The data show that, across sectors, 66 percent of new establishments were still in existence 2 years after their birth, and 44 percent were still in existence 4 years after.
It looks like the unemployment rate for college graduates hovers around 4.2%, even in this economy. And it is more than twice that without a college degree.
I'm not sure how to directly compare those numbers though. Maybe you'd have to look at how long entrepreneurs spend between new businesses... or making money from new businesses? Anyway, the point is that starting a successful business is HARD. Unless you have a clear business plan for new business, it doesn't really make much sense ot drop out of college. SUre, if you have some vision and real skills, by all means, drop out and follow your dream. But to just drop out because that's what Gates and Zuckerberg did is just stupid. Just because you CAN be successful without college doesn't mean that's how you become successful (something you might learn in a logic class in college, BTW;-)
Sorry, I meant to say that personal computing was held back by DOS. Not just the PC. Obviously only IBM of the competitors I mentioned were competing on the PC market.
Mind explaining the distinction between between Gates and Zuckerberg that you're trying to highlight? Is it a personal judgment or are you valuing the products that the corresponding companies have put out? Because if it is personal... seems to me that both characters have a lot in common. As for the products it is more difficult to judge. Personally, I think the PC was held back for a long long time by DOS and, then, early versions of Windows which Gates brought to the scene. DOS was a shitty excuse for operating system from the start and remained dominant on the PC for far too long without ever really improving while companies like Commodore (AmigaOS), Apple (MacOS), and even IBM (OS/2) were doing much more interesting and innovative things. Imagine if they weren't held back by Microsoft's dominance in the OS market. Sure, Windows is an acceptable operating system now (it is mostly stable, finally), but that's only after 2 decades of utter crap. I'm not sure the world needs more Gateses. They're no better than the Zuckerbergs anyway.
I would hope that if you're dropping $150k on college, you're ready to start a lucrative career in something. If you don't, however, have plans to do anyting big, $150k is probably way too much to be spending on school.
Anyway, the failure rate of new businesses is MUCH higher than that of college graduates.
Maybe he feels like student loans hold people back from starting new businesses... compelling people to get a full time job as soon as possible. Just a thought.
Didn't it still have a slide out keyboard though? I think the big thing with the iPhone was going all touch. But anyway, I've never really been hung up on who came up with ideas first. It certainly isn't worth filing a lawsuit over. Tech should be more like fashion where copying ideas is not just allowed, but expected. Sometimes it stupid how companies are encouraged to do something different just for the sake of being different and not because it is really better.
Honestly, I'd be more worried about the quality of the streaming capabilities. Typically when they bundle these features into TVs they're subpar. In the past you could buy TVs with VHS and/or DVD built in, but this was usually reserved for low end TVs or TVs that had installation constraints such that having external players was just inconvenient. But if you've got an entertainment center, there's little drawback to having an external player.
Your new TV set contains a computer that performs the functions provided by the external box. The firmware for that computer can be reprogrammed.
Yes and no. Depends on whether or not the TV relies on hardware acceleration for video playback. TV manufacturers coudl have built DVD players into every TV, but people either already had a DVD player or standalone DVD players were cheap enough that it wasn't a compelling feature in a TV. Also, All-in-One electronics are typically low end or they do one thing well, but not the other. Many people would rather have the option of pairing their player of choice with their TV of choice. It can be really difficult to get everything you want in a single package. There will always be a market for devices that do one thing very well... or a few related things very well. Like I would combine a game console and video streaming, but I wouldn't build a gaming console into a TV. Sure, you can update the software in teh TV, but you're always going to be limited by the processing power in the TV.
Or TVs with VCR/DVD built in. Pretty much the same thing here except that you could theoretically upgrade your streaming capability with a firmware update. So you're a little more modular in that respect. That could be the important difference. As long as your TV has sufficient processing power and and isn't limited to specific codecs because of hardware accelerators, you should be good to go. That last bit is what I'd be mainly concerned about when buying any hardware streaming device. Can it be upgraded? Or is it is like a TiVo stuck using MPEG-2 because that's all the hardware is capable of.
What settings? You mean your login information for Netflix? Presumably you're buy a new TV because it is better. So why would you want to transfer any settings other than your account information? The amount of information stored locally should be minimal.
Yeah, but you won't be taking advantage of your shiny new 1080p TV. Netflix is alright for casual TV watching, but you're lucky if you can get DVD quality. If you really want to have a home theater, BluRay is the way to go. It is not a dead format. My home wireless (802.11n) can barely push full HD video, forget about my internet connection doing it reliably any time soon.
The cool thing about the PS3 is that you get a solid streaming box AND a blu-ray player in one. So it is almost worth buying just for that. As for gaming, I'm normally a PC gamer but sometimes it is just nice to pop in a disc and sit on the couch to play a game. I could plunk a gaming PC behind my HD TV, but that's not really where I want it.
I have a feeling next generation consoles will do a much better job of streaming.
Eh? My PS3 streams Netflix great. In fact, it is one of the main reasons I bought it. That and the blu-ray player. The problem is limited bandwidth from Netflix. I can't get Blu-Ray quality or 5.1 sound. But that has nothing to do with the potential of my console as far as I know.
Well, maybe I'm mistaken, but the iPhone did seem to define the look and feel of modern smart phones. The idea of ditching the physical keyboard and making the screen the full size of the devices was somewhat revolutionary. That's not to say that anyone should be prevented from implementing that general design because I don't think people can own special rights to that sort of thing, but I think it is appropriate to give Apple some credit for its innovation.
Because they provide different functionality and advantages. A lot of the time distros will guess which is best for thier user. kitchen-sink or media-server distros will use pulse, slim will use just plain alsa, and and studio/artsy distros will use jack.
But almost nobody cares. That's what I'm saying. I'd rather just have it work and not have to think about it. I have never worried about the underlying sound architecture on any other major platform besides LInux. Linux makes it an issue where it should not be. It is choice for the sake of choice, not for the sake of the user.
That's not even the biggest advantage though. Constructing from a small base up, along with being able to define system wide use flags makes it relatively easy to customize a system to your needs and expectations.
Meh, it is overrated. I spent years and years tinkering with Linux, making it do what I wanted, customizing it to my "needs" until I realized that I wasn't actually getting that much done. I had fun and I learned a lot, but ultimately what I really needed was something that just worked. I also learned that sometimes it is more efficient to adapt to a system than try to make the system adapt to you. So... I use OS X. One of the least flexible systems in terms of doing things differently than Apple intended, but I love it. It just fucking works.
Nope, but both packages will link or compile against the newest libraries for it's dependencies rather than keeping old and potentially insecure libraries.
A theoretical problem, not a practical concern in the vast majority of cases. So basically Gentoo makes me jump through a dozen hoops for only theoretical gains. Yay.
Or... I could copy the app bundle with my mouse. Hmmm..
Sure if you really had the hankering to do so, but I'm not even sure of the advantages unless you don't have root access of the matching.
The advantage is that it is way simpler and requires far less commitment. I think it is one of those things where you don't see the use in something until it is just so easy to do that you find ways to take advantage of it. Want to take an app on the go? Just copy the app bundle to a thumbdrive. No installers, no package managers, no compilers. Run it right from the thumbdrive. Could not get simpler.
Will install the daily build of firefox. Sure the first time takes a little longer, but the the second or third time you go to update it your out ahead because you don't have to go find the bundle, and copy/install it where you want it. In fact from that point on you don't have to even think about it. Every time you update the system, you will have the daily build of Firefox with it. See where I'm going here?
First of all, I'm not normally going to keep the nightly build. More often than not I'm just trying it out and I want to run it side by side with the stable version of the browser. I don't really want to commit to installing it. I just want to plunk it on my desktop and run. I cetainly dont' want wait 30 minutes for the damn thing to compile. It is an unnecessary step. Let someone else do the compiling. The compiling is what annoyed me most about Gentoo. The most I ever have to think about when using the Firefox app bundle is backing up my profile.
I could be mistaken, but it is my understanding that Windows is rather sophisticated in terms of utilizing audio hardware wheres on Linux they're still trying to figure out a common way to mix audio from different applications. So I'm not saying Linux should just copy or emulate Windows, but all that "choice" is crippling Linux on the desktop. And it isn't like there's no choice on Windows. There's a dozen different libraries and APIs for audio.
Because if developers choose what libraries to use mainly based on whether the library is likely to be already installed, then new but useful libraries are unlikely to become popular enough to become a standard install.
So? I'd rather have a consistent development target, even if it is mediocre, than a mishmash of incomparable libraries like we find on Linux. That's definitely not a good situation to be in. You end up installing all the libraries and all of you apps look and behave differently.
In addition it becomes tempting for the developer to just implement the libraries functionality, but of course that is a drain on application developer time and is likely to have bugs that were already shaken out of the library. In an ideal world, developers would choose what library to use solely based on how well the library does the job.
Except that doesn't happen. People choose libraries for all sorts of reasons. Mainly it is just whatever you're comfortable with or what you want to tinker with or whatever aligns with your programming language of choice. Like C? GTK. Like C++? Qt. That's about as far as most developers get, I think. Better just to tell them what they have to use if they want to develop for the platform. Apple has done such a great job at this. Though Apple does support languages other than Objective-C out of the box.
The real world already falls short of that (e.g. API stability and likelihood of the library still being maintained down the road are major concerns), but we should try to avoid making the situation worse than it already is.
OS X does not really fall very short in this regard. Cocoa is a sane, stable development platform.
On Linux the situation is pretty bad and it is precisely because there is so much fragmentation and "choice." The solution is to declare one definitive set of libraries that makes up the core that everyone uses and you can add auxiliary libraries as needed. This is kind of what LSB does, but it doesn't go nearly far enough. Linux needs a well defined target for desktop apps. No more of this Qt/GTK split. Hell, linux apps can't even decide on a sound library for crying out loud.. What is this, 1995? It doesn't have to be perfect. Just decide on SOMETHING.
If they give equity in the form of stock options, you become "vested" over time. So the company gets some extra assurance that you'll be around and you get your equity. That's what they get out of it.
Eh, you do know you can build your apps from source against whatever damn libraries you please, right?
Sure, and then you suddenly lose any benefit that you might have gained from package management. In fact, this is one of the things that annoyed me so much about running Linux. If I wanted to do anything outside of the package manager, I was prtty much on my own and my system suddenly became 10 times harder to maintain. If for some reason I wanted to upgrade the base system, I'd have to go back and recompile all that custom shit by h and.
We've got a few nifty little things like make, gcc, and gnu autopuke that can make that happen for you. And at least on Debian and slackware systems, it's barely any more work at all to rebuild an entire package, then install it through your existing package manager, so you're not even confronted with the issue of dependency resolution not knowing about locally installed apps.
Have you actually tried backporting packages on Debian systems? You know there are strict dependencies, so you can't usually just backport your main app. You also have to backport many of the dependencies. And then you might find out that one f those dependencies won't build clearly and you spend half a day trying to figure that out. It is hell.
So this "straitjacket" is completely optional, and you can get out of it at any time, it doesn't seem much like a straitjacket.
Except that it turns out that being out of the straight jacket is worse than the straight jacket. OUt of the frying pan and into the fire, so to speak.
Maybe you've been using one system that makes this stuff hard (Redhat, maybe? I haven't so much as diddled with an RPM-based distro since '99), but the oldschool distroes all make it easy, because they're made by hackers who wouldn't want to be constrained that way themselves.
I've used half a dozen or more distributions including Debian, Gentoo, Ubuntu, Redhat, CentOS, and Slackware. These hackers have designed themselves into the straight jacket of package management because the alternative is worse and they are too disorganized to do any better. Package management is all about constraining yourself. That's why it works. That's how it works.
I don't see how.app packaging is any different from statically compiled binaries.
It is different because they are dynamically linked (for the most part) against system libraries. THe point is that OS X is a broad, stable target unlike the highly fragmented Linux distributions. A Linux developer can't even depend on you having GTK+ on your system, must less the correct version of it. If I target OS X 10.5, on the other hand, I know exactly what GUI libraries will be available and there's no need to statically link or code in libraries dependencies for a package manager to work out.
The one draw back to all this is that applications are not terribly backward compatible. Currently it seems like many applications don't' support anything older than OS X 10.5, which is fine with me since I have no reason NOT to run at least 10.5.
Besides, that greatly favors popular libraries and ignores libraries which may not be pre-installed whereas package management with external libraries are neutral to this.
Keywords let you use pulse audio on every application if you want, or not have that library at all on the system.
Right, because the sound library my applications use is exactly the kind of thing I should be thinking about as a user. Why the hell should I have to care what audio library I'm using? Why is that still an issue on Linux? It is 2011 for Chrissake.
Yes, I used Gentoo years ago and it was the biggest waste of tme of all the Linux distributions... once I got past the novelty of compiling everything tuned for my CPU. I'm getting too old for that shit.
It's fairly straightfoward how to keep multiple versions of a package as well.
More straight forward than just dragging, dropping, and renaming app bundles?
To get custom locations you can copy the ebuild to an overlay and change a couple of the options that get passed to the configuration tool if your into that sort of thing.
Or... I could copy the app bundle with my mouse. Hmmm.
There are also methods to use versions that are newer of older than the latest stable
Are any of those methods easier than just, you know, copying an app bundle? You see where I'm going here?
Still amounts to a straight jacket. You get latest software, but you still are forced to run whatever happens to be in the repo and what happens to be current for you your distribution and it installs only where the package dictates that it be installed. Also, you are forced to update things you may not want to because that's what your latest and greatest software happened to built against. It is all so tightly coupled. Overall it is just a bad user experience for anyone who may not want to do things exactly as the package maintainers think you should.
But the "latest and greatest" aspect is just part of it. I want to be able to install any version I want/need, sometimes with two different versions side by side (Firefox 3 and FIrefox 4, for example). Also, I like being able to run an app from wherever I want. Hell, on OS X, I can launch most apps right from the compressed disk image it came in. I do that kind of things with little games I want to try out or whatever. Or I can carry an app around on a USB stick. It is just so handy, especially if you have to work on other Macs. I would never ever go back to centralized package management. As an end user, centralized package mangement actually gives me very little other than a single place to look for apps. But who cares? The Google is strong with me.
Slashdot Mirror
According to: http://www.bls.gov/opub/mlr/2005/05/ressum.pdf
The data show that, across sectors, 66 percent of new establishments were still in existence 2 years after their birth, and 44 percent were still in existence 4 years after.
According to: http://www.bls.gov/news.release/empsit.t04.htm
It looks like the unemployment rate for college graduates hovers around 4.2%, even in this economy. And it is more than twice that without a college degree.
I'm not sure how to directly compare those numbers though. Maybe you'd have to look at how long entrepreneurs spend between new businesses... or making money from new businesses? Anyway, the point is that starting a successful business is HARD. Unless you have a clear business plan for new business, it doesn't really make much sense ot drop out of college. SUre, if you have some vision and real skills, by all means, drop out and follow your dream. But to just drop out because that's what Gates and Zuckerberg did is just stupid. Just because you CAN be successful without college doesn't mean that's how you become successful (something you might learn in a logic class in college, BTW ;-)
But paying them directly gives them money to start a business....
Sorry, I meant to say that personal computing was held back by DOS. Not just the PC. Obviously only IBM of the competitors I mentioned were competing on the PC market.
Mind explaining the distinction between between Gates and Zuckerberg that you're trying to highlight? Is it a personal judgment or are you valuing the products that the corresponding companies have put out? Because if it is personal... seems to me that both characters have a lot in common. As for the products it is more difficult to judge. Personally, I think the PC was held back for a long long time by DOS and, then, early versions of Windows which Gates brought to the scene. DOS was a shitty excuse for operating system from the start and remained dominant on the PC for far too long without ever really improving while companies like Commodore (AmigaOS), Apple (MacOS), and even IBM (OS/2) were doing much more interesting and innovative things. Imagine if they weren't held back by Microsoft's dominance in the OS market. Sure, Windows is an acceptable operating system now (it is mostly stable, finally), but that's only after 2 decades of utter crap. I'm not sure the world needs more Gateses. They're no better than the Zuckerbergs anyway.
I would hope that if you're dropping $150k on college, you're ready to start a lucrative career in something. If you don't, however, have plans to do anyting big, $150k is probably way too much to be spending on school.
Anyway, the failure rate of new businesses is MUCH higher than that of college graduates.
Maybe he feels like student loans hold people back from starting new businesses... compelling people to get a full time job as soon as possible. Just a thought.
Didn't it still have a slide out keyboard though? I think the big thing with the iPhone was going all touch. But anyway, I've never really been hung up on who came up with ideas first. It certainly isn't worth filing a lawsuit over. Tech should be more like fashion where copying ideas is not just allowed, but expected. Sometimes it stupid how companies are encouraged to do something different just for the sake of being different and not because it is really better.
Honestly, I'd be more worried about the quality of the streaming capabilities. Typically when they bundle these features into TVs they're subpar. In the past you could buy TVs with VHS and/or DVD built in, but this was usually reserved for low end TVs or TVs that had installation constraints such that having external players was just inconvenient. But if you've got an entertainment center, there's little drawback to having an external player.
Your new TV set contains a computer that performs the functions provided by the external box. The firmware for that computer can be reprogrammed.
Yes and no. Depends on whether or not the TV relies on hardware acceleration for video playback. TV manufacturers coudl have built DVD players into every TV, but people either already had a DVD player or standalone DVD players were cheap enough that it wasn't a compelling feature in a TV. Also, All-in-One electronics are typically low end or they do one thing well, but not the other. Many people would rather have the option of pairing their player of choice with their TV of choice. It can be really difficult to get everything you want in a single package. There will always be a market for devices that do one thing very well... or a few related things very well. Like I would combine a game console and video streaming, but I wouldn't build a gaming console into a TV. Sure, you can update the software in teh TV, but you're always going to be limited by the processing power in the TV.
Or TVs with VCR/DVD built in. Pretty much the same thing here except that you could theoretically upgrade your streaming capability with a firmware update. So you're a little more modular in that respect. That could be the important difference. As long as your TV has sufficient processing power and and isn't limited to specific codecs because of hardware accelerators, you should be good to go. That last bit is what I'd be mainly concerned about when buying any hardware streaming device. Can it be upgraded? Or is it is like a TiVo stuck using MPEG-2 because that's all the hardware is capable of.
What settings? You mean your login information for Netflix? Presumably you're buy a new TV because it is better. So why would you want to transfer any settings other than your account information? The amount of information stored locally should be minimal.
Yeah, but you won't be taking advantage of your shiny new 1080p TV. Netflix is alright for casual TV watching, but you're lucky if you can get DVD quality. If you really want to have a home theater, BluRay is the way to go. It is not a dead format. My home wireless (802.11n) can barely push full HD video, forget about my internet connection doing it reliably any time soon.
The cool thing about the PS3 is that you get a solid streaming box AND a blu-ray player in one. So it is almost worth buying just for that. As for gaming, I'm normally a PC gamer but sometimes it is just nice to pop in a disc and sit on the couch to play a game. I could plunk a gaming PC behind my HD TV, but that's not really where I want it.
I have a feeling next generation consoles will do a much better job of streaming.
Eh? My PS3 streams Netflix great. In fact, it is one of the main reasons I bought it. That and the blu-ray player. The problem is limited bandwidth from Netflix. I can't get Blu-Ray quality or 5.1 sound. But that has nothing to do with the potential of my console as far as I know.
Well, maybe I'm mistaken, but the iPhone did seem to define the look and feel of modern smart phones. The idea of ditching the physical keyboard and making the screen the full size of the devices was somewhat revolutionary. That's not to say that anyone should be prevented from implementing that general design because I don't think people can own special rights to that sort of thing, but I think it is appropriate to give Apple some credit for its innovation.
Because they provide different functionality and advantages. A lot of the time distros will guess which is best for thier user. kitchen-sink or media-server distros will use pulse, slim will use just plain alsa, and and studio/artsy distros will use jack.
But almost nobody cares. That's what I'm saying. I'd rather just have it work and not have to think about it. I have never worried about the underlying sound architecture on any other major platform besides LInux. Linux makes it an issue where it should not be. It is choice for the sake of choice, not for the sake of the user.
That's not even the biggest advantage though. Constructing from a small base up, along with being able to define system wide use flags makes it relatively easy to customize a system to your needs and expectations.
Meh, it is overrated. I spent years and years tinkering with Linux, making it do what I wanted, customizing it to my "needs" until I realized that I wasn't actually getting that much done. I had fun and I learned a lot, but ultimately what I really needed was something that just worked. I also learned that sometimes it is more efficient to adapt to a system than try to make the system adapt to you. So... I use OS X. One of the least flexible systems in terms of doing things differently than Apple intended, but I love it. It just fucking works.
Nope, but both packages will link or compile against the newest libraries for it's dependencies rather than keeping old and potentially insecure libraries.
A theoretical problem, not a practical concern in the vast majority of cases. So basically Gentoo makes me jump through a dozen hoops for only theoretical gains. Yay.
Or... I could copy the app bundle with my mouse. Hmmm..
Sure if you really had the hankering to do so, but I'm not even sure of the advantages unless you don't have root access of the matching.
The advantage is that it is way simpler and requires far less commitment. I think it is one of those things where you don't see the use in something until it is just so easy to do that you find ways to take advantage of it. Want to take an app on the go? Just copy the app bundle to a thumbdrive. No installers, no package managers, no compilers. Run it right from the thumbdrive. Could not get simpler.
Sure.
echo "www-client/firefox-9999" >> /etc/portage/package.USE
emerge -uD world
Will install the daily build of firefox. Sure the first time takes a little longer, but the the second or third time you go to update it your out ahead because you don't have to go find the bundle, and copy/install it where you want it. In fact from that point on you don't have to even think about it. Every time you update the system, you will have the daily build of Firefox with it. See where I'm going here?
First of all, I'm not normally going to keep the nightly build. More often than not I'm just trying it out and I want to run it side by side with the stable version of the browser. I don't really want to commit to installing it. I just want to plunk it on my desktop and run. I cetainly dont' want wait 30 minutes for the damn thing to compile. It is an unnecessary step. Let someone else do the compiling. The compiling is what annoyed me most about Gentoo. The most I ever have to think about when using the Firefox app bundle is backing up my profile.
Why would a professional wrestling channel care about space flight?
I could be mistaken, but it is my understanding that Windows is rather sophisticated in terms of utilizing audio hardware wheres on Linux they're still trying to figure out a common way to mix audio from different applications. So I'm not saying Linux should just copy or emulate Windows, but all that "choice" is crippling Linux on the desktop. And it isn't like there's no choice on Windows. There's a dozen different libraries and APIs for audio.
Because if developers choose what libraries to use mainly based on whether the library is likely to be already installed, then new but useful libraries are unlikely to become popular enough to become a standard install.
So? I'd rather have a consistent development target, even if it is mediocre, than a mishmash of incomparable libraries like we find on Linux. That's definitely not a good situation to be in. You end up installing all the libraries and all of you apps look and behave differently.
In addition it becomes tempting for the developer to just implement the libraries functionality, but of course that is a drain on application developer time and is likely to have bugs that were already shaken out of the library. In an ideal world, developers would choose what library to use solely based on how well the library does the job.
Except that doesn't happen. People choose libraries for all sorts of reasons. Mainly it is just whatever you're comfortable with or what you want to tinker with or whatever aligns with your programming language of choice. Like C? GTK. Like C++? Qt. That's about as far as most developers get, I think. Better just to tell them what they have to use if they want to develop for the platform. Apple has done such a great job at this. Though Apple does support languages other than Objective-C out of the box.
The real world already falls short of that (e.g. API stability and likelihood of the library still being maintained down the road are major concerns), but we should try to avoid making the situation worse than it already is.
OS X does not really fall very short in this regard. Cocoa is a sane, stable development platform.
On Linux the situation is pretty bad and it is precisely because there is so much fragmentation and "choice." The solution is to declare one definitive set of libraries that makes up the core that everyone uses and you can add auxiliary libraries as needed. This is kind of what LSB does, but it doesn't go nearly far enough. Linux needs a well defined target for desktop apps. No more of this Qt/GTK split. Hell, linux apps can't even decide on a sound library for crying out loud.. What is this, 1995? It doesn't have to be perfect. Just decide on SOMETHING.
If they give equity in the form of stock options, you become "vested" over time. So the company gets some extra assurance that you'll be around and you get your equity. That's what they get out of it.
Eh, you do know you can build your apps from source against whatever damn libraries you please, right?
Sure, and then you suddenly lose any benefit that you might have gained from package management. In fact, this is one of the things that annoyed me so much about running Linux. If I wanted to do anything outside of the package manager, I was prtty much on my own and my system suddenly became 10 times harder to maintain. If for some reason I wanted to upgrade the base system, I'd have to go back and recompile all that custom shit by h and.
We've got a few nifty little things like make, gcc, and gnu autopuke that can make that happen for you. And at least on Debian and slackware systems, it's barely any more work at all to rebuild an entire package, then install it through your existing package manager, so you're not even confronted with the issue of dependency resolution not knowing about locally installed apps.
Have you actually tried backporting packages on Debian systems? You know there are strict dependencies, so you can't usually just backport your main app. You also have to backport many of the dependencies. And then you might find out that one f those dependencies won't build clearly and you spend half a day trying to figure that out. It is hell.
So this "straitjacket" is completely optional, and you can get out of it at any time, it doesn't seem much like a straitjacket.
Except that it turns out that being out of the straight jacket is worse than the straight jacket. OUt of the frying pan and into the fire, so to speak.
Maybe you've been using one system that makes this stuff hard (Redhat, maybe? I haven't so much as diddled with an RPM-based distro since '99), but the oldschool distroes all make it easy, because they're made by hackers who wouldn't want to be constrained that way themselves.
I've used half a dozen or more distributions including Debian, Gentoo, Ubuntu, Redhat, CentOS, and Slackware. These hackers have designed themselves into the straight jacket of package management because the alternative is worse and they are too disorganized to do any better. Package management is all about constraining yourself. That's why it works. That's how it works.
I don't see how .app packaging is any different from statically compiled binaries.
It is different because they are dynamically linked (for the most part) against system libraries. THe point is that OS X is a broad, stable target unlike the highly fragmented Linux distributions. A Linux developer can't even depend on you having GTK+ on your system, must less the correct version of it. If I target OS X 10.5, on the other hand, I know exactly what GUI libraries will be available and there's no need to statically link or code in libraries dependencies for a package manager to work out.
The one draw back to all this is that applications are not terribly backward compatible. Currently it seems like many applications don't' support anything older than OS X 10.5, which is fine with me since I have no reason NOT to run at least 10.5.
Besides, that greatly favors popular libraries and ignores libraries which may not be pre-installed whereas package management with external libraries are neutral to this.
And I should care... why?
Keywords let you use pulse audio on every application if you want, or not have that library at all on the system.
Right, because the sound library my applications use is exactly the kind of thing I should be thinking about as a user. Why the hell should I have to care what audio library I'm using? Why is that still an issue on Linux? It is 2011 for Chrissake.
Yes, I used Gentoo years ago and it was the biggest waste of tme of all the Linux distributions... once I got past the novelty of compiling everything tuned for my CPU. I'm getting too old for that shit.
It's fairly straightfoward how to keep multiple versions of a package as well.
More straight forward than just dragging, dropping, and renaming app bundles?
To get custom locations you can copy the ebuild to an overlay and change a couple of the options that get passed to the configuration tool if your into that sort of thing.
Or... I could copy the app bundle with my mouse. Hmmm.
There are also methods to use versions that are newer of older than the latest stable
Are any of those methods easier than just, you know, copying an app bundle? You see where I'm going here?
Still amounts to a straight jacket. You get latest software, but you still are forced to run whatever happens to be in the repo and what happens to be current for you your distribution and it installs only where the package dictates that it be installed. Also, you are forced to update things you may not want to because that's what your latest and greatest software happened to built against. It is all so tightly coupled. Overall it is just a bad user experience for anyone who may not want to do things exactly as the package maintainers think you should.
But the "latest and greatest" aspect is just part of it. I want to be able to install any version I want/need, sometimes with two different versions side by side (Firefox 3 and FIrefox 4, for example). Also, I like being able to run an app from wherever I want. Hell, on OS X, I can launch most apps right from the compressed disk image it came in. I do that kind of things with little games I want to try out or whatever. Or I can carry an app around on a USB stick. It is just so handy, especially if you have to work on other Macs. I would never ever go back to centralized package management. As an end user, centralized package mangement actually gives me very little other than a single place to look for apps. But who cares? The Google is strong with me.