The tip and the eraser of a pencil are identifiably unique; different shape, texture, configuration, and color. Experiment and you will see that they perform two functions; one marks, and the other rubs out.
A mouse with two buttons does not have this trait, especially across different applications and contexts. The multiple mice buttons, as I said earlier, are not identifiably unique. In some situations you will double click the far left button, or single click it. In other situations you will single click the far right button, yet training would imply that if the buttons are indistinguishable, where then do you double click the far right button?
And what is the middle button for?
No, you bring in games: That is a specific application, and I'm sorry, but an OS is not a game. If you play a game, you need the tools to play the game with, whether it be a joypad, joystick, pedals and throttle, or a multiple button *mouse*. That means it's irrelevant to the discussion because Apple is not selling you a game, it is selling you an OS and a set of Applications that come with the OS.
The *logic* to why a single button mouse is better is design. If you design an OS for two buttons, then a single button mouse by definition is inferior. If you design an OS for a single button, then a single button mouse is sufficient (my prior argument), while a multiple button mouse is superior (again my prior argument). The only reason a single button mouse can be considered superior is because the *OS* is designed to make it superior; it's not the mouse or the number of buttons, it's the user interface!
The same way that a game that only uses 3 of the 4 cardinal directions on a joypad means that a joypad with 4 directions, 6 directions, or 12 directions is (while not pointless), not immediately useful.
Think of it that way; a single button is sufficient if designed that way.
Each key on a keyboard is unique, with a different symbol, and thus a different use.
On a single button mouse, a single click is unique, and a double click is unique, as is a click and drag or a click and hold.
With two buttons, then there's a question: Which button to use in any situation? With three buttons, you've also got to worry about two button combos (keyboards have combos!)
So in a sense, it's just less training. The Mac OS is designed to be sufficient with a single mouse button, and every additional button and scrollwheel is acceleration.
The Windows OS is *not* designed to be sufficient with a single mouse button. Rather, it's extremely inconvenient to use only a single mouse button.
On the *flip* side, the Mac has not traditionally been designed to be run mouseless (OS X may be more so, but I haven't tested that capability), while Windows has been designed from the ground up to be navigable without a mouse. Not terribly pretty, but it works.
So the bitching about a single mouse button is wasted energy; if you're using a Mac, you don't need more, though you are certainly welcome to use more if you want it, while on Windows (and Linux) it's just different, not worse, not better.
As long as you understand that research and knowledge cannot be measured until they have fruited; how studies of dead frogs a hundred years ago give us computers now, and how studies of multiple languages and cultures now give us ??? a hundred years later. Indifference is okay, as long as it's not a hostile indifference:)
Sure, isn't that the whole point of the linguists and scientists who want to capture all these dying languages?
My point was to reply to the guy who thought it was a waste; it isn't a waste, there is value, and there is insight if you want to look, and some people want to look. No one is stopping people from dying or moving to suburbia; it's more like people are stopping scientists from studying by saying "That's worthless, why are you trying to study a dead language?"
That is entirely why differing languages are important. Different languages capture *differences* in culture, and differences in mindset.
These differences explore the breadth and depth of what it is to be human.
So different people having different opinions are good, and therefore having different cultures with different worldviews are good; language is just one part of that equation.
As I explain in another post: Capturing different languages helps to capture different cultures, and differences in culture help to teach us how different and similar people are, and how the brain works.
We will be losing anthropological information, something we will want 1000 years from now, and something we *know* we will want. Think of all the old lost civilizations we study, and think of the fact that we are watching the same thing happen in front of our very eyes.
You're using the definition of children in a circular manner.
My postulate was that when a child was treated as a non child they would act like a non child, and when treated like a child they thusly acted like one.
You saying "But the are CHILDREN" doesn't refute my logic, just reinforce it.
They aren't responsible, mature, or knowledgable because they don't have the training, something we both agree on; what we disagree is how to instill this into them. You believe in discipline and a structured military-like process.
I believe in reciprocity and templating; that they act the way they are expected to act (because people are normally and generally social creatures, and children are really just small people), and that they act in ways that reward them and avoid acting in ways that punish them.
In the extreme I will concede that a military training fulfills all of the above, but I don't yet believe that the military method is necessary; it is definitely sufficient, but I am not yet convinced it is necessary.
About your view on how to treat children, but I can at least say I don't agree with the way I see many parents treating children right now.
A lot of parents, and adults in general, treat kids like little monsters, pets, or stupid people.
They may not have the experience or the total understanding, but I don't see why they shouldn't be treated as *people*.
The kids that I've seen treated as people tend to act like them; responsible, mature, knowledgable, and intelligent. The ones who are treated like monsters or rejects or annoyances, they act like them.
So if you think kids should be treated as a subordinate, I guess that's your call for your kids. I think I'm gonna treat my kids like people, from day 1, with needs, demands, desires, and the ability to reason, even if they lack the vocabulary or eloquence that comes with study and experience.
If you were truly gonna invest in a high bandwidth digital soundsystem, shouldn't the system be doing something like *firewire* sound?
That way you can do pure digital (and no noise) from source (CD or DVD or whatnot) to the speaker (2, 2.1, 3.1, 5.1, 6.1, etc) system?
And when you phrase it that way, doesn't that immediately mean the sound card becomes irrelevant, unless it transforms into the equivalent of a video card, and does digital mixing, resampling, effects, and transforms, while it's left to the speakers to do DAC?
Yes, but there's a decidely different quality between tree farms and old growth forests.
There's the nature of ecological diversity, which aids in preventing disease and forest fires, there's the increase in diversity in the wildlife due to different ecological niches, and thus a similar increase in the aesthetics since the forest is varied and not uniform.
Replanting *works* but it's a solution to an entirely different problem. You'd want, ideally, *both* solutions so both problems get tackled.
Problem: We don't have enough lumber Problem: We don't have enough forest
Replanting does not a forest (yet, at least) make, and saving forest does not satisfy our need for lumber.
So both solutions in some sort of balance seems best.
The point of tree hugging, whether they believe it or understand it or not, is so that 100 years from now, 1000 years from now, 5000 years from now, we've still got trees (to cut down, view, explore, study, whatever).
If you cut down all the good trees now, there won't be any later. Duh!
You got it backwards, it's the Anti-Lucas Gambit. They cancelled, not announced, another version. They're consolidating rather than incrementally updating and releasing.
I just got my PowerBook battery replaced, and this was the situation:
Charge to 100% Unplug At about 70%, the battery would instantaneously go to zero and sleep.
Rinse and repeat.
However, what the article *seems* to describe is batteries not charging fully, or batteries with reduced lifespan.
My new replacement battery has a lifespan of 4 hours and 20 minutes My old spare battery has a lifespan of 3 hours and 40 minutes The dead replaced battery had a lifespan of about 20 minutes before dropping to zero, even though it reported a full charge.
Is this what others are seeing? The 10.2.4 problem doesn't seem to sound like any of these. The new battery is fine, the spare battery just sounds old, and the dead battery sounds like it was broken.
And what are you going to do about it if someone claims that this 'middle ground' has been achieved, and is Mac OS X?
You get your Unix software, your Mac software, and your Windows software. You've got your pretty fonts, you get your 'Out of box experience', you get your IE, you get your Mozilla, you get your Safari...
You get your whole mantra that "simple is more important" than powerful.
As you say, you cannot always select what you consider the perfect tool for the job.
This case, however, is an exception, one of the times where Weed *could* select what he considered the perfect tool for his job.
You have to trust his judgement, since he's the one out in the desert doing his photo-manip stuff (probably), and as such, if his need for a PowerBook will let us win, and the military agrees, it's really out of our hands other than to backseat drive.
The issue here is that we're comparing a 1.8GHz (max) P4 to a 1.0GHz G4, where the speed delta is small enough that cache, ram, Altivec, and code optimization might make a difference. I mean, we don't know how fast the Toughbooks they can get, they may possibly get 1.3GHz P4s, or 1.2GHz P3s, in which case the performance/benefit analysis is much different than if they could get 3.0GHz P4s.
Also note, I think there's ram limitations; a G4 can get 1.0GB, I don't think the Toughbooks can. And a Toughbook actually costs more, to the public, than a G4.
Now, could he have gotten a Dell instead? Sure. However, he is also most comfortable with a Mac; he *knows* he can get his job done with the PowerBook-at this point then, we have to trust him that he is performing his soldierly duties, and performing them well.
Um... well, the fastest Toughbooks out there are like 1.8GHz P4s.
If he's utilizing Altivec optimized code (quite possible), it's quite possible that his 1GHz PowerBook can outperform a 1.8GHz P4.
Which tasks are Altivec optimized? Photoshop, for one. Certain encryption/decryption tasks are another. Certain video tasks, as well.
It's certainly within the realm of likely possibility, given the description that "Weed declined to specify what he does exactly, but said he works with giant satellite and reconnaissance images,"
Sure, a P4 is fast but when you're talking about a 800MHz difference, the other things (like cache, registers, Altivec, pipeline depth, etc) make more of a difference.
Now, if they were talking about 2.4GHz P4Ms or 3.0GHz P4 (desktops), that would be different. However, Toughbooks don't scale that fast (yet).
There's no reason he couldn't be doing imagine manipulation, enhancement, and analysis using Photoshop; so long as the image is in a standard image format (and why not?), it's just pixels and filters.
I mean, even if it's just simple stuff: Overlays of two images taken in different spectrums (IR and visible) Time-lapse animation (multiple layers transformed into an animation, not unlike an animated GIF) Edge detection/feature enhancement Cropping to remove useless data Rotation, perspective, and skewing to transform poorly captured or framed images into more easily understood images Overlay of before/after shots (perhaps using difference blending) Comparison of two different photos with an identical feature (perhaps identifing buildings, known vs unknown, performed again with overlays and blends) Scaling of a photo so a comparison to a similar photo, taken with different settings, can be accomplished Enhancement of a photo to compensate for low light levels (levels, etc) Normalization of a photo (perspective, levels, colors, scale) so comparisons between two different photos can be accomplished
1) But since we aren't talking about any of the other problems (I will glance at your site, since I'm curious), there's nothing to refute or rebutt
2) So you no longer have a problem with code sharing and code reuse? Now you want to argue instead about two different mechanisms for two different problems? Apps are Apps, and Frameworks are Frameworks, and one does not necessarily rely on the other, depending on the direction of transitivity. Applications can be installed by drag and drop, universally, as can Frameworks, but in the situation where you've got multiple of either, or you're installing a different version of a Framework, or you have to do something else that I'm not aware of, you have an installer to automatically and programmatically perform all those tasks because humans make mistakes and programs, at least, will make those mistakes consistently. Or as your FAQ puts it, this allows for dependency checking, among other things.
3) It discourages code sharing on the developer's end, you mean? Use, Mac users like AppFolders (Bundles). Regardless of whether you use a Framework or not, you're supposed to use a Bundle. Even if you don't like the way Mac uses Bundles, a Bundle as a generic container is not a bad idea, wherein disparate related files are maintained in a structured manner.
I'm still reading your FAQ, but I think your objections aren't with Bundles per se, but with placing shared code and libraries in Bundles.
In which case, the solution is to place them in Frameworks (which are themselves a variation of Bundles I do believe).
Then the next problem you have is that this makes installation messy because there are two installation methods: Drag and drop and Installers (which are really wrappers for Drag and drop functionality anyway). The point is that it isn't a mess. If a user opens a CD/DMG/Folder and finds an App, they can double click it, or drag it to/Applications, or ~/Desktop, or where-ever, and that's not confusing. If instead they find an application called Vise Installer, then the name makes it self evident too, and they double click, and follow the installation wizard.
But if you really feel that's a problem, you as a developer can simply choose to be consistent and release everything in an Installer, whether or not the pre/post scripts are necessary, whether dependency checking and Framework installtion is necessary.
There's nothing broken about appfolders, I don't think. It's just a container and a structure. You can throw shared code into it, and therefore make it redundant, or you can create a Framework, and therefore increase efficiency, but it's really in the developer's hands and choice. Are you saying that Apple tells developers to use Bundles instead of Frameworks? Are you saying Users are telling developers to use Bundles instead of Frameworks?
For the common case installers are unnecessary, points over Microsoft's current installation.
For all of your concerns, there is Mastercard. Wait, no, there are mechanisms that exist to deal with all your concerns.
So when an installer is necessary, an installer can be used, with Perl scripts, pre-install scripts, post-install scripts, Frameworks, etc.
So at the *worst*, we are a step above Microsoft (with the Unixish linking and library versioning), and at the best, we are far away simpler and more convenient (ne better) than Microsoft.
I'm not sure why you say only Apple can updated the OS. You mean other people cannot install system critical Frameworks?
Because DirectX would analagously be a Framework in OS X, and a user could place/install it in/Library or/System/Library. What is stopping this mechanism?
Or do you mean things like QT? Or QuickTime? I know I can update Quicktime codecs by placing them in/Library/Quicktime
And then about Code Reuse. You've got two mechanisms in play:
Bundles stifle code reuse, sacrificed in favor of code inclusiveness. Frameworks promote code reuse. In OS X, DirectX would be a Framework. Look in/System/Library/Frameworks or/Library/Frameworks. The user can add and update/Library/Frameworks, and I think with a special sanction, can also add to the/System folder
Modules are as coarse as you wish. I've got Frameworks that are about 260k in size (for ObjC unit testing), up to 2.8mb in size for the SharedMenu framework for Camino/Chimera.
Then the Quicktime codec in/Library/Quicktime is only 276k as well.
You've already got a couple replies regarding this.
Apple gets around the problem of shared DLLs by using the concept of Frameworks.
You can place Frameworks in/System/Library,/Library, or ~/Library (each user has one), depending on whether the Framework should have global, local, or user scope (and security).
So if you want to make QT available to all Apps, you'd put the appropriate QT libraries in/Library, for example. Same with a DivX package. And you could have multiple versions of QT in/Library, assuming whoever created the QT package kept that in mind.
Then there is Bundles, in which local resources are kept local.
So with both systems, you get the inclusive solution so that each application has it's own structure (a Bundle) and the system has it's own library structure (a Framework).
Quicktime, for example, exists as a Framework. So does WebCore and Cocoa.
Slashdot Mirror
Until you start making DV videos and DVDs, which take gbs and gbs of storage each...
Then you'll want a 64 bit (Apple?) machine.
Sorry, but you sound like an idiot.
The tip and the eraser of a pencil are identifiably unique; different shape, texture, configuration, and color. Experiment and you will see that they perform two functions; one marks, and the other rubs out.
A mouse with two buttons does not have this trait, especially across different applications and contexts. The multiple mice buttons, as I said earlier, are not identifiably unique. In some situations you will double click the far left button, or single click it. In other situations you will single click the far right button, yet training would imply that if the buttons are indistinguishable, where then do you double click the far right button?
And what is the middle button for?
No, you bring in games: That is a specific application, and I'm sorry, but an OS is not a game. If you play a game, you need the tools to play the game with, whether it be a joypad, joystick, pedals and throttle, or a multiple button *mouse*. That means it's irrelevant to the discussion because Apple is not selling you a game, it is selling you an OS and a set of Applications that come with the OS.
The *logic* to why a single button mouse is better is design. If you design an OS for two buttons, then a single button mouse by definition is inferior. If you design an OS for a single button, then a single button mouse is sufficient (my prior argument), while a multiple button mouse is superior (again my prior argument). The only reason a single button mouse can be considered superior is because the *OS* is designed to make it superior; it's not the mouse or the number of buttons, it's the user interface!
The same way that a game that only uses 3 of the 4 cardinal directions on a joypad means that a joypad with 4 directions, 6 directions, or 12 directions is (while not pointless), not immediately useful.
Think of it that way; a single button is sufficient if designed that way.
Each key on a keyboard is unique, with a different symbol, and thus a different use.
On a single button mouse, a single click is unique, and a double click is unique, as is a click and drag or a click and hold.
With two buttons, then there's a question: Which button to use in any situation? With three buttons, you've also got to worry about two button combos (keyboards have combos!)
So in a sense, it's just less training. The Mac OS is designed to be sufficient with a single mouse button, and every additional button and scrollwheel is acceleration.
The Windows OS is *not* designed to be sufficient with a single mouse button. Rather, it's extremely inconvenient to use only a single mouse button.
On the *flip* side, the Mac has not traditionally been designed to be run mouseless (OS X may be more so, but I haven't tested that capability), while Windows has been designed from the ground up to be navigable without a mouse. Not terribly pretty, but it works.
So the bitching about a single mouse button is wasted energy; if you're using a Mac, you don't need more, though you are certainly welcome to use more if you want it, while on Windows (and Linux) it's just different, not worse, not better.
Exactly right.
And some gentle soul mentioned Sapir-Whorf. Man, what a doozy!
Sapir-Whorf Theory
As long as you understand that research and knowledge cannot be measured until they have fruited; how studies of dead frogs a hundred years ago give us computers now, and how studies of multiple languages and cultures now give us ??? a hundred years later. Indifference is okay, as long as it's not a hostile indifference :)
What does your view have to do with people studying dead and dying languages?
:)
No one said that these dying languages should be ressurrected or 'frozen'. Common language is good, and so is different language. Both have value.
To illustrate, Chinese language doesn't have tenses. Try and puzzle that out
Sure, isn't that the whole point of the linguists and scientists who want to capture all these dying languages?
My point was to reply to the guy who thought it was a waste; it isn't a waste, there is value, and there is insight if you want to look, and some people want to look. No one is stopping people from dying or moving to suburbia; it's more like people are stopping scientists from studying by saying "That's worthless, why are you trying to study a dead language?"
That is entirely why differing languages are important. Different languages capture *differences* in culture, and differences in mindset.
These differences explore the breadth and depth of what it is to be human.
So different people having different opinions are good, and therefore having different cultures with different worldviews are good; language is just one part of that equation.
No, the loss will be scientific, as well.
As I explain in another post: Capturing different languages helps to capture different cultures, and differences in culture help to teach us how different and similar people are, and how the brain works.
We will be losing anthropological information, something we will want 1000 years from now, and something we *know* we will want. Think of all the old lost civilizations we study, and think of the fact that we are watching the same thing happen in front of our very eyes.
You're using the definition of children in a circular manner.
My postulate was that when a child was treated as a non child they would act like a non child, and when treated like a child they thusly acted like one.
You saying "But the are CHILDREN" doesn't refute my logic, just reinforce it.
They aren't responsible, mature, or knowledgable because they don't have the training, something we both agree on; what we disagree is how to instill this into them. You believe in discipline and a structured military-like process.
I believe in reciprocity and templating; that they act the way they are expected to act (because people are normally and generally social creatures, and children are really just small people), and that they act in ways that reward them and avoid acting in ways that punish them.
In the extreme I will concede that a military training fulfills all of the above, but I don't yet believe that the military method is necessary; it is definitely sufficient, but I am not yet convinced it is necessary.
About your view on how to treat children, but I can at least say I don't agree with the way I see many parents treating children right now.
A lot of parents, and adults in general, treat kids like little monsters, pets, or stupid people.
They may not have the experience or the total understanding, but I don't see why they shouldn't be treated as *people*.
The kids that I've seen treated as people tend to act like them; responsible, mature, knowledgable, and intelligent. The ones who are treated like monsters or rejects or annoyances, they act like them.
So if you think kids should be treated as a subordinate, I guess that's your call for your kids. I think I'm gonna treat my kids like people, from day 1, with needs, demands, desires, and the ability to reason, even if they lack the vocabulary or eloquence that comes with study and experience.
If you were truly gonna invest in a high bandwidth digital soundsystem, shouldn't the system be doing something like *firewire* sound?
That way you can do pure digital (and no noise) from source (CD or DVD or whatnot) to the speaker (2, 2.1, 3.1, 5.1, 6.1, etc) system?
And when you phrase it that way, doesn't that immediately mean the sound card becomes irrelevant, unless it transforms into the equivalent of a video card, and does digital mixing, resampling, effects, and transforms, while it's left to the speakers to do DAC?
Yes, but there's a decidely different quality between tree farms and old growth forests.
There's the nature of ecological diversity, which aids in preventing disease and forest fires, there's the increase in diversity in the wildlife due to different ecological niches, and thus a similar increase in the aesthetics since the forest is varied and not uniform.
Replanting *works* but it's a solution to an entirely different problem. You'd want, ideally, *both* solutions so both problems get tackled.
Problem: We don't have enough lumber
Problem: We don't have enough forest
Replanting does not a forest (yet, at least) make, and saving forest does not satisfy our need for lumber.
So both solutions in some sort of balance seems best.
Oh, come on, I hope you're joking.
The point of tree hugging, whether they believe it or understand it or not, is so that 100 years from now, 1000 years from now, 5000 years from now, we've still got trees (to cut down, view, explore, study, whatever).
If you cut down all the good trees now, there won't be any later. Duh!
You got it backwards, it's the Anti-Lucas Gambit.
They cancelled, not announced, another version. They're consolidating rather than incrementally updating and releasing.
I just got my PowerBook battery replaced, and this was the situation:
Charge to 100%
Unplug
At about 70%, the battery would instantaneously go to zero and sleep.
Rinse and repeat.
However, what the article *seems* to describe is batteries not charging fully, or batteries with reduced lifespan.
My new replacement battery has a lifespan of 4 hours and 20 minutes
My old spare battery has a lifespan of 3 hours and 40 minutes
The dead replaced battery had a lifespan of about 20 minutes before dropping to zero, even though it reported a full charge.
Is this what others are seeing? The 10.2.4 problem doesn't seem to sound like any of these. The new battery is fine, the spare battery just sounds old, and the dead battery sounds like it was broken.
And what are you going to do about it if someone claims that this 'middle ground' has been achieved, and is Mac OS X?
You get your Unix software, your Mac software, and your Windows software. You've got your pretty fonts, you get your 'Out of box experience', you get your IE, you get your Mozilla, you get your Safari...
You get your whole mantra that "simple is more important" than powerful.
As you say, you cannot always select what you consider the perfect tool for the job.
This case, however, is an exception, one of the times where Weed *could* select what he considered the perfect tool for his job.
You have to trust his judgement, since he's the one out in the desert doing his photo-manip stuff (probably), and as such, if his need for a PowerBook will let us win, and the military agrees, it's really out of our hands other than to backseat drive.
The issue here is that we're comparing a 1.8GHz (max) P4 to a 1.0GHz G4, where the speed delta is small enough that cache, ram, Altivec, and code optimization might make a difference. I mean, we don't know how fast the Toughbooks they can get, they may possibly get 1.3GHz P4s, or 1.2GHz P3s, in which case the performance/benefit analysis is much different than if they could get 3.0GHz P4s.
Also note, I think there's ram limitations; a G4 can get 1.0GB, I don't think the Toughbooks can. And a Toughbook actually costs more, to the public, than a G4.
Now, could he have gotten a Dell instead? Sure. However, he is also most comfortable with a Mac; he *knows* he can get his job done with the PowerBook-at this point then, we have to trust him that he is performing his soldierly duties, and performing them well.
Unless the military has special connections :P
Who knows? Perhaps they get them first?
Um... well, the fastest Toughbooks out there are like 1.8GHz P4s.
If he's utilizing Altivec optimized code (quite possible), it's quite possible that his 1GHz PowerBook can outperform a 1.8GHz P4.
Which tasks are Altivec optimized? Photoshop, for one. Certain encryption/decryption tasks are another. Certain video tasks, as well.
It's certainly within the realm of likely possibility, given the description that "Weed declined to specify what he does exactly, but said he works with giant satellite and reconnaissance images,"
Sure, a P4 is fast but when you're talking about a 800MHz difference, the other things (like cache, registers, Altivec, pipeline depth, etc) make more of a difference.
Now, if they were talking about 2.4GHz P4Ms or 3.0GHz P4 (desktops), that would be different. However, Toughbooks don't scale that fast (yet).
There's no reason he couldn't be doing imagine manipulation, enhancement, and analysis using Photoshop; so long as the image is in a standard image format (and why not?), it's just pixels and filters.
I mean, even if it's just simple stuff:
Overlays of two images taken in different spectrums (IR and visible)
Time-lapse animation (multiple layers transformed into an animation, not unlike an animated GIF)
Edge detection/feature enhancement
Cropping to remove useless data
Rotation, perspective, and skewing to transform poorly captured or framed images into more easily understood images
Overlay of before/after shots (perhaps using difference blending)
Comparison of two different photos with an identical feature (perhaps identifing buildings, known vs unknown, performed again with overlays and blends)
Scaling of a photo so a comparison to a similar photo, taken with different settings, can be accomplished
Enhancement of a photo to compensate for low light levels (levels, etc)
Normalization of a photo (perspective, levels, colors, scale) so comparisons between two different photos can be accomplished
All of those are trivial with Photoshop.
1) But since we aren't talking about any of the other problems (I will glance at your site, since I'm curious), there's nothing to refute or rebutt
/Applications, or ~/Desktop, or where-ever, and that's not confusing. If instead they find an application called Vise Installer, then the name makes it self evident too, and they double click, and follow the installation wizard.
2) So you no longer have a problem with code sharing and code reuse? Now you want to argue instead about two different mechanisms for two different problems? Apps are Apps, and Frameworks are Frameworks, and one does not necessarily rely on the other, depending on the direction of transitivity. Applications can be installed by drag and drop, universally, as can Frameworks, but in the situation where you've got multiple of either, or you're installing a different version of a Framework, or you have to do something else that I'm not aware of, you have an installer to automatically and programmatically perform all those tasks because humans make mistakes and programs, at least, will make those mistakes consistently. Or as your FAQ puts it, this allows for dependency checking, among other things.
3) It discourages code sharing on the developer's end, you mean? Use, Mac users like AppFolders (Bundles). Regardless of whether you use a Framework or not, you're supposed to use a Bundle. Even if you don't like the way Mac uses Bundles, a Bundle as a generic container is not a bad idea, wherein disparate related files are maintained in a structured manner.
I'm still reading your FAQ, but I think your objections aren't with Bundles per se, but with placing shared code and libraries in Bundles.
In which case, the solution is to place them in Frameworks (which are themselves a variation of Bundles I do believe).
Then the next problem you have is that this makes installation messy because there are two installation methods: Drag and drop and Installers (which are really wrappers for Drag and drop functionality anyway). The point is that it isn't a mess. If a user opens a CD/DMG/Folder and finds an App, they can double click it, or drag it to
But if you really feel that's a problem, you as a developer can simply choose to be consistent and release everything in an Installer, whether or not the pre/post scripts are necessary, whether dependency checking and Framework installtion is necessary.
There's nothing broken about appfolders, I don't think. It's just a container and a structure. You can throw shared code into it, and therefore make it redundant, or you can create a Framework, and therefore increase efficiency, but it's really in the developer's hands and choice. Are you saying that Apple tells developers to use Bundles instead of Frameworks? Are you saying Users are telling developers to use Bundles instead of Frameworks?
Essentially it boils down to this:
For the common case installers are unnecessary, points over Microsoft's current installation.
For all of your concerns, there is Mastercard. Wait, no, there are mechanisms that exist to deal with all your concerns.
So when an installer is necessary, an installer can be used, with Perl scripts, pre-install scripts, post-install scripts, Frameworks, etc.
So at the *worst*, we are a step above Microsoft (with the Unixish linking and library versioning), and at the best, we are far away simpler and more convenient (ne better) than Microsoft.
We being those who can afford Macs.
I'm not sure why you say only Apple can updated the OS. You mean other people cannot install system critical Frameworks?
/Library or /System/Library. What is stopping this mechanism?
/Library/Quicktime
/System/Library/Frameworks or /Library/Frameworks. The user can add and update /Library/Frameworks, and I think with a special sanction, can also add to the /System folder
/Library/Quicktime is only 276k as well.
/usr/lib
Because DirectX would analagously be a Framework in OS X, and a user could place/install it in
Or do you mean things like QT? Or QuickTime? I know I can update Quicktime codecs by placing them in
And then about Code Reuse. You've got two mechanisms in play:
Bundles stifle code reuse, sacrificed in favor of code inclusiveness.
Frameworks promote code reuse. In OS X, DirectX would be a Framework. Look in
Modules are as coarse as you wish. I've got Frameworks that are about 260k in size (for ObjC unit testing), up to 2.8mb in size for the SharedMenu framework for Camino/Chimera.
Then the Quicktime codec in
And of course, I still have
You've already got a couple replies regarding this.
/System/Library, /Library, or ~/Library (each user has one), depending on whether the Framework should have global, local, or user scope (and security).
/Library, for example. Same with a DivX package. And you could have multiple versions of QT in /Library, assuming whoever created the QT package kept that in mind.
Apple gets around the problem of shared DLLs by using the concept of Frameworks.
You can place Frameworks in
So if you want to make QT available to all Apps, you'd put the appropriate QT libraries in
Then there is Bundles, in which local resources are kept local.
So with both systems, you get the inclusive solution so that each application has it's own structure (a Bundle) and the system has it's own library structure (a Framework).
Quicktime, for example, exists as a Framework. So does WebCore and Cocoa.