As we all know from the intergalactic truck-stop, there are three forms: Regular Matter. Dark Matter and Wassa Matter. (Don't eat the bathroom egg sandwiches.)
Smooth scrolling is one of many fixes needed for 4. In general I agree with your notions although I don't think that the overlapping window model is bankrupt.
I'll be the first to admit that I'm not knowledgeable about power adapters to a degree that I can conclude who is and who isn't. With that out of the way, I stand by everything else in my comment. I reckon the bulk of the article was informative (even if not directly insightful or original research) and brought up a new angle of an ongoing "Apple's hardware has been problematic" arc since about January.
I'm admitting to being tired of the kind of answer your original comment brought up - roughly, "this is on some dude's web site, which according to my lookup chart means that it's not worth reading" - and I wish people could bring up real actual problems with the article instead of deriding it by proxy. There's a big difference between saying "we should post everything on personal web sites" and "we should not post anything from personal web sites". An electrical engineer posted a few threads away about how the logic in the article is busted, and that I can relate to.
I don't approve of guilt-by-association. Nowhere does anyone claim this is science, not even the guy himself.
If someone knowledgeable about power adapters takes one apart and sees that it's horribly constructed and writes about it, that text means the same thing no matter where it's published - a napkin, his personal web site, the local rag, Playboy, NY Times or Popular Science.
The text itself doesn't change no matter where it's published. If it's crap and it's in Popular Science (although thanks to peer review, this doesn't happen very often), it's still crap. If it's genius and it's on some dude's napkin, it's still genius.
I personally enjoyed the article and I was glad to have been shown it. I don't care about it being on a personal web site, it's the text I'm interested in. If you don't approve of the article itself, why don't you just say so?
Personally, I think that everyone with their head on straight knows at this point that copying is something that everyone does on this level. You may be one of the holdouts to think that this is bad, but if it was bad and nobody would do it, where would we be today? Just taking the Mother of All Demos, nobody except the by Engelbart designated innovators would be given access to the stuff presented there: the mouse ("Bug"), video conferencing, email, hypertext... When people are bitching about "copying" or "stealing", I don't think they consider the alternative and how much more crappy it is.
There's also a thing as overdoing it and not inventing enough on your own, but I don't think any major vendor (Microsoft, Apple, Sun, Red Hat, Ubuntu, and so on) are doing that as of today. Apple's poking fun at Vista to rally the troops (it's a developer conference!) and to twiddle Microsoft's nose once more while they have the chance - it's marketing, not the universal truth.
I also think that 10.5 is misunderstood at this level. Take Time Machine: even if we discount the smoke-and-mirrors display of the thing or the fact that the OS helps you backup efficiently with a non-boot volume and four UI controls in its preference pane, the big innovation here is really that you can restore not only one file but that there are *built-in hooks* for "here's this old file and here's this new file" which means that you can cherry-pick old items from old database files. This is something very neat and very useful that in 99 cases of 100 couldn't be done before without resorting to poking and prodding the database files themselves; and now that it's built-in to some of Apple's apps, it's not only going to be tremendously useful there but there's going to be an onus on third-party developers to provide support for this, which means a better user experience for everyone.
As a developer, I'm very excited about 10.5. There's all sorts of new APIs, the old APIs have been extended in better ways, and the developer tools have reportedly gotten the biggest facelift since, well, *ever*. Xcode 3.0 may even trump the step from OS X's Project Builder to Xcode 1.0, and the Interface Builder has finally received some much-needed love. Gruber's right: "Complaining that the announcements at WWDC only appealed to 'the geeks' is like going to a rock concert and complaining that all they did was play loud music."
That sounds like it'd make sense, but it doesn't describe why normal applications and, say, drivers work differently (which is why I'm betting that Windows has some sort of "portal executable").
Sure, but if I recall correctly, the apps run under some sort of emulation app. If I get the material from Apple straight, the big idea with 64 on ppc64 and x86_64 on OS X is that those shenanigans aren't needed, and so any kind of executable - be they a driver or an application - would run natively.
Every 32-bit program will run on the respective 64-bit architecture (ppc will run on ppc64 and i386 will run on x86_64) directly without any emulation as far as I've heard. If that's true, that eliminates the Windows x64-style issues.
I know most people here (Sweden) who *must* use xx/xx/xx use dd/mm/yy.
I don't see why we can't all just switch to YYYY-mm-dd. You can't confuse it with any existing format. It's logical (they all come in order). It keeps mm/dd/yy's only benefit (as far as I'm concerned; I don't see keeping the year at the back a benefit) by letting months come ahead of days. With AM/PM vs 24 hour, at least there's a valid argument in that people think "five in the afternoon", not "seventeen"; with YYYY-mm-dd there's no such factor.
The FSF has always had restrictions on their licenses in order to reinforce and make possible other freedoms. The FSF has chosen to use this to protect the *end user*, not *whoever uses the license on their product*, in order to try to make the Free Software utopia come true. This utopia requires not having things like DRM as it runs counter to the entire principle.
I'm a BSD/MIT license man myself so by definition I don't agree 100% with the principle and I agree on a personal level with your sentiments, but I also understand the way the GPL works towards the idea of Free Software by using these restrictions in their licenses. If you're here to complain about FSF "start[ing to add] restrictions to its licenses", you're at least a decade late to the party. Applying this to DRM in specific is what's new, not the underlying concept.
There are a small number of differences between each instruction set. Compilers generally produce binaries that target both AMD64 and EM64T, making the differences mainly of interest to compiler developers and operating system developers.
So yes, I did read the page, and I did see that there are differences, but some bullets refer to "early xx processors" or begin with "originally". I do admit that they're not identical, but as the quote above says, not to a degree that hinders compiled programs from working on both Intel and AMD processors.
About the indenting issue - I've found one very apparent discrepancy at least in Objective-C.
A line like this: NSDictionary *d = [NSDictionary dictionary];// one-line comment starting directly after the semicolon
causes the line following it to be indented.
A line like this: NSDictionary *d = [NSDictionary dictionary];// one-line comment starting after the semicolon and a space or a tab
does *not* cause the line following it to be indented.
There probably are other holes in this logic but I thought I'd post it since you mentioned indentation specifically. I'm going to go report this on Radar now.
Xcode is a rethink of Project Builder, but I don't think it's a rewrite. I didn't use Project Builder for long on Mac OS X, and I never used it on NeXT, but Xcode is so much faster than Project Builder it's freaky. If you think Xcode is slow, thank your lucky star you never had to use Project Builder on OS X 10.2.:)
I would be very surprised if Xcode didn't include optional support for the new free Intel compilers in its next major version (3.0 with OS X 10.5 "Leopard", probably), but I don't think it'll be entirely replaced since Apple has made a heavy commitment to GCC (including some more or less experimental efforts like garbage collection and LLVM), and more importantly since the Intel compiler doesn't do Objective-C at all. I don't think it's impossible at all to replace the compiler; even if Xcode *is* hardcoded to gcc, which I don't think, just write a shell script that calls the other compiler and translates the command line options as necessary and put that in place. And, after all, if you're doing command line development and see yourself just doing that for the foreseeable future, there's no reason to just use text editors and Make, especially not if that's the way you've traditionally done it.
To answer questions in your comment, you don't tell gcc to make a Universal binary, you build an i386 build and a ppc build and use lipo to make a fat binary, and my Xcode build log tells me that you supply framework search paths or explicit framework names to ld and gcc directly.
Great, now my head hurts.:) Backwards compatibility point gladly supported.
I obviously know about a 2% of what you know regarding actual details here, but my main point was that Apple might have worked into a special case for HFS+ into the very lowest layer - this would make sense since they have to make accommodations for HFS+ support anyway and have a system for supporting file systems that from the looks of it is non-standard (check/System/Library/Filesystems/).
Lesson learned: never mess with someone with an alum.mit.edu email address and a four digit Slashdot ID.:)
There's no such thing, relatively, and in this context, as a "pure Cocoa app", because as far as I know Cocoa builds on Carbon in order to simplify and not repeat very low-level primitives. In this vein, the Cocoa file system primitives (NSFileManager and so on) use either the Carbon library functions or C (POSIX library) functions directly. I have no idea if Apple's implementation of the POSIX layer does this automatically, but it'd make sense.
Well, you're right about the Shift+Command+4 thing. One thing that really lingers about Apple's NIH years is the alternative keyboard layouts (no Print Screen, for one). Far from all Windows laptops have separate pgup/pgdn/home/end keys, but you're right that they *should* have. The zing about ctrl+click being less obvious than right-click is well-deserved, and I've propagated in favor of at least a two-button trackpad *option* on the laptops. I have a new MacBook which has "two fingers on trackpad + click means right-click", which is still not obvious but is opt-in and far more convenient.
I'm not denying that the shell is the right tool for some jobs, or that "mere mortals" can get stuff done in it too, but it's far from obvious unless you really do know your shit how to even get started, and I don't think anyone would win anything by making the shell more mainstream just so they could leave it outside of the Utilities folder (which I agree is a wrong label, now that you mention it).
The funny thing about the package management is that one of either FInk or DarwinPorts is supposed to work in a Debian-ish manner, and one in a BSD-ish manner (I forget which, and I'm not a wizard here). You're totally right that it sucks how most projects are available on one or the other and many projects are only available in ordinary downloads, however. It's a mess for one reason: there are several systems and Apple hasn't blessed one of them - I'm thinking that once they do, you'll have it a lot easier.
It's not the first time I've heard the Amiga comparison and it's not going to be the last. I can't claim to know a whole lot about Amigas either, but I do know that the whole market is more stable overall now and that Apple has stayed in the top 10 computer companies (by volume) for longer than Amiga's "twilight years" ever lasted, so they're not in any danger. If everything goes to shit it'll still take at least five years to break down the company unless there's an Enron-style crash. The other side of this coin is that Mac OS X 10.5 could cook your breakfast and gas up your car and still attract no potential switchers. A stable market is also often a stagnant market, for better and worse. (Also, if you ask me, Apple caters well to both consumers and professionals in one go, simply by producing different products. This is reason alone why they won't go down in flames that easy. Their business model works and they rake in profits. )
Sure, lots of people use shells, but by a wide margin more people will open, say, iWeb on OS X or Windows Movie Maker on XP before they ever open "Terminal" or "Command Prompt". "Hiding" it in Utilities doesn't do a thing - if you need it, it's there. And contrary to your belief, lots of people legitimately never *need* to use a shell for anything. Sticking tools where people who need them will find them and people who don't need them can blissfully ignore them isn't a good ground to dismiss an OS as "frustratingly bad" in this aspect, unless it's your conviction that every OS should assume that their primary user writes shell scripts or at least is a heavy shell user, something that with the advent of Ubuntu and other more "friendly" alternatives is increasingly hard to claim of even Linux distros anymore.
As an OS X user, I'm finding your claim about bizarre keystrokes to get everyday things done, well, bizarre. Is this about the Command key? Is it about the Home/End keys moving the viewport and not the text cursor? Is it about selections being extended instead of shortened in some cases? Here, we're on the same page, but keystrokes sounds like "shortcut keys" - list five of your bizarre shortcut keys and then we'll talk.
I can't complain about the way you feel shortchanged expecting a bundled OS X command line package manager (because package managers do exist - DarwinPorts and Fink) and more open-source software (Rosetta, the PPC-to-x86 translator, is currently holding back the kernel source release for x86; meanwhile such blockbuster hits as TextEdit (sample code), Chess (GPLed) and the Safari rendering engine WebKit (supposedly to facilitate integration with KDE but probably a smart recruiting move) are already open-source beyond Darwin). I have high hopes for the next revision of OS X. Apple has some genuinely smart *NIX-heads on board who really do get it, even if a small bit of Not-Invented-Here syndrome still lingers in the organization, and 10.4 is so much better than 10.2 (the first version I really tried) in all aspects that it's hard to believe Microsoft was toiling away at Vista in the same time span.
Maybe because introducing a unit that *shouldn't* scale along with the rest of the page (and the rest of the units) would complicate the spec even more? Look at how IE6 implements the pixel unit in conjunction with scaling. It's not consistent and it's not pretty.
An absolute unit does make sure that, say, a sidebar stays a certain width when scaled up, I'll give you that. But it also hinders the sidebar to move beyond stamp-sized in high pixel density displays, which for all I know could have been in the marketplace in droves by now hadn't there still been big software issues with regards to scaling.
I'd say that taking all this into account, it would be a good idea to have a unit that at its baseline occupies one pixel, and it logically follows that it would be called a pixel.
Slashdot Mirror
As we all know from the intergalactic truck-stop, there are three forms: Regular Matter. Dark Matter and Wassa Matter. (Don't eat the bathroom egg sandwiches.)
Smooth scrolling is one of many fixes needed for 4. In general I agree with your notions although I don't think that the overlapping window model is bankrupt.
I'll be the first to admit that I'm not knowledgeable about power adapters to a degree that I can conclude who is and who isn't. With that out of the way, I stand by everything else in my comment. I reckon the bulk of the article was informative (even if not directly insightful or original research) and brought up a new angle of an ongoing "Apple's hardware has been problematic" arc since about January.
I'm admitting to being tired of the kind of answer your original comment brought up - roughly, "this is on some dude's web site, which according to my lookup chart means that it's not worth reading" - and I wish people could bring up real actual problems with the article instead of deriding it by proxy. There's a big difference between saying "we should post everything on personal web sites" and "we should not post anything from personal web sites". An electrical engineer posted a few threads away about how the logic in the article is busted, and that I can relate to.
I don't approve of guilt-by-association. Nowhere does anyone claim this is science, not even the guy himself.
If someone knowledgeable about power adapters takes one apart and sees that it's horribly constructed and writes about it, that text means the same thing no matter where it's published - a napkin, his personal web site, the local rag, Playboy, NY Times or Popular Science.
The text itself doesn't change no matter where it's published. If it's crap and it's in Popular Science (although thanks to peer review, this doesn't happen very often), it's still crap. If it's genius and it's on some dude's napkin, it's still genius.
I personally enjoyed the article and I was glad to have been shown it. I don't care about it being on a personal web site, it's the text I'm interested in. If you don't approve of the article itself, why don't you just say so?
Personally, I think that everyone with their head on straight knows at this point that copying is something that everyone does on this level. You may be one of the holdouts to think that this is bad, but if it was bad and nobody would do it, where would we be today? Just taking the Mother of All Demos, nobody except the by Engelbart designated innovators would be given access to the stuff presented there: the mouse ("Bug"), video conferencing, email, hypertext... When people are bitching about "copying" or "stealing", I don't think they consider the alternative and how much more crappy it is.
There's also a thing as overdoing it and not inventing enough on your own, but I don't think any major vendor (Microsoft, Apple, Sun, Red Hat, Ubuntu, and so on) are doing that as of today. Apple's poking fun at Vista to rally the troops (it's a developer conference!) and to twiddle Microsoft's nose once more while they have the chance - it's marketing, not the universal truth.
I also think that 10.5 is misunderstood at this level. Take Time Machine: even if we discount the smoke-and-mirrors display of the thing or the fact that the OS helps you backup efficiently with a non-boot volume and four UI controls in its preference pane, the big innovation here is really that you can restore not only one file but that there are *built-in hooks* for "here's this old file and here's this new file" which means that you can cherry-pick old items from old database files. This is something very neat and very useful that in 99 cases of 100 couldn't be done before without resorting to poking and prodding the database files themselves; and now that it's built-in to some of Apple's apps, it's not only going to be tremendously useful there but there's going to be an onus on third-party developers to provide support for this, which means a better user experience for everyone.
As a developer, I'm very excited about 10.5. There's all sorts of new APIs, the old APIs have been extended in better ways, and the developer tools have reportedly gotten the biggest facelift since, well, *ever*. Xcode 3.0 may even trump the step from OS X's Project Builder to Xcode 1.0, and the Interface Builder has finally received some much-needed love. Gruber's right: "Complaining that the announcements at WWDC only appealed to 'the geeks' is like going to a rock concert and complaining that all they did was play loud music."
That sounds like it'd make sense, but it doesn't describe why normal applications and, say, drivers work differently (which is why I'm betting that Windows has some sort of "portal executable").
Sure, but if I recall correctly, the apps run under some sort of emulation app. If I get the material from Apple straight, the big idea with 64 on ppc64 and x86_64 on OS X is that those shenanigans aren't needed, and so any kind of executable - be they a driver or an application - would run natively.
Every 32-bit program will run on the respective 64-bit architecture (ppc will run on ppc64 and i386 will run on x86_64) directly without any emulation as far as I've heard. If that's true, that eliminates the Windows x64-style issues.
There was a slide for performance. The G5 was base, the Opteron was 1.6x and the Xeon was 3x. I don't think these are the Xeons of yesteryear anymore.
I know most people here (Sweden) who *must* use xx/xx/xx use dd/mm/yy.
I don't see why we can't all just switch to YYYY-mm-dd. You can't confuse it with any existing format. It's logical (they all come in order). It keeps mm/dd/yy's only benefit (as far as I'm concerned; I don't see keeping the year at the back a benefit) by letting months come ahead of days. With AM/PM vs 24 hour, at least there's a valid argument in that people think "five in the afternoon", not "seventeen"; with YYYY-mm-dd there's no such factor.
Host NASA servers? :)
It's hard to wreck a nice beach. :)
The FSF has always had restrictions on their licenses in order to reinforce and make possible other freedoms. The FSF has chosen to use this to protect the *end user*, not *whoever uses the license on their product*, in order to try to make the Free Software utopia come true. This utopia requires not having things like DRM as it runs counter to the entire principle.
I'm a BSD/MIT license man myself so by definition I don't agree 100% with the principle and I agree on a personal level with your sentiments, but I also understand the way the GPL works towards the idea of Free Software by using these restrictions in their licenses. If you're here to complain about FSF "start[ing to add] restrictions to its licenses", you're at least a decade late to the party. Applying this to DRM in specific is what's new, not the underlying concept.
Two can quote from that page:
So yes, I did read the page, and I did see that there are differences, but some bullets refer to "early xx processors" or begin with "originally". I do admit that they're not identical, but as the quote above says, not to a degree that hinders compiled programs from working on both Intel and AMD processors.
Isn't EM64T exactly AMD64 under a different "hey-look-our-competitors-totally-didn't-invent-th is" name?
The gist of the differences:
Regular install: Copy over DLLs, run scripts to register them.
Image install: Copy over DLLs and the 'registered DLLs' database thing.
About the indenting issue - I've found one very apparent discrepancy at least in Objective-C.
A line like this:
NSDictionary *d = [NSDictionary dictionary];// one-line comment starting directly after the semicolon
causes the line following it to be indented.
A line like this: // one-line comment starting after the semicolon and a space or a tab
NSDictionary *d = [NSDictionary dictionary];
does *not* cause the line following it to be indented.
There probably are other holes in this logic but I thought I'd post it since you mentioned indentation specifically. I'm going to go report this on Radar now.
You already do. You just keep the lens cap on.
Xcode is a rethink of Project Builder, but I don't think it's a rewrite. I didn't use Project Builder for long on Mac OS X, and I never used it on NeXT, but Xcode is so much faster than Project Builder it's freaky. If you think Xcode is slow, thank your lucky star you never had to use Project Builder on OS X 10.2. :)
I would be very surprised if Xcode didn't include optional support for the new free Intel compilers in its next major version (3.0 with OS X 10.5 "Leopard", probably), but I don't think it'll be entirely replaced since Apple has made a heavy commitment to GCC (including some more or less experimental efforts like garbage collection and LLVM), and more importantly since the Intel compiler doesn't do Objective-C at all. I don't think it's impossible at all to replace the compiler; even if Xcode *is* hardcoded to gcc, which I don't think, just write a shell script that calls the other compiler and translates the command line options as necessary and put that in place. And, after all, if you're doing command line development and see yourself just doing that for the foreseeable future, there's no reason to just use text editors and Make, especially not if that's the way you've traditionally done it.
To answer questions in your comment, you don't tell gcc to make a Universal binary, you build an i386 build and a ppc build and use lipo to make a fat binary, and my Xcode build log tells me that you supply framework search paths or explicit framework names to ld and gcc directly.
This is very interesting stuff and I thank you for sharing.
Great, now my head hurts. :) Backwards compatibility point gladly supported.
/System/Library/Filesystems/).
:)
I obviously know about a 2% of what you know regarding actual details here, but my main point was that Apple might have worked into a special case for HFS+ into the very lowest layer - this would make sense since they have to make accommodations for HFS+ support anyway and have a system for supporting file systems that from the looks of it is non-standard (check
Lesson learned: never mess with someone with an alum.mit.edu email address and a four digit Slashdot ID.
There's no such thing, relatively, and in this context, as a "pure Cocoa app", because as far as I know Cocoa builds on Carbon in order to simplify and not repeat very low-level primitives. In this vein, the Cocoa file system primitives (NSFileManager and so on) use either the Carbon library functions or C (POSIX library) functions directly. I have no idea if Apple's implementation of the POSIX layer does this automatically, but it'd make sense.
Well, you're right about the Shift+Command+4 thing. One thing that really lingers about Apple's NIH years is the alternative keyboard layouts (no Print Screen, for one). Far from all Windows laptops have separate pgup/pgdn/home/end keys, but you're right that they *should* have. The zing about ctrl+click being less obvious than right-click is well-deserved, and I've propagated in favor of at least a two-button trackpad *option* on the laptops. I have a new MacBook which has "two fingers on trackpad + click means right-click", which is still not obvious but is opt-in and far more convenient.
I'm not denying that the shell is the right tool for some jobs, or that "mere mortals" can get stuff done in it too, but it's far from obvious unless you really do know your shit how to even get started, and I don't think anyone would win anything by making the shell more mainstream just so they could leave it outside of the Utilities folder (which I agree is a wrong label, now that you mention it).
The funny thing about the package management is that one of either FInk or DarwinPorts is supposed to work in a Debian-ish manner, and one in a BSD-ish manner (I forget which, and I'm not a wizard here). You're totally right that it sucks how most projects are available on one or the other and many projects are only available in ordinary downloads, however. It's a mess for one reason: there are several systems and Apple hasn't blessed one of them - I'm thinking that once they do, you'll have it a lot easier.
It's not the first time I've heard the Amiga comparison and it's not going to be the last. I can't claim to know a whole lot about Amigas either, but I do know that the whole market is more stable overall now and that Apple has stayed in the top 10 computer companies (by volume) for longer than Amiga's "twilight years" ever lasted, so they're not in any danger. If everything goes to shit it'll still take at least five years to break down the company unless there's an Enron-style crash. The other side of this coin is that Mac OS X 10.5 could cook your breakfast and gas up your car and still attract no potential switchers. A stable market is also often a stagnant market, for better and worse. (Also, if you ask me, Apple caters well to both consumers and professionals in one go, simply by producing different products. This is reason alone why they won't go down in flames that easy. Their business model works and they rake in profits. )
Sure, lots of people use shells, but by a wide margin more people will open, say, iWeb on OS X or Windows Movie Maker on XP before they ever open "Terminal" or "Command Prompt". "Hiding" it in Utilities doesn't do a thing - if you need it, it's there. And contrary to your belief, lots of people legitimately never *need* to use a shell for anything. Sticking tools where people who need them will find them and people who don't need them can blissfully ignore them isn't a good ground to dismiss an OS as "frustratingly bad" in this aspect, unless it's your conviction that every OS should assume that their primary user writes shell scripts or at least is a heavy shell user, something that with the advent of Ubuntu and other more "friendly" alternatives is increasingly hard to claim of even Linux distros anymore.
As an OS X user, I'm finding your claim about bizarre keystrokes to get everyday things done, well, bizarre. Is this about the Command key? Is it about the Home/End keys moving the viewport and not the text cursor? Is it about selections being extended instead of shortened in some cases? Here, we're on the same page, but keystrokes sounds like "shortcut keys" - list five of your bizarre shortcut keys and then we'll talk.
I can't complain about the way you feel shortchanged expecting a bundled OS X command line package manager (because package managers do exist - DarwinPorts and Fink) and more open-source software (Rosetta, the PPC-to-x86 translator, is currently holding back the kernel source release for x86; meanwhile such blockbuster hits as TextEdit (sample code), Chess (GPLed) and the Safari rendering engine WebKit (supposedly to facilitate integration with KDE but probably a smart recruiting move) are already open-source beyond Darwin). I have high hopes for the next revision of OS X. Apple has some genuinely smart *NIX-heads on board who really do get it, even if a small bit of Not-Invented-Here syndrome still lingers in the organization, and 10.4 is so much better than 10.2 (the first version I really tried) in all aspects that it's hard to believe Microsoft was toiling away at Vista in the same time span.
Maybe because introducing a unit that *shouldn't* scale along with the rest of the page (and the rest of the units) would complicate the spec even more? Look at how IE6 implements the pixel unit in conjunction with scaling. It's not consistent and it's not pretty.
An absolute unit does make sure that, say, a sidebar stays a certain width when scaled up, I'll give you that. But it also hinders the sidebar to move beyond stamp-sized in high pixel density displays, which for all I know could have been in the marketplace in droves by now hadn't there still been big software issues with regards to scaling.
I'd say that taking all this into account, it would be a good idea to have a unit that at its baseline occupies one pixel, and it logically follows that it would be called a pixel.