an excuse, but bugs do happen, and this has not been reported during the beta – meaning it's either exceptionally rare or a very recent bug. I'd bet on the former.
On a different note, the CNET article takes a very sensationalist approach with using the phrase "plagued with bugs". There's a few bugs, reported by a vocal minority of users (one of which they list – incompatibilities – isn't really a bug, just a consequence of being a new OS version with new features, changed features, and a few removed features*). I've been using Snow Leopard for the past month-and-a-half, and have experienced only a tiny handful of non-damaging crashes. One kernel panic, about three or four Safari crashes. It's around the average number of problems I've experienced on most OS/version combinations.
* One such removal is a relatively undocumented 'hack' called "InputManagers" which loads code into every Cocoa application that starts up. These no longer work in 64-bit applications, and such plugin functionality has to be re-implemented using either an application-specific plugin format (where available) or as a mach_inject background process.
Daniel Jackson's repeated Ascensions/Falls
What made this acceptable, even enjoyable, is how this was played with and referenced. Even early on in the series he had a tendency to get killed and revived (initially using the sarcophagus) -- and the references made by the whole team, particularly O'Neill, made the entire thing into a running gag, as opposed to a running annoyance for me.
That's part of what made SG-1 and Atlantis so enjoyable for me -- the humorous, somewhat self-aware edge to the characters and the series as a whole.
The same is done in Canada. My personal domain name has hidden information (just shows registrar and DNS servers), whereas domains registered as belonging to a corporation have full information available. I believe this was changed sometime in the past five years -- as I clearly recall personal domain names having fully visible information back in 2004, and maybe even as late as 2007 or 2008.
So they have enough resources to have four vehicles chase down a single car with a family in it, but not to actually man their security checkpoints?
Idiots.
I'd test as a mix of everything, so they'd get my nationality right, eh?
Cloud is bad enough. Starting up bullshit analogies with clouds and rain just muddy whatever you're talking about far, far more than is necessary.
Sure, most users would probably immediately delete it, but it would be the users CHOICE too.
You haven't really worked with many users, have you? Their program stops working and gives them an option to make it work again, do you really think they'll read/follow why they shouldn't make it work?
Or back up copies encrypted in another way. Lose one set? No problem, a different key, program, algorithm, or some combination of those unlocks another copy. If it's important enough to encrypt and important enough to backup, it's important enough to backup right.
Flashblock (or equivalent - something is available for most browsers) is even better for this - it blocks legitimate (non-ad) Flash until you want it playing. Great if you have a few tabs open to something like YouTube.
Likely it's identical to the device that comes with/works with some treadmills. It detects BPM (beats per minute) and that's pretty much it. That's about all the data that's useful for pure exercise monitoring anyway. If this is a public middle school and they're just asking you to buy the strap and not the device, then that's likely the most sophisticated they could afford, even if there was 'evil' motivations behind it. Seen physical education budgets lately?
So yeah, just a little paranoid...
Haiku has an edge over the old BeOS - from what I understand it's mostly POSIX compliant, meaning Linux and other applications won't be anywhere near as difficult to port over in a least a somewhat useful (though maybe not BeOS in philosophy) fashion.
You're seriously suggesting that 4GiB of RAM is functionally necessary for Win7? Most laptops didn't come with more than 1GiB standard until about two years ago. What's it doing that it's chewing up more than 2GiB for active data, and not just spare drive cache?
GlimmerBlocker is similar and somewhat easier to configure, but less mature and runs on OSX only.
Works great with Safari, though, with not having to worry about things like SafariAdblock causing stability problems.
From what I've seen, Apple definitely got a good deal out of the WebKit community interaction - there's a huge number of people doing bug reporting, snapshot testing, and submitting patches as a result.
How about we go with the library being a developer tool, but the results of what the library does being of benefit to everyone? It both makes utilizing threads easier for developers, as well as making the OS and applications (when most/all of them use this library) significantly 'smarter' about how much it can be doing at once, generally keeping responsiveness at an acceptable level. It doesn't just improve the development experience - it also improves the end-user experience.
Trial and error?