Then you would be hacking the OS instead, becaus AFAIK Mac OS is limited to the chips that Apple want to run it on.
This not really true anymore. Remember that the Darwin kernel is open-source, the code that makes up OS X runs on top of Darwin. If you have drivers that enable Darwin to run on the hardware, OS X will run. In fact OS X already runs on systems not supported by Apple.
In fact, if OS X would check for the presence of approved chips, how the hell would Pear PC be able to boot OS X, except by duplicating the functionality of said chips which would be a) a lot of work b) probably illegal.
To run OS X on a XBox 2, you would basically need the following things:
Install some kind of bootloader on the XBox2.
Write Darwin drivers for the XBox2's hardware.
Somehow build a bootable volume that contains OS X + those drivers.
Have the bootloader build a device tree and start the Darwin kernel.
"Given the huge number of Linux drivers, those are the kind of changes that Apple can pull off but a near to impossible to do in the Linux world."
You trade the freedom to use lots of different hardware (Linux) for the convenience of using Apple hardware. Deciding which is better is personal.
I was not arguing about ideological question like open vs closed, or free or not, simply pointing out technical innovation that the Darwin kernel has, and Linux has not.
kernel are not that important anymore"
I have difficulty agreeing with this comment. Certainly there may be a great deal of interoperability between Linux and OSX (perhaps ignoring the GPL in some cases) but this is not the same as saying that the particular kernels are not important. If your comment were correct, would we need Free Darwin? (Also here and here.)
Actually, I think that Free Darwin is exactly an example that shows that Kernels are not that important. Free Darwin is not a different kernel, it is simply the Darwin kernel shipped with GNU user-land programs (the same than Linux). What people use, see and care for are the user-land programs - if they want the GNU user-land, they can install it.
Actually, I don't know who needs Free Darwin, I don't even know if people actually use it. People who want to install GNU programs and libraries usually rely on less political and more practical distributions, like fink, or open darwin (which is done with some help from Apple) or even gentoo's portage.
A related example is Quake3; it runs much faster under Linux than under Windows. Sometime Windows (applications) run faster under wine than under Windows.
I don't think that characterizing the performance of an operating system by the performance of a game that relies little on the operating system's service is very relevant. I suspect that Quake performance is heavily influenced by the quality of the graphic card's driver more than anything else.
However, OSX has an extra layer which probably does not help speed things up.)
What extra layer? Are you talking about the Mach micro-kernel? The BSD layer and the Mach kernel are running in the same memory space, so the overhead is limited. Also this separation between Mach kernel and BSD personality make the design of the kernel cleaner. This might or not be an advantage in the future, it is difficult to decide now, but there must be a reason people from the Hurd are trying to build a similar architecture...
I believe that the development and improvement of Linux will increase much faster than that of "MacOS" or Windows and that the "4.0 Linux kernel" will be much better and more powerful than "OS12" or "Windows XYZ".
Actually, Darwin has quite a few cool ideas that would be nice in the Linux kernel like IOKit (object oriented driver system) or the unified system to launch service that should be in Tiger. Given the huge number of Linux drivers, those are the kind of changes that Apple can pull off but a near to impossible to do in the Linux world.
You also have to realise that, all in all, kernel are not that important anymore, what is important are libraries and applications. At this level, the differences between OS X and Linux are not that large anyway, as most Linux libraries and applications can run on OS X.
I was there a few weeks ago, visiting a friend who does research there, while travelling in Switzerland. It's definitely an interesting place (although it lacks the futuristic "aura" that I somehow expected). Near the main gate is a massive wooden sphere called the Innovation Globe. It's still in construction, but it looks like it will be a great and interesting facility (with public exhibits and theaters), and its organic look is a stark contrast to the mostly drab buildings inside.
Re:It's not "Cern"
on
Happy 50th Cern!
·
· Score: 4, Informative
First, nobody uses the acronym C.E.R.N. (I know, I work there). Second, CERN is not really an acronym anymore, as it stands for Conseil de Recherche Nucléaire Européen, and is the name of the council that founded the CERN and was disbanded in the fifties. Third as somebody else pointed out, rules for capitalization change depending on the country, so depending on the reference language (UK English, French, maybe Swiss French) it might be correct or not, in any case US English rules do not apply.
The official name is, in French Organisation Européenne de Recherche Nucléaire (which would be OERN), and in English European Organization for Nuclear Research (which would be EONR). The name CERN simply stuck because it sounds nice and people are used to it, perhaps also because of the German word Kern that means nucleus.
In the Geneva area many people believe that CERN stands for Centre de Recherche Nucléaire Européen (I learnt that at school), although this was never true.
Meanwhile, the poor Babel fish, by effectively removing all barriers to communication between different races and cultures, has caused more and bloodier wars than anything else in the history of creation.
Maybe this explains the poor quality of the other fish, it is not that machine translation does not work, but a valiant effort to prevent wars caused by better understanding.
Actually, I don't think it is a drawback for the market the iMac is aimed at. In this case, the lack of upgradability is indeed not a problem. My mother has one (G3) iMac, and had monoblocs Macintoshes before (my trusty old SE/30 and a classic before). So I understand this approach and thing it is reasonable.
On the other hand, to play the devil's advocate, I suspect that for people, it is a drawback, no that they will really need the extension features, but because they might. This includes of course many in the/. crowd, but also some 'semi-technical people', who might want to keep one open option.
You don't need a internal extension now, but what about tomorrow? The real question is, will the 'next big thing (tm)' be usable using a firewire or an USB 2.0 connection (and if yes will this be convenient / expensive).
You might argue that FW has enough bandwidth to handle anything you can connect to a simple PCI connector, and you would be right. You could also argue that there is no 'next big thing' that would connect via PCI to the machine. I would agree, the next big thing is probably going to have its own processor and have an ethernet connector, but if you look in the past, all the external interfaces could mostly be added to machines using internal connectors:
SCSI
Ethernet
SCSI 2
USB
Fast Ethernet
Firewire
USB 2.0
Gigabit ethernet
Firewire 2
Notice a pattern?
The second factor is convenience. You can add a serial port to a Mac via an USB dongle. You can add a memory stick card reader using USB. This is nice but means cables, lots of cables. Between an PCI USB card and an USB hub, there is little difference in price, but one is compact and does not need a power supply.
If you look at single block Macintoshes, many had a single extension slot:
the SE line (my SE/30 had an ethernet card).
the SI
the LC line
the pizza box PPCs (6xxx)
the one in all PPCs (5xxx)
20th anniversary machine
All I wanted to say is that I feel, for many, extension slots are not so much a direct need than an insurance against future change. To a large extent, I agree that this is irrational, but on the other hand there are many irrational things about computers that decide if we will or not buy them.
Really, they have taken a laptop, removed the keyboard and touch pad and given it a stand.
If you a look at the inside of the machine, you will notice that while it is a compact design it is much thicker than a normal laptop mother board, the hard-drive is also a 3 inch model and the power-supply is included in the box. The design is much closer to a pizza-box design as a laptop design. There have been other models done in this way (like for instance the 20th anniversary Mac.
When you think of it this way, one really does have to ask the questions, "Why the hell hasn't this been done to death already?".:P
Extensibility, this kind of design means that the machine will not be extensible, no PCI slots, no possibility of changing the video card, in short most of the drawbacks of the laptop design.
Except for Microsoft themselves. They've already dumped PowerPC, MIPS and Alpha support to release solely on x86.
This can change, once you only have managed code, going to another architecture is reasonably easy. So going with another processor is not completely impossible.
While you respond in disgust, what happens if one day science does indeed discover that biology trumphs freewill? What if almost all of out behaviors are predetermined by chemistry?
You have a very naive idea of science if you think such a thing can be proven scientifically. It would first imply that research on the subject could be done without religious/political/cultural meddling. Then it would imply a categorical answer. Even if human behaviour could be determined at 55% by chemistry, there still is freewill (and please explain how you are coing to quantify the percentage).
Science can at best indicate factors. No doubts human behaviour is influenced by a myriad of factors (including what humans believe are the factors that influence their behaviour) so reaching a definitive answer on the question of determinism makes little sense.
You don't listen to music because it is difficult....interesting.
No, this is quite legitimate.
There are times when you want to listen to music, without giving it much attention. This means you don't want to navigate a filesystem every three minutes to select the next tune, which would amount to interrupt what you are doing. Playlists don't work well with me, because this means selecting stuff in advance. You can do the analogy to how you listen to music in your car, playlists are like burning your own compilation on a CD, selecting files is like inserting a new tape after each tune. The first one implies work and to be organised (I'm not), the second would not be very safe. Of course, you listen to whole albums sequentially, but this is often not what I feel like.
I used to have a Sun station with XMMS, but not listening a lot to music because of this. With iTunes, I usually browse either by genre or artist or even use the search facility and play the set of songs that come out. The party mode is also, I think, quite a good idea, although I have to figure out how to prevent certain tunes from ending up in there.
They are very similar from the user's point of view, but in every other way (underlying architecture, overall goals, design philosophy), there are significant differences that make the idea of a merger of the projects rather silly.
I'm not so sure, if you take OS X, there are basically two toolkits Cocoa and Carbon. They are both very different in terms of architecture and design philosophy. Carbon is a C (originally Pascal) API that was designed to run on the original 8 Mhz Macintoshes. Cocoa is an objective-C framework that relies heavily on oo features like reflection and dynamic bindings.
The difference between the two APIs is probably similar in scope to the difference between KDE and Gnome. Yet Apple is actively merging both APIs, they share the same low-level drawing engine, have the same look and feel and in general are not easily distinguishable by end-users (they are telltales of course, but not the kind the average user notices).
Can that actually be done? I have never seen any controlls in the panel that refer to turning DX off.
I dont think there is a control panel to disable DX on windows, but then again, I never saw a control panel in Gnome to disable 3D operations in the X server.
I would be astonished if you could not cripple DX by changing the permissions of the DLLs and thus prevent the execution of games. For me this is the same kind of operations than disabling DRI in the X11 server.
All in all, I'm simply no convinced that the whole 'we can cripple the X server' is really a strong point of Unix vs Windows.
Exactly. I dont people playing Quake or Unreal Tournament during working hours..
Ok, I can agree to the general idea, yet you will only block one category of games, this will not prevent anybody of playing solitaire, bubble trouble, or any of those silly flash games, so this is a complicated measure for a very limited effect.
Also I wonder if you cannot get a similar result on windows by disallowing access to direct X?
1: Allowing graphical interface but NOT allowing 3D graphic card operations used (Simple with X, deny access to DRI)
Could you explain why you would like to do this? I mean what security gain you get by doing this? I cannot imagine a scenario were a person should be allowed use of the display, but should not use high-speed 3D operations, or where using those operations would be damageable to the system.
The difference would be that Apple's R&D, for the most part, comes up with good products, and the ones that don't make the grade don't get announced, let alone shipped.
Apple had more of its shared of failed project, shipped or not. Many of those ideas were great, and failed, not so much because of their shortcoming, but because they came at the wrong time, or where not marketed correctly. In fact elements of those failed projects are often reappearing in Mac OS X.
The problem is not so much coming out with good ideas, but to manage to sell them...
I think the reason why see those Sun is dying articles is because the 'Apple is dying' articles are starting to be really difficult to take seriously.
Actually, the article is eerily similar to the 'Apple should have' articles. Basically, what was done wrong was to try to do new things, invest in research. Instead the company should have built wintel boxes like Dell and fired a maximum of people.
How many companies have been successful in imitating Dell except Dell?
Constant standards are what is needed to make software last that long.
True.
Language standards don't even last 200 years, how do we expect something as new as software standards to be more uniform than language standards?
Most of the standards we use a quite old: the numbering system, angles, time units and well the metric unit system...
Human languages are IHMO a bad reference point, first because they are not simply communication tools, but also social tools. Also while languages evolve quite a lot, we can still read languages written a few millennia ago. Humans are quite good at understanding unknown human languages - they can even bootstrap with an unknown language. Some human languages have even built-in variability and an initial handshake mechanism.
For data, I think XML is a step in the right direction, between trying to parse and unknown binary file or an unknown XML file (with the DTD), I certainly would prefer the second. For code, the situation is more messy, how many people can still understand FORTRAN or COBOL code?
Wouldn't that be a good reason to study it? We know language evolves - we can study and compare historical documents from different time periods to see that. But when have we ever been able to see the evolution happening right before our eyes, at such a rapid pace?
Actually, there are languages that evolve very fast. One example are the Swiss German dialects. As they have no fixed written form, those language tend to change in short spans of time (mine is twenty years old for instance). One the interesting side-effects of this is that two speakers have to do some initial synchronisation in order to communicate. Because it has no written form, Swiss German dialects cannot be used for formal, written communication. High-German is used for that, so this means people basically use two language, one for day to day communication, and another, formal language. The linked paper offers good insights about the divergence of the languages.
One thing I noticed is that as mails are not considered 'serious' written matter, people write Swiss-German in a phonetic way. The resulting text is as legible for a German as l33t-speak. On the other hand this form also enables the reader to figure out where the writer is from. Of course, this is mostly done by young people. I would not be surprised to see blogs written in Swiss-German...
Actually, Apple did the same for the original Macintosh:
Mac 128K: 8Mhz
Mac 512K: 8Mhz
Mac+: 8Mhz
Mac SE: 8Mhz
Mac Classic: 8Mhz
The first machine was introduced in 1984, the Mac classic was retired in 1992. The main improvement between these machine is in the memory and the I/O system. The Macintosh Classic could fully boot with a disk image stored in ROM.
Slashdot Mirror
In fact, if OS X would check for the presence of approved chips, how the hell would Pear PC be able to boot OS X, except by duplicating the functionality of said chips which would be a) a lot of work b) probably illegal.
To run OS X on a XBox 2, you would basically need the following things:
Actually, I don't know who needs Free Darwin, I don't even know if people actually use it. People who want to install GNU programs and libraries usually rely on less political and more practical distributions, like fink, or open darwin (which is done with some help from Apple) or even gentoo's portage.
I don't think that characterizing the performance of an operating system by the performance of a game that relies little on the operating system's service is very relevant. I suspect that Quake performance is heavily influenced by the quality of the graphic card's driver more than anything else. What extra layer? Are you talking about the Mach micro-kernel? The BSD layer and the Mach kernel are running in the same memory space, so the overhead is limited. Also this separation between Mach kernel and BSD personality make the design of the kernel cleaner. This might or not be an advantage in the future, it is difficult to decide now, but there must be a reason people from the Hurd are trying to build a similar architecture...You also have to realise that, all in all, kernel are not that important anymore, what is important are libraries and applications. At this level, the differences between OS X and Linux are not that large anyway, as most Linux libraries and applications can run on OS X.
It is a perfectly valid name, but it is not an acronym anymore.
Acronym n : a word formed from the initial letters of a multi-word name
The official name is, in French Organisation Européenne de Recherche Nucléaire (which would be OERN), and in English European Organization for Nuclear Research (which would be EONR). The name CERN simply stuck because it sounds nice and people are used to it, perhaps also because of the German word Kern that means nucleus. In the Geneva area many people believe that CERN stands for Centre de Recherche Nucléaire Européen (I learnt that at school), although this was never true.
I think you are misunderstood, the goal of CERN is to see the Higgs boson.
On the other hand, to play the devil's advocate, I suspect that for people, it is a drawback, no that they will really need the extension features, but because they might. This includes of course many in the /. crowd, but also some 'semi-technical people', who might want to keep one open option.
You don't need a internal extension now, but what about tomorrow? The real question is, will the 'next big thing (tm)' be usable using a firewire or an USB 2.0 connection (and if yes will this be convenient / expensive).
You might argue that FW has enough bandwidth to handle anything you can connect to a simple PCI connector, and you would be right. You could also argue that there is no 'next big thing' that would connect via PCI to the machine. I would agree, the next big thing is probably going to have its own processor and have an ethernet connector, but if you look in the past, all the external interfaces could mostly be added to machines using internal connectors:
- SCSI
- Ethernet
- SCSI 2
- USB
- Fast Ethernet
- Firewire
- USB 2.0
- Gigabit ethernet
- Firewire 2
Notice a pattern?The second factor is convenience. You can add a serial port to a Mac via an USB dongle. You can add a memory stick card reader using USB. This is nice but means cables, lots of cables. Between an PCI USB card and an USB hub, there is little difference in price, but one is compact and does not need a power supply.
If you look at single block Macintoshes, many had a single extension slot:
- the SE line (my SE/30 had an ethernet card).
- the SI
- the LC line
- the pizza box PPCs (6xxx)
- the one in all PPCs (5xxx)
- 20th anniversary machine
All I wanted to say is that I feel, for many, extension slots are not so much a direct need than an insurance against future change. To a large extent, I agree that this is irrational, but on the other hand there are many irrational things about computers that decide if we will or not buy them.Is it me, or does this picture show a lot of condensers?
Most motherboards I have seen don't have so many, or not so big. Or am I wrong?
Just wondering...
Science can at best indicate factors. No doubts human behaviour is influenced by a myriad of factors (including what humans believe are the factors that influence their behaviour) so reaching a definitive answer on the question of determinism makes little sense.
There are times when you want to listen to music, without giving it much attention. This means you don't want to navigate a filesystem every three minutes to select the next tune, which would amount to interrupt what you are doing. Playlists don't work well with me, because this means selecting stuff in advance. You can do the analogy to how you listen to music in your car, playlists are like burning your own compilation on a CD, selecting files is like inserting a new tape after each tune. The first one implies work and to be organised (I'm not), the second would not be very safe. Of course, you listen to whole albums sequentially, but this is often not what I feel like.
I used to have a Sun station with XMMS, but not listening a lot to music because of this. With iTunes, I usually browse either by genre or artist or even use the search facility and play the set of songs that come out. The party mode is also, I think, quite a good idea, although I have to figure out how to prevent certain tunes from ending up in there.
Why not use iperf, which is meant for this usage?
The difference between the two APIs is probably similar in scope to the difference between KDE and Gnome. Yet Apple is actively merging both APIs, they share the same low-level drawing engine, have the same look and feel and in general are not easily distinguishable by end-users (they are telltales of course, but not the kind the average user notices).
I would be astonished if you could not cripple DX by changing the permissions of the DLLs and thus prevent the execution of games. For me this is the same kind of operations than disabling DRI in the X11 server.
All in all, I'm simply no convinced that the whole 'we can cripple the X server' is really a strong point of Unix vs Windows.
Also I wonder if you cannot get a similar result on windows by disallowing access to direct X?
Apple had more of its shared of failed project, shipped or not. Many of those ideas were great, and failed, not so much because of their shortcoming, but because they came at the wrong time, or where not marketed correctly. In fact elements of those failed projects are often reappearing in Mac OS X.
The problem is not so much coming out with good ideas, but to manage to sell them...
Actually, the article is eerily similar to the 'Apple should have' articles. Basically, what was done wrong was to try to do new things, invest in research. Instead the company should have built wintel boxes like Dell and fired a maximum of people.
How many companies have been successful in imitating Dell except Dell?
Human languages are IHMO a bad reference point, first because they are not simply communication tools, but also social tools. Also while languages evolve quite a lot, we can still read languages written a few millennia ago. Humans are quite good at understanding unknown human languages - they can even bootstrap with an unknown language. Some human languages have even built-in variability and an initial handshake mechanism.
For data, I think XML is a step in the right direction, between trying to parse and unknown binary file or an unknown XML file (with the DTD), I certainly would prefer the second. For code, the situation is more messy, how many people can still understand FORTRAN or COBOL code?
One thing I noticed is that as mails are not considered 'serious' written matter, people write Swiss-German in a phonetic way. The resulting text is as legible for a German as l33t-speak. On the other hand this form also enables the reader to figure out where the writer is from. Of course, this is mostly done by young people. I would not be surprised to see blogs written in Swiss-German...
- Mac 128K: 8Mhz
- Mac 512K: 8Mhz
- Mac+: 8Mhz
- Mac SE: 8Mhz
- Mac Classic: 8Mhz
The first machine was introduced in 1984, the Mac classic was retired in 1992. The main improvement between these machine is in the memory and the I/O system. The Macintosh Classic could fully boot with a disk image stored in ROM.