Post a Comment
If you read the article he's not attacking Apple. He has criticized various things publicly, probably because he loves the platform and wishes for it to shine. Here he is reflecting on the past and offering opinions that at least sound somewhat naive when put into a news article.
there is this even more famous computer vip who seems to think memory protection not really necessary see 3.7 of
http://www.cs.inf.ethz.ch/~wirth/Articles/GoodIdeas_origFig.pdf
Wirth only says that it memory protection "should" have been superceded by safe languages and correctly implemented compilers.
While creating a safe language is certainly possible (if you're prepared to pay the cost in flexibility and/or performance), ensuring a correct implementation that is safe not only from accidental bugs but also from malicious exploits is a huge problem.
Meanwhile, hardware memory protection may not be as fain-grained as compiler-based schemes, but it costs very little in performance and one can be fairly certain that it's correct.
But in this context we are really talking about a specific system, not that you, in theory, can make a secure system without protected memory. I don't think anyone would doubt that it is possible even without reading Wirths work, it just wouldn't make sense that using protected memory was the one true way.
Now about MacOS 9 being secure without it, hell, it wasn't even close to being stable, so i seriously doubt it would be, in any way, secure. I personally think the Macintosh would be dead if they hadn't switched OS. I just don't see how they could fix up an OS that in important ways were technologically behind even Windows 95. (of course it would be possible, but would be a slow and expensive process)
Another question would be, how many of the current MacOS X users would have used a system based on OS 9. I know i wouldn't have touched it with a 10' pole :-)
Well you see the latest 3rd episode (I think) of channel 9's coverage of the Singularity project http://research.microsoft.com/os/singularity/ . IIRC they had about 10% increase in perform. To put that in perspective that would your safer compiled language would run faster than your C code. Java could be beating it regularly (but not always of course, because under some situations Java just starts to suck shit).
> hardware memory protection [cut] costs very little in performance
That's not what the guys from the Singularity project said: they had the same level of performance for system call with their unoptimised implementation as Linux or Windows.
The reason why they were able to do so is that they are not using the hardware memory protection.
I have to respectfully disagree with the great and powerful Woz.
First, security through obscurity is never a plus of an operating system. I mean, let's say I was running an OS that I developed that was full of security holes, but because no one had access to a copy of it (the only copy in existence being on my computer), it would be hard if not impossible to crack. Maybe obscurity makes you less prone to attack, but it doesn't mean that it's more secure.
Second, OS9 didn't live to see the days when computer attacks were so common. The internet was still pretty safe, people hadn't migrated to cable modems and such as much. It was simply living in a different reality. OS9 was pretty much out as the attacks on WinXP started. I remember that fall at my school - nearly half of windows machines got compromised - but OS9 wasn't really used anymore. It just never saw a time when attacks were common.
Now this is an example where security by obscurity DOES work: Apple OS 9 and earlier were closed-source systems, but they were also relatively obscure.
Subtract the internet, and you have a system which is more secure from attacks than the open-source and well-understood BSD. It has holes of a different colour to those of Windows.
Change one of those elements (so that OS <9 becomes an open-source, but obscure, OS, a closed-source OS as ubiquitous as Windows, or an OS as wide open to the net as OS X is now, and you have all the potential for this non-memory-protected OS to be every bit as insecure as Windows.
sure, obscurity works agains script kiddies doing wide spread generic attacks to make botnets.
But those sort of attacks aren't the one people should be worried about. They are easy to prevent on any OS.
Obscurity is pointless when a direct attack is made on you company by someone(competitor, theif, etc.) to steal your data or spy on you. The attack would have researched you company and know that you were running OS 9.
They would get a copy of OS 9 and do some random attacks on it's network services to attempt to uncover bugs, and surely there would be many because being obscure also means having less resources to search for bugs.
Or, the attacker might just send you an email with a malicious program attached that expoilts the lack of memory protection in OS 9 and gain complete control of your system.
"Subtract the internet, and you have a system which is more secure from attacks than the open-soemse urce and well-understood BSD."
Uh, if you subtract the internet and any outside connectivity all systems are equally "secure".
You cant compare a connected BSD system to a disconnected OS9 system.
Two things.
Watching these security updates for some time, many of them come from the bundled open-source applications that are included as part of the OS. With such a large range of developers working on all of the different components, its no wonder that they're releasing more updates as they discover more issues.
Secondly, you have to admit that OSX is ten times the OS that OS9 was, you really can't compare them. The more features you have the more little bugs will sneak in the door.
What you need to be more concerned with are the operating systems that don't get regular updates.
"Secondly, you have to admit that OSX is ten times the OS that OS9 was, you really can't compare them. The more features you have the more little bugs will sneak in the door."
I'm not a programmer, nor am I a mac user, but I did learn C for a couple of months on PowerMacs, a few years back. In CodeWarrior, running on top of (I think) OS8. Well any memory allocation mistake in my (crappy) code used to bring the entire system down at runtime. Fatal error message with the bomb and all.
It may have been more secure at the time, but there was probably not much that could be saved under the hood...
First, security through obscurity is not a security. The holes exist and are neglected on the assumption that no one will try to exploit them... Until someone exploits them.
Also, there were less possible holes in Classic because it didn't have near the functionality that Mac OS X offers. This is tantamount to selling a rickshaw on how little engine care is involved. What the salesman isn't telling you is that the thing hits 7 MPH and carries two people, tops.
As for Apple continuing on with the Mac OS 7/8/9 lineage, it may not have been as much of a mess as Apple or others made it seem in the late Nineties, but it certainly was not up to the task of doing everything a more "modern" OS could do. BeOS and OPENSTEP had it beat in all of those departments. Shunting in that functionality to Classic would have been a larger headache with about the same amount of risk as the Rhapsody/Mac OS X plan was. So, why not just go to a new platform? And that's exactly what Apple did.
Those who know the history of Apple easily recognize that Steve Wozniak was the man behind the technical splendor that was emerging. Apple I, Apple II was due to his labour.
Nevertheless, I think that Steve has lost a little bit sense of reality. Times are different, way different and raw comparasion is not that easy.
Also, the option to base Mac OS X on *nix grants access to a lot of Free Software apps already available for *nix. On the other hand, there are a lot more programmers with knowledge of a *nix than compared to those that have expertize on Mac OS 9.
Continuing to develop over Mac OS 9 would be senseless. I'm sorry Steve, you've missed this one.
Edited 2006-10-25 20:17
Those who know the history of Apple easily recognize that Steve Wozniak was the man behind the technical splendor that was emerging. Apple I, Apple II was due to his labour.
Yes, that is the common myth, but read the book "On the Edge" about the history of Commodore and you get a very, very different perspective of the Woz. He is not and was not the god that some people make him out to be.
"He is not and was not the god that some people make him out to be."
Indeed. Commodore, Acorn and various other companies were making better engineered 6502-based microcomputers than Apple were, although Woz's "from scratch" hardware hacks do afford him some respect.
But, Wozniak concedes, the switch to OS X wasn’t a bad thing. “Now looking back it’s a good thing we have a better more stable more complete better operating system, from the ground up,” he said.
So Woz don't think it was needed, but he also don't think it was a bad choice...
There is (and there was) much more to OPENSTEP than preemptive multitask..
Of course, it provided proper multitask, obviously an improvement on OS9. But the main advantages of NeXT's OPENSTEP were the absolutely incredible development tools (InterfaceBuilder and the OpenStep framework), the really object-oriented environment (sadly, OS X tonned that down -- all the capacities are here but aren't as exposed to the user as they were on OPENSTEP) full vector display (DisplayPostScript), true multiuser, etc.
OS X is an evolution of these concepts. In addition to that, beeing based on PostScript, OPENSTEP provided a perfect environment for DTP -- one of the big Apple markets.
BeOS was a very cool OS, indeed, with excellent ideas. But OPENSTEP was pragmatically better (if only because it supported printer, pantone colors, was more robust, had a proved track record, etc), and really a better choice in the end. The BeOS toolkit was nice, but Cocoa is quite a few steps above (beside Objective-C > C++ .. but hey).
Edited 2006-10-25 20:32
How much does the Woz know about software (in comparison to his knowledge of hardware?).
My impression is he is a hardware guru, can anyone shed light on this? Having owned an Apple IIe, Mac SE (still working), and now a G5 (with a couple of things in between), I have emense respect for the man and what he helped create but is this really his forte?
I think what Woz is hinting at is that Apple needed not ditch the structure of OS 7-9 in favour of a UNIX-based one. I think he wants to say, hey, we could've extended and improved our existing OS, so that it would be capable of doing all the things OS X does now.
And I think he has a point.
...need one say more?
How many years of development was it before Apple finally discarded the idea of the '32-bit, multi threaded, multi tasking, memory protected, smp-aware modern OS based on OS 7.x'???
I personally don't think they'd ever get it right... It'd be too much work, too many changes had to be done.
Edited 2006-10-25 21:32
SO they could have spent millions trying to get their not-so-great older OS proper memory protection, SMP support, preemptive multitasking etc etc or purchase a very nice, already completed OS and add a VM that would run a version of the older os for backwards compatibility?
I think NeXT was a MUCH better choice. They would have had to completely rewrite classic and still had to use a VM (most likely) to be properly compatible with older apps.
or maybe im totally wrong.. who knows...
Erm, it is not as if OSX is NeXT 2.0. They used components and ideas from NeXT.
Having actually used NeXT for a while back in the day, saying OSX uses components of NeXT is silly. It is not a fundamentally different OS, OSX IS NeXT. It is heavily modified under the hood and has a spiffy new paintjob, but it is still NeXT. OSX is NeXT with components of System 9 in it, not the other way around.
"I think NeXT was a MUCH better choice. They would have had to completely rewrite classic and still had to use a VM (most likely) to be properly compatible with older apps."
Indeed. Apple's operating system rewrite projects were just tar pits, and the infighting and politics that make up 90% of the Apple story would have made the whole strategy unviable. If you have to credit Steve Jobs with one thing it's cutting right through all that and giving the Mac the operating system it should have had all along.
I agree that Woz has a point.
Apple's failure to produce a "modern" operating system was the result of bad management, rather than the shortcomings of the classic Mac OS. It took Apple many years after the acquisition of NeXT to recover from those management issues and produce a mediocre operating system (Mac OS X 10.0 and Mac OS X 10.2).
Carbon itself should be ample evidence that Mac OS could have been modernized. It is the classic Mac OS API, adapted for memory protection and preemptive multitasking and other such goodies. But Apple felt that they had to reinvent itself, so NeXT became the keystone.
... read this article (and cry) http://en.wikipedia.org/wiki/Mac_OS_memory_management
[bomb] "Sorry, an error of Type 11 occurred." Which I saw constantly on my Powerbook 520c, running Systems 7.5.3 - 7.5.5.
There are some things I liked better, appearance-wise, on the Classic Mac OS, such as labels coloring the entire folder, not just the title, but for the most part System 7 was not very stable. I am a Mac user, and supported Apple throughout their troubles in the mid-1990's, but a lot of their efforts during the years weren't really that wonderful. Better than Windows, but not up to OS X standards.
"In this book Wozniak claims that the OS 7 crashes were linked to running Microsoft Internet Explorer, rather than the Apple operating system. He says that his own research made it clear that people with IE installed on their Macs were suffering from crashes, but those who used alternative browsers, such as Netscape, were experiencing no problems...There was nothing wrong with OS 7."
HA HA HA
I remember using OS 7 before the internet was even a reality. To say that the mysterious crashes were a post-IE problem is laughable to anyone who used those systems. The lack of proper memory managament caused the system to crash all the time when using industrial apps like Adobe Photoshop or QuarkXPress. One sloppily written plugin could bring the whole machine down. That says nothing for the cooperative multitasking model causing ridiculously poor multitasking performance. I recall running large print jobs intentionally while playing Tetris because it made the advanced levels slow as molasses. I'd love to sample a little of what Steve is smoking.
Edited 2006-10-25 21:56
The internet pre-dates the MAC, so I wonder what hardware platform you were using OS 7 on.
Are you confusing the original ARPAnet with internet? The pubicly held concept of the internet is a 1990's thing, and last time I checked the Mac was first made in 1984.
Regardless, the world wide web was several years after the introduction of the internet (back then it was all the proprietary BBS's, AOL, Compuserve, GEnie, et cetera) and Internet Explorer for the Mac came even later than that. It doesn't change the original point that there were stability problems with OS 7 long before Internet Explorer first graced the platform with it presence.
Edited 2006-10-26 12:28
Are you confusing the original ARPAnet with internet?
In '84, we split MILnet from ARPAnet with only a few trusted gateways allowed. At that time we started calling it the internet, although it was always an internet. The "public" internet started when we split off NFS net and DARPA stopped funding. At that time it became possible to access the net via a limited number of commercial ISPs.
You're "90s" thing was when we allowed the internet to be used for commercial purposes. Bad mistake, but there's no going back.
In '84, we split MILnet from ARPAnet with only a few trusted gateways allowed. The "public" internet started when we split off NFS net and DARPA stopped funding. The "public" internet started when we split off NFS net and DARPA stopped funding. At that time it became possible to access the net via a limited number of commercial ISPs.
Merger of NSFnet and ARPAnet occured in 1986, but it was still for government and academic uses, if I recall correctly that was after the release of the original Macintosh wasn't it? Likewise widespread public use of the internet, specifically outside the government and academia, wasn't several years later. Furthermore Woz's contention about the IE browser caused the stability problem is more an issue of the world wide web aspect of the internet, not the network itself. While the first web server may date to 1990, it wasn't until 1993/1994 that anyone was really using it. At that point System 7 had been around with stability problems galore for some time.
You're "90s" thing was when we allowed the internet to be used for commercial purposes. Bad mistake, but there's no going back.
I know, it's been so horrible that we've managed to connect so many people to each other and information resources while at the same time letting people make some money in the process.
Actually Steve Wozniak's correct.
Except for two (?) caveats, "inits" and A5 worlds.
Mac OS 8 (Copland), the real one:
http://en.wikipedia.org/wiki/Copland
http://www.amazon.com/gp/product/0201479559/qid=1152563948/sr=1-16/...
was on it's way to addressing much of what we enjoy today.
If you have ever seen any of those Copland builds "floating" around, as crash prone as they were, were beautiful, the ResEdit they shipped with it was functional and they had new resources/extentions thingys (<-- technical term) that I never quite figured out, kinda, maybe an OpenDoc modular model.
The OS 8 book sited above clears up a lot, it's quite through.
The programmers that built the Macintosh were artist
System 7 was a work of art.
What became Mac OS 7.6, 8.x (7.7) and 9,x, were the remnants of that failed attempt/experiment.
I'd really like to have seen what Copland could have become, that whole OS model was a revolution.
http://myoldmac.net/webse-e-flash.htm
http://www.nd.edu/~jvanderk/sysone/
http://main.system7today.com/
http://www.macdrivermuseum.net/disk.html
http://www.mackido.com/
I have owned my own Mac since the days of the Performa 550 running os 7.5.3. During which time I was constantly online (eWorld then AOL). My friends all had x86 boxes running WFW3.11. During this time my friends were always getting some kind of virus or trojan on their computers and I NEVER HAD ONE. Not because people didnt write them for Macs, just because it was harder to compromise a Mac. True, if you compare an OS9 box to a new XP SP2 or Mac OSX box, it probably isnt as secure, but for its time I believe they were very secure.
OS X is an incredible OS, but I think that Apple, with the right direction, counld have taken their own os and combined it with the best of the others around (BeOS and OpenStep) and made an OS that would have destroyed everything in sight
Edited 2006-10-25 22:35
At least you could make heads or tails of what was going on then.
Now:
http://developer.apple.com/documentation/Darwin/Reference/ManPages/...
You don't have a chance of figuring out who is really controlling what's going on.
http://www.securityfocus.com/columnists/402
CarbonErrors
http://www.versiontracker.com/dyn/moreinfo/macosx/19236
Meaning of Error Type 11
http://docs.info.apple.com/article.html?artnum=16004
http://docs.info.apple.com/article.html?artnum=18912
http://developer.apple.com/documentation/macos8/mac8.html
ResEdit
http://www.versiontracker.com/dyn/moreinfo/macos/417
MacsBug
http://developer.apple.com/tools/debuggers/MacsBug/
http://developer.apple.com/documentation/macos8/DevTools/MacsBug/Ma...
Stuff
http://download.info.apple.com/Apple_Support_Area/Apple_Software_Up...
The title of this article is soooo wrong!! Sorry to say but this feels like an OS News flamebait.
Lets break it in 4 small parts:
1- Most of the crashes in OS7 were because of Microsoft's IE even if it wasn't running:
QUOTE:
"OS 7 crashes were linked to running Microsoft Internet Explorer, rather than the Apple operating system(...)those who used alternative browsers, such as Netscape, were experiencing no problems(...)IE did not have to be running for the crashes to happen"
2- Woz never says OS9 is more secure than OS X, not once! He only says he has warned Steve Jobs that Apple didn't need the huge amount of money needed to buy NeXT.
QUOTE
"we didn’t really need a new operating system for that amount of money involved in the purchase of NeXT"
3- But he does compare both OS9 and OS X, just to say the oposite of the article's title:
QUOTE:
"Looking back it’s a good thing we have a BETTER MORE STABLE MORE COMPLETE BETTER operating system, from the ground up"
Again:
- Better
- More stable
- More complete
- Better operating system
4- The only thing he says is that Unix is more prone to attacks because people are familiar with the holes in Unix. Well duh! But Unix is around for more than 30 years, this is a proven base for any 'internet ready' Operating System. The core of OS X is NOT made of unproven technologies!
5- The rest of the article is just boring stuff about his past experiences, talking about the type of person who seeks to break into an operating system and such.
Apple's project 'Copland' gave us nothing. They were just spending money and resources while Win95 was eating Apple's market share. Gil Amelio and even Ellen Hancock realized the situation was hopeless. So they did the right thing. NeXT was bought, Steve Jobs returned more matured to his 'home' and Apple in right on track.
Again, the only reason I can find for such a bad chosen title is, I'm afraid to said, that this is just a flamebait. Even the original article only says 'Apple didn't need NeXT' - Woz.
Edited 2006-10-26 01:38
Woz never says OS9 is more secure than OS X, not once!
Oh? He clearly says OS 9 was more secure than OSX:
"Not that Mac OS 9 - the operating system that predated OS X - was not stable. According to Wozniak. “That was something to say for Mac OS 9. It was just so secure,” he said, denying that the security was based on the fact that being a minority operating system meant that nobody was interested in attacking it. He explained: “Mac OS 9 was differently constructed to anything else. It wasn’t because it was minor and unknown.”
Mac OS X on the other hand is built in Unix and is therefore is more prone to attacks because people are familiar with the holes in Unix, explained Woznaik. “Some of the holes in Unix are well known. So keeping Firewalls on is more important. And we keep announcing, even our own security fixes, not as many as Microsoft but still we never really had those in the OS 9 days."
I see the distortion field is still up and running
.
Edited 2006-10-26 08:11
I think you got it all wrong. Ok, he does say Mac OS X is more prone to attacks. Its Unix and it has been around for more than 30 years. Sure its faults are well know. OS 9 was not an 'internet OS' and security via obscurity is no security at all. The worst thing is not being unsafe, is thinking you're safe when you're not.
He does compare OS9 to OSX when he says that, and I quote again "Looking back it’s a good thing we have a better more stable more complete better operating system, from the ground up"
Once again, comparing to OS9 in Woz's own words, OSX is:
- Better
- More stable
- More complete
- Better operating system
No distortion whatsoever. I really think you got it all wrong.
No distortion whatsoever. I really think you got it all wrong.
No, YOU get it wrong. He says CLEARLY that OSX is less secure than OS9. It's RIGHT THERE in plain sight.
"That was something to say for Mac OS 9. It was just so secure," he said. [...] Mac OS X on the other hand is built in Unix and is therefore is more prone to attacks [than OS 9] because people are familiar with the holes in Unix, explained Woznaik."
Governa, it's right there in plain sight: he says OSX is more prone to attacks than OS9. In other words, Woz says that OSX is less secure than OS9. Anyone not affected by the RTF understands that.
> No, YOU get it wrong.
It doesn't need to be an expert to realize Mac OS X is a better OS in every aspect (and I surely include security) than previous incarnations of Mac OS. I don't think Woz meant what you think he meant, and from what I've read, most of the replies seem to express the same feeling as mine.
One thing we agree is that we disagree. No hard feelings.
I still think you're flamebaiting though...
Edited 2006-10-26 14:01
Its Unix and it has been around for more than 30 years. Sure its faults are well know.
You know, UNIX actually evolved during these 30 years. Also, most "UNIX" exploits are actually application exploits (eg. SSH exploits, sudo exploits). The few that are kernel based are often specific to a particular kernel (linux exploits will not work on FreeBSD, FreeBSD kernel exploits will not work on NetBSD, NetBSD kernel exploits will not work on OpenBSD and so on and so forth)
The worst thing is not being unsafe, is thinking you're safe when you're not.
That applies to any OS, Mac OS classic included.






