The problem is that you still cannot eliminate human factors. Humans are fickle, and may have changed their minds between telling you what they want and when your engineering team finishes the feature. Existing customers may be able to tell you clearly which bugs are important to them, but they can't necessarily tell you what new features are needed to attract new customers. There is still going to be a level of guesswork between the customer and the engineer, what changes is who (customer or marketing) looks at the crystal ball.
The "bells and whistles" may indeed be useless to an old customer who just wants a better (faster, fewer bugs) version of the one he bought, but what does a potential customer (who didn't see enough in your old feature set to put down money) want? Also, are you really sure the customer will not be stolen from you by a competitor with bells and whistles of their own, while you're fixing those old bugs?
Do you actually sit there--maybe even in the U.S.--writing pretty much whatever you want on whatever you topic you want, and then claim that the U.S. and Communist China are "the same"?
I wrote nothing of the sort, so don't put words in my mouth. I live in the US, which is a relatively free country. I am merely questioning the point of comparing human rights records in the US to those in China, instead of another country that is actually free. It should surprise and flatter no one that the US wins!
It's like saying you're a better person than Saddam Hussein. Very likely true, but so what?
Microsoft won because they were COMPATIBLE. Pure and simple. Windows 3.1 killed all the competition at the time because it was the most compatible with DOS.
No, I don't think that's true. The competitors of the day, such as GEM and GEOS, were both very compatible with MS-DOS. In fact, GEM was something you ran from DOS and quit back to DOS. One key difference among the competitors for a windowing system was applications. GEM had Ventura Publisher, which was eventually defeated by the Windows/PageMaker pair, and when Lotus 1-2-3 and the other big names signed on (nevermind Excel and Word), there was no more hope for the Windows competitors. At that time, Windows could not even really run DOS in a window properly, and was not particularly fast or stable.
Each competitor probably fell because of a different reason, but I don't believe compatibility with DOS was the major issue for any of them. I believe the biggest factor was who the heavyweight software developers (one of which is Microsoft) chose to support.
That's like saying the Olympics aren't a real contest because they only test the prowess of athletes, not their ability to tidy up the locker room after use, their politeness towards other clients at the gym or how nice their outfits look on TV.
No, that's like saying the Olympics isn't a real contest of athletics because you're only testing how fast they can run 100 meters. The results don't show who was fastest at 10 meters, 50 meters, or who would be fastest at 150 or 1,000 meters. Recognizing this shortcoming, the Decathlon adds up the scores from multiple events to find the best all-around track and field athlete.
A programming contest is the equivalent of a single track and field event. There's nothing wrong with that, but we have to be careful what conclusions we draw from its results.
Many complaints I hear about moving over to unix like systems is the filesystem hierarchy. I think for a really easy-for-noobs and for grabbing windows users, I'd like to see a patched distro where/dev,/proc,/sys etc are moved to (for example)/system.
Check out MacOS X sometime. If you use the Finder (the GUI), you see your drives (technically, a partition on a drive, but they are displayed with a drive icon). Clicking on the main one shows you four directories: Applications, Library, System, and Users. However, if you bring up the terminal and cd to the root directory, you see all the other Unixy directories, along with the four named above.
Thus, newbies who don't bring up the terminal never see the Unix directories, people who want to can still do it, and software don't have to be patched.
Re:Dual boot? How about virtualization, too!
on
Going To Boot Camp
·
· Score: 1
If it lacks access to common hardware like sound and the CD, then it's not even code complete, much less beta quality. I'm guessing this is meant to ride on the waves of the Boot Camp announcement to gain recognition, but it is certainly a product I will be watching closely.
Everything can be quantified. If not, Vegas would be out of business. They create odds on everything from people dying to if construction projects will be done on time.
Odds are not the same as statistics. For example, two baseball teams may have identical records, which gives them about even probability of winning against each other, but the odds in actual betting can be quite different. Odds involve perception, and can be created even when no valid statistical model exists, because most people don't go to Vegas to practice their knowledge of probability.
Secondly, there are valid statistics on human lifespans. The average age at death in the First World is somewhere around 76 years old, higher for females, lower for smokers, etc.
As for this case, I can think of one valid hypothetical statistical model. If TiVo had been sued for identical claims in 100 independent jurisdictions and won 70 of those cases, then you could say that they have a 70% chance of winning the next one. As it stands, "70%" is meaningless.
What exactly is a "70% chance of winning"? A court case is not probabilistic, and the likelihood of TiVo winning cannot be quantified. It's one thing for the Washington Post to be dumb, but News for Nerds should not perpetuate it.
So you distribute that key via assymetric encryption, very soon before you send the actual message. That narrows the keyspace a bit, but means that if the attacker doesn't have the computing power to brute-force the assymetric encryption between the time that the key is sent, and the time that the quasar is monitored, that the attacker has failed.
I start monitoring as many quasars as I can the moment I intercept the key message. That way, when I finally decode the key message I can also read the actual message. The secrecy of your message then depends on whether my choices of quasars get lucky, which is not nearly as good as a real one-time pad.
An Apple based cellphone will only appeal to a small market segment and thus could not dominate the way iPod did.
Sure, but you forget that Apple has a profitable computer business that appeals only to a small market segment. The iPod itself was certainly not dominant for a good while, entering the market at the very top end of price ranges.
Unless Apple dominates a market like with the iPod, Apple typically struggles (like with 30 years of computers).
Is there a company that doesn't struggle when it's not dominating the market? The companies you mentioned - Nokia, Motorola, Samsung, Sony - do they not struggle in the cell phone market?
Upgrade! When you do an upgrade you upgrade everything! When you upgrade KDE 3.5.1 to 3.5.2, you upgrade the libraries AND the applications.
One of us is severely confused. On the one hand, you wrote "You're talking about every KDE version since 1.0!" which sounds like I need to have every version of KDE in RAM, which is indeed a pain. Now you're saying I upgrade the apps as well, to minimize on the RAM cost.
Let's rewind a bit. Shared libraries only make sense when most or all of our apps are using the same version, because otherwise they don't get to share any code. However, to get all the apps to use the same version when only one of them really needs the newer one, you have to upgrade them all, which risks bringing in bugs.
Thus, our choices are to waste memory or risk bugs. Compared to the MacOS X approach, which minimizes memory waste by having a large set of core libraries, and minimizes the risk of bugs to unrelated apps by keeping the core libraries relatively constant, how is the "package manager" superior?
You're talking about every KDE version since 1.0! And not just the KDE libraries, but all the Qt and miscellaneous support libraries as well.
So? Why should I care as an end user? If I need more RAM, they're about $100 per GB. If I need more disk space, they're about $1 per GB. The question here is whether binary application bundles are useful, not a laundry list of pseudo-technical excuses why they can't be done. (Also, if there are so many versions of everything to talk about, how exactly is shared libraries saving me any RAM compared to the MacOS approach?)
The fact is, distro builders have failed to either backport bleeding-edge apps to the older libraries they use, or port older apps to bleeding-edge libraries. I use "failed" very very lightly here, because many distro builders are volunteers who make little or no money, and there are just so many libraries and apps. I'm just pointing out that either port can result in minimizing the duplication of libraries in the system, which then makes binary application bundles more feasible.
Now, maybe such a distro won't sell, which would be an entirely valid reason to not do it. But so far you've been trying to put up technical hurdles that somebody has clearly worked through.
Not being a free operating system, it doesn't have to worry about several dozen fundamental third party system libraries.
Free as in speech or beer? If the former, MacOS is built on top of BSD, so I'll have to disagree that it makes any difference. If the latter, many Linux distros cost money, and they still don't do this.
That's because OSX is a closed system, whose "fundamental" libraries are defined by Apple. Linux and BSD cannot do this because they are open. What do you do if the system has all the KDE 3.4 libraries installed, but you want to run a KDE 3.5.1 application?
Until the relevant people figure out that this is not the user's problem, it will remain something that makes Linux harder to use. Why should they care if they have enough disk space and RAM to hold both versions of KDE?
Or worse yet, how do you upgrade without a package manager when KDE 4.0 comes out?
You mean like when OS X 10.1, 10.2, 10.3, and 10.4 came out?
That's why app bundles won't work under free Linux and BSD systems.
That's partly why I got tired of maintaining my Linux box, even though I started using Linux around 1992.
But it won't matter because package managers are a better solution.
I just explained the technical reasons why massively shared libraries are bad for desktop, end-user applications.
Even though the post I was responding to was phrased poorly, I understood - I was talking about wanting to use something not in the OS X core libraries.
If you understood, then your choice of gtk as example was poor both in its size and popularity.
Installing an app in the stable version of Debian is less likely to break another app then under any other OS.
Only because of diligent work by volunteers, not because of inherent technical merit. As I have explained, using shared libraries increases the risk of bringing bugs far beyond what you think you're updating, and that's a bad thing from a technical perspective. Human diligence mitigates that somewhat, but don't hide behind that.
I take your point, but its a trade-off easier to make when you control as much as Apple does - and its not a trade-off that everyone is willing to (or able to) make.
Not everyone, sure. But how many Linux users still really care about disk space or RAM for end user applications? More importantly, what Linux distro gives you the choice to ship binary application bundles for the users that are willing and able to make that trade-off?
They're sound business tactics, but that doesn't necessarily mean they're right or make sense. Leveraging a monopoly is also a very sound business tactic.
You might have a point if your analogy wasn't so awful. Suing your competitor for infringing on your patents happens to be entirely legal.
So - you install a version of the gimp - you get a copy of gtk, install eog - you get a copy of gtk, install.... well I think you get my point.
No, the post you were responding to phrased it poorly.
What is inherent in a MacOS X version (say, 10.4.5) already covers most of what an application might need, including Apple's "equivalent" of gtk. Go browse the Apple Developer docs sometime, and you'll see a much richer set of libraries that come with the OS by default. Thus, if an application wants an "obscure" library and decides to bundle it, the cost to the system is minimal even if another app has the same one. The MacOS bundle is not at all the equivalent of statically linking against gtk and other "fundamental" libraries in Linux.
On the Mac you have scour the web deciding Free or Paid, [...] Once you've downloaded it, things are pretty sweet, but finding the download can be a PITA.
Have you ever heard of versiontracker?
On the linux box (I am going to choose Debian as I'm familiar with it). Fire up synaptic from the gnome menu. Search for barcode. Two results returned. Both of these programs I know to be free of trojans, compatable with my system & configured for it. To install, I double click.
Which can bring in new versions of a library, which in turn brings in new versions of another application, which may be broken in some way (like any app can be). In other words, installing one app in Linux can effectively break another one. You pay for the isolation available to MacOS X apps in the form of disk space and RAM, both of which can be relatively cheap depending on what you're doing with your computer.
Because they don't own the song or the movie, and because the owner says so, and because we expect even the poor to obey laws at least when survival is not in question.
it costs next to nothing to make a copy so selling it at next to nothing + a little bit and you'll get more money back than the poor buying a pirated copy
I completely agree with you, but that doesn't give any poor person the moral right to infringe on the copyright of a song or a movie. We're discussing orthogonal issues here.
I think it's sort of implied that when you license code under the GPL, you have set it "free". What this means is that the code is no longer really yours, it belongs to the collective pool of free software, from which anyone may draw freely.
No, you still retain copyright and ownership. That means if a company approaches you and says they'll pay you some money in exchange for not having to open source their product based on your original code, you are free to make that deal. No other person who got the code via GPL has this right.
A twenty percent tax credit is hardly a "subsidy".
Anything that means I pay a bigger share of taxes than somebody who earns the same amount is a subsidy. Why do you think I should be required to shoulder a bigger share of taxes compared to somebody who makes game mods? (Of course, I would be a bit surprised if a finalized law actually would permit the credit to game mods, so this is really just an academic discussion.)
Slashdot Mirror
The "bells and whistles" may indeed be useless to an old customer who just wants a better (faster, fewer bugs) version of the one he bought, but what does a potential customer (who didn't see enough in your old feature set to put down money) want? Also, are you really sure the customer will not be stolen from you by a competitor with bells and whistles of their own, while you're fixing those old bugs?
I wrote nothing of the sort, so don't put words in my mouth. I live in the US, which is a relatively free country. I am merely questioning the point of comparing human rights records in the US to those in China, instead of another country that is actually free. It should surprise and flatter no one that the US wins!
It's like saying you're a better person than Saddam Hussein. Very likely true, but so what?
Did you just compare human rights in the US versus the same in China, and congratulate yourself on winning?
Have you noticed how iPods tend to be smaller than their competition, at least at the time of release? That's hardware engineering.
No, I don't think that's true. The competitors of the day, such as GEM and GEOS, were both very compatible with MS-DOS. In fact, GEM was something you ran from DOS and quit back to DOS. One key difference among the competitors for a windowing system was applications. GEM had Ventura Publisher, which was eventually defeated by the Windows/PageMaker pair, and when Lotus 1-2-3 and the other big names signed on (nevermind Excel and Word), there was no more hope for the Windows competitors. At that time, Windows could not even really run DOS in a window properly, and was not particularly fast or stable.
Each competitor probably fell because of a different reason, but I don't believe compatibility with DOS was the major issue for any of them. I believe the biggest factor was who the heavyweight software developers (one of which is Microsoft) chose to support.
No, that's like saying the Olympics isn't a real contest of athletics because you're only testing how fast they can run 100 meters. The results don't show who was fastest at 10 meters, 50 meters, or who would be fastest at 150 or 1,000 meters. Recognizing this shortcoming, the Decathlon adds up the scores from multiple events to find the best all-around track and field athlete.
A programming contest is the equivalent of a single track and field event. There's nothing wrong with that, but we have to be careful what conclusions we draw from its results.
I don't think I've ever heard anybody complain that Microsoft doesn't ship Windows for the Mac.
Check out MacOS X sometime. If you use the Finder (the GUI), you see your drives (technically, a partition on a drive, but they are displayed with a drive icon). Clicking on the main one shows you four directories: Applications, Library, System, and Users. However, if you bring up the terminal and cd to the root directory, you see all the other Unixy directories, along with the four named above.
Thus, newbies who don't bring up the terminal never see the Unix directories, people who want to can still do it, and software don't have to be patched.
If it lacks access to common hardware like sound and the CD, then it's not even code complete, much less beta quality. I'm guessing this is meant to ride on the waves of the Boot Camp announcement to gain recognition, but it is certainly a product I will be watching closely.
To connect to the optic nerve? No, thanks, I think I'll use a goggle.
I stand corrected! However, I'm still uncomfortable with people putting numbers where no rigorous analysis appears to have taken place.
Odds are not the same as statistics. For example, two baseball teams may have identical records, which gives them about even probability of winning against each other, but the odds in actual betting can be quite different. Odds involve perception, and can be created even when no valid statistical model exists, because most people don't go to Vegas to practice their knowledge of probability.
Secondly, there are valid statistics on human lifespans. The average age at death in the First World is somewhere around 76 years old, higher for females, lower for smokers, etc.
As for this case, I can think of one valid hypothetical statistical model. If TiVo had been sued for identical claims in 100 independent jurisdictions and won 70 of those cases, then you could say that they have a 70% chance of winning the next one. As it stands, "70%" is meaningless.
What exactly is a "70% chance of winning"? A court case is not probabilistic, and the likelihood of TiVo winning cannot be quantified. It's one thing for the Washington Post to be dumb, but News for Nerds should not perpetuate it.
I start monitoring as many quasars as I can the moment I intercept the key message. That way, when I finally decode the key message I can also read the actual message. The secrecy of your message then depends on whether my choices of quasars get lucky, which is not nearly as good as a real one-time pad.
Sure, but you forget that Apple has a profitable computer business that appeals only to a small market segment. The iPod itself was certainly not dominant for a good while, entering the market at the very top end of price ranges.
Unless Apple dominates a market like with the iPod, Apple typically struggles (like with 30 years of computers).
Is there a company that doesn't struggle when it's not dominating the market? The companies you mentioned - Nokia, Motorola, Samsung, Sony - do they not struggle in the cell phone market?
One of us is severely confused. On the one hand, you wrote "You're talking about every KDE version since 1.0!" which sounds like I need to have every version of KDE in RAM, which is indeed a pain. Now you're saying I upgrade the apps as well, to minimize on the RAM cost.
Let's rewind a bit. Shared libraries only make sense when most or all of our apps are using the same version, because otherwise they don't get to share any code. However, to get all the apps to use the same version when only one of them really needs the newer one, you have to upgrade them all, which risks bringing in bugs.
Thus, our choices are to waste memory or risk bugs. Compared to the MacOS X approach, which minimizes memory waste by having a large set of core libraries, and minimizes the risk of bugs to unrelated apps by keeping the core libraries relatively constant, how is the "package manager" superior?
So? Why should I care as an end user? If I need more RAM, they're about $100 per GB. If I need more disk space, they're about $1 per GB. The question here is whether binary application bundles are useful, not a laundry list of pseudo-technical excuses why they can't be done. (Also, if there are so many versions of everything to talk about, how exactly is shared libraries saving me any RAM compared to the MacOS approach?)
The fact is, distro builders have failed to either backport bleeding-edge apps to the older libraries they use, or port older apps to bleeding-edge libraries. I use "failed" very very lightly here, because many distro builders are volunteers who make little or no money, and there are just so many libraries and apps. I'm just pointing out that either port can result in minimizing the duplication of libraries in the system, which then makes binary application bundles more feasible.
Now, maybe such a distro won't sell, which would be an entirely valid reason to not do it. But so far you've been trying to put up technical hurdles that somebody has clearly worked through.
Not being a free operating system, it doesn't have to worry about several dozen fundamental third party system libraries.
Free as in speech or beer? If the former, MacOS is built on top of BSD, so I'll have to disagree that it makes any difference. If the latter, many Linux distros cost money, and they still don't do this.
So educate us.
Until the relevant people figure out that this is not the user's problem, it will remain something that makes Linux harder to use. Why should they care if they have enough disk space and RAM to hold both versions of KDE?
Or worse yet, how do you upgrade without a package manager when KDE 4.0 comes out?
You mean like when OS X 10.1, 10.2, 10.3, and 10.4 came out?
That's why app bundles won't work under free Linux and BSD systems.
That's partly why I got tired of maintaining my Linux box, even though I started using Linux around 1992.
But it won't matter because package managers are a better solution.
I just explained the technical reasons why massively shared libraries are bad for desktop, end-user applications.
If you understood, then your choice of gtk as example was poor both in its size and popularity.
Installing an app in the stable version of Debian is less likely to break another app then under any other OS.
Only because of diligent work by volunteers, not because of inherent technical merit. As I have explained, using shared libraries increases the risk of bringing bugs far beyond what you think you're updating, and that's a bad thing from a technical perspective. Human diligence mitigates that somewhat, but don't hide behind that.
I take your point, but its a trade-off easier to make when you control as much as Apple does - and its not a trade-off that everyone is willing to (or able to) make.
Not everyone, sure. But how many Linux users still really care about disk space or RAM for end user applications? More importantly, what Linux distro gives you the choice to ship binary application bundles for the users that are willing and able to make that trade-off?
You might have a point if your analogy wasn't so awful. Suing your competitor for infringing on your patents happens to be entirely legal.
No, the post you were responding to phrased it poorly.
What is inherent in a MacOS X version (say, 10.4.5) already covers most of what an application might need, including Apple's "equivalent" of gtk. Go browse the Apple Developer docs sometime, and you'll see a much richer set of libraries that come with the OS by default. Thus, if an application wants an "obscure" library and decides to bundle it, the cost to the system is minimal even if another app has the same one. The MacOS bundle is not at all the equivalent of statically linking against gtk and other "fundamental" libraries in Linux.
On the Mac you have scour the web deciding Free or Paid, [...] Once you've downloaded it, things are pretty sweet, but finding the download can be a PITA.
Have you ever heard of versiontracker?
On the linux box (I am going to choose Debian as I'm familiar with it). Fire up synaptic from the gnome menu. Search for barcode. Two results returned. Both of these programs I know to be free of trojans, compatable with my system & configured for it. To install, I double click.
Which can bring in new versions of a library, which in turn brings in new versions of another application, which may be broken in some way (like any app can be). In other words, installing one app in Linux can effectively break another one. You pay for the isolation available to MacOS X apps in the form of disk space and RAM, both of which can be relatively cheap depending on what you're doing with your computer.
Because they don't own the song or the movie, and because the owner says so, and because we expect even the poor to obey laws at least when survival is not in question.
it costs next to nothing to make a copy so selling it at next to nothing + a little bit and you'll get more money back than the poor buying a pirated copy
I completely agree with you, but that doesn't give any poor person the moral right to infringe on the copyright of a song or a movie. We're discussing orthogonal issues here.
No, you still retain copyright and ownership. That means if a company approaches you and says they'll pay you some money in exchange for not having to open source their product based on your original code, you are free to make that deal. No other person who got the code via GPL has this right.
Anything that means I pay a bigger share of taxes than somebody who earns the same amount is a subsidy. Why do you think I should be required to shoulder a bigger share of taxes compared to somebody who makes game mods? (Of course, I would be a bit surprised if a finalized law actually would permit the credit to game mods, so this is really just an academic discussion.)