MacOSX Vs BeOS ShootOut 416
Jolie writes: "After Palm purchased Be's assets, the future of BeOS became uncertain and a lot of users have left the platform. One of these users was Scot Hacker, mostly known for his 'BeOS Bible' book among other things. Scot tried to stick to Windows, then to Linux but he ended up with MacOSX. He has written a long and detailed article comparing, from the user's point of view, his beloved BeOS to his new favorite, MacOSX."
bah humbug... (Score:1, Interesting)
The perfect user (Score:2, Interesting)
Scot Hacker seems like the ideal OS X user. Unlike hard-core Mac users, like most of the OS X audience, he doesn't have Mac desktop environment that's tweaked exactly the way he wants and his hands don't automatically issue Finder commands. He's extremely at home at the command-line and can tap the power of the Unix underneath but still appreciates an elegant, consistent GUI. (Unlike desktop Linux fans, who consider middle-button text pasting that may or may not work between apps from different toolkits to be perfectly satisfactory integration.) And, as he said, when you're coming from Be, it doesn't take a lot of software to look like a vast cornucopia of available apps.
The one thing that surprises me is that the speed didn't bother him more. The biggest thing BeOS had going for it, besides that file system, was blazing, silky-smooth speed, whereas all the OS X systems I've seen dragged their butts. (Admittedly, I haven't used 10.1.) He did have a really fast box, though.
Re:BeOS... (Score:5, Interesting)
Frankly, no. For a system with such a small user base and development team as BeOS, the product was *highly* polished. It contained virtually every feature of a modern operating system with outstanding features ranging from the kernel (true multithreaded processing) to the interface (the textual "move to" and "copy to" options are some of the most ingenious interface considerations in a long time). Tet it's obvious that BeOS wasn't a finished product, but it was definitely going places quite fast-- and if the company was actually able to get money, the rate and quality of development would have been quite impressive. Ever hear of BONE or BeOpenGL? Besides, does an OS really need to have "polish" to market? Think of a little company in Redmond and define "polished".
The real reason BeOS wasn't "a contender" is because Microsoft screwed Be over with the fine print in its OEM contract. I suggest that you read this article [byte.com] by Scot Hacker with an accurate description of Be's demise.
Re:I hate these arguments (Score:3, Interesting)
2) "with a GUI similar to windows" --- It would be more realistic to describe Windows as having a GUI similar to the Mac, considering which came first.
Re:OS Preferences (Score:2, Interesting)
Graphics, layout, the whole thing.
Maybe you need a paper tape punch/reader. That's a proven, reliable means of cutting and pasting plain ASCII characters from one program to another. I was doing that back in 1975 on a Teletype ASR-33. Running a time sharing system, kinda similar to Unix.
Re:OS Preferences (Score:0, Interesting)
Also from the article:
After a few false starts, I had a running Mandrake box. But contrary to its reputation, Linux was crashing and freezing on me left and right. I had made the mistake of thinking that Linux had evolved enough by then to offer dual-processor capabilities as sophisticated as Be's. Wrong. Moving to a single-processor box fixed the stability issues, and I was free to explore the OS.
First off, if the guy has trouble installing something as point-n-clicky as mandrake, then his technical skills may have not been up to the challenge of writing an article such as this.
Secondly, when will people like this brainiac ever get that Linux != XFree86+Window manager.
Sure things like netscape on KDE or gnome crash every now and then. Hell, just X and a bad video driver crash so hard sometimes that it knocks out the keyboard.
But I've never had a case where I couldn't just telnet or ssh in and fix things. Try doing that on windows next time you get a BSOD.
The applications you are running on a OS are just that - applications. And yes some are buggy, but saying"Linux was crashing and freezing on me left and right" is nothing more than FUD perprated by someone who has nothing better to do than push his new favorite OS
Re:OS Preferences (Score:2, Interesting)
Right, I misspoke. Thanks for clearing me up.
You may or may not have used DOS systems, but every application in DOS that had a GUI looked (and worked) differently.
And boy, were they confusing, too.
Re:OS Preferences (Score:2, Interesting)
Yes. Most emphatically yes.
This does not neccesarily mean that my desktop will look (or behave) anything like yours. To me, it means that when I configure my system so that "shift-rightclick" means "copy the current selection to the clipboard", all my applications pay attention to my configured preferences.
This is a real basic issue of *nix user-friendliness (primarily for X apps - GNU tools have gne along way towards helping "standardize" command line interfaces.) I expect my computer to do what I tell it to do, and what I have configured it to do, not what some l33t hax0r d00d thinks it should be doing.
Re:OS Preferences (Score:3, Interesting)
Well, personally, I want everything done my way. =P I, for instance, don't like to use the mouse, and if CTRL-C copies highlighted text in all the applications I use and CTRL-V pastes it, I'm a happy person. I'm not tied to CTRL-C and CTRL-V (that's just what Windows uses, so everyone else does too), but I would like it to be consistent. IMO, the best way to handle this would be to allow the universal keystrokes to be definable so that I could make, say, CTRL-P be the "paste" shortcut in all of my applications. The OS (or it's GUI shell) would catch the preferred keystroke and pass on system-defined messages, which the applications would look for, instead of keystrokes. Not going to happen anytime soon, but still nice to think about.
As to making Linux internally consistent: I'd like it to be so, yes. I prefer that all of my computer knowledge become obsolete only with major upgrades, as opposed to each time I install a new application.
Re:BeOS... (Score:5, Interesting)
Most of this applications section isn't really about operating systems, but about the apps available for the operating systems, so you might want to skip it if you're just looking for the OS comparisons. However, I believe that the applications landscape is an integral part of the total OS experience, so included it here.
The problem is that apps are not "an integral part" of a computing experience. They are almost the totality of it. With the exception of some supergeeks, nobody buys a computer in order to run the operating system. People buy computers to run apps. No matter how lickable the shutdown/adduser/finder screen interface is, without real apps a system is doomed. If Be had all the killer apps that people buy computers for, it would still be alive today.
Nobody cares about threading, "multimedia support", or POSIX. Users want Photoshop, MS Word, Quicken, Halo and that goofy little custom VB app that runs your small company's entire finance department.
Spare me the "OS Shootouts." Gimme the apps.
Use the Force, or Linux+Unix vs. BeOS/OSX (Score:4, Interesting)
these users do not really know what the hell they're doing in front of a command line interface. They may think they've mastered the shell of Unix or Linux, but they haven't --
because once you have, you will never really have a use for anything else -- the beauty of the shell is that all things and all functions are subsumed below it in consistent fashion, in one magnificent world-view, and all things no matter how complex become possible with a single, well-constructed command, almost like magic.
Some of my fellow Linux or Unix users will understand what I am talking about here -- using the command-line interface is not, as this author says, like carrying around a heavy toolbelt all day when none is needed. Instead, once one has truly mastered the CLI, one is like a Jedi master -- all acts are balanced, rapid, skilled, both intricate and simple at the same time -- and all things are possible and as simple as one another. I can get more work done in ten minutes with my CLI -- including editing video streams and audio streams! -- than most users can get done in days using GUI-only tools.
Of course, OSX and BeOS both have a CLI -- but neither is very useful because much of the rest of the system and the set of standard tools is gutted or malformed in peculiar OSX and BeOS ways. Users of BeOS and OSX think they are getting a CLI, but it's as though they've been trained only by Obi-Wan and never by Yoda -- the real essence of the system is muddied and lost and the benefits are never realized -- or worse -- they are driven from the concept of a CLI unduly.
That is my belief: that users who claim to want a desktop in which CLI use is normally avoided really don't understand and haven't yet mastered the CLI -- because once you have, anything else feels like a straightjacket.
MHO
Re:Metadata Reviewed (Score:3, Interesting)
Thank you for the clarification. Here is what caused my confusion:
"The official Apple recommendation to developers regarding the storage of file type metadata in Mac OS X (as expressed in the Mac OS X System Overview document at the time of this writing) is as follows
In Mac OS X, you indicate the type of a document by specifying two things:
...
The "consequences" of removing a file name extension are actually determined by Mac OS X applications, not by the operating system itself. If I email a Photoshop document named "Logo(Second Revision)" to a Windows user and my email application does not encode the file type information in the file name by appending the appropriate ".psd" file name extension, then the recipient may have trouble opening the file.
d ata-8.html#macosx-file-types
l .html
Unfortunately, Apple does not recommend that applications that move files across platforms behave in this manner. Instead, as we've seen, Apple recommends that Mac OS X applications encode file type metadata in the file name as soon as the file is created. This "solves" the interoperability problem in that any file created in this manner can be sent to another platform without encoding file type metadata in the file name at the time of the transfer. But it requires Mac users to live with file name extension the rest of the time as well.
From: http://arstechnica.com/reviews/01q3/metadata/meta
More info is also available at:http://people.ne.mediaone.net/siracusa/proposa
In any event, I apologize for my stupidity. In any event, what I want is to view files in the Finder and be able to sort by attributes similat to Hacker's Be equivalent.
Hey, I admitted I was wrong, surely a /. first!
Re:Use the Force, or Linux+Unix vs. BeOS/OSX (Score:3, Interesting)
On a techncial level, poor engineering is evident in the CLI's lack of consistency. Nobodies quite sure how formatted output should look. ifconfig looks different from host that looks different to route. Any good CLI should seperate content from presentation, but this is never the case (unless talking about runlevels). Hence `text processing' which is as nasty way of dealing with data in the order of Microsoft Word.
But more importantly: an ordinary computer user writes documents, send email, does archiving, has PDFs top be printed of shown on screen, wants to view web sites with plugins, etc etc etc. Some people just want to get their work done. Sure, they could learn tar, zip, bzip, lha, lhx, their various switches, and learn about piping and redicrection, but maybe they're got actual work to do (remember, the computer is an means to an end, and most people want their means to be easy to pick up and use. I'm know all these command lines switches of the top of my head myself, but remeberingtyping tar -zxvf "whatever" takes longer than clicking the file and hitting enter or clicking three times in KDE to extract it. yes, the GUI saves time. Something that takes multiple uses of ls, sort, and wc is easily accompilished with a single click using Konq's sorted list widget.
You might be a mechanic, others want to drive. And if you didn't build your own car fram scratch I'll bite your troll and call you a hipocrite.
Re:OS Preferences (Score:3, Interesting)
Yes!! Most people have been pointing out that consistency is important in the UI particularly for first time users. Of course UI consistency is usefull even for advanced users - after all even the most advanced user might on occasion use a piece of software that he is not familiar with - if there is no consistency he is not able to take all the knowledge and skills that make him an "advanced user" and apply it to the new unkown application. For that application he is essentially a "first time user" and must struggle through the learning curve all over again. If the UI is consistent he probably already knows how to use it even though he has never laid eyes on it before.
But internal consistency goes beyond just the UI. Consistency is important under the hood too. Why do you think the Linux crowd is always pushing open standards? A standard is simply a way of maintaining consistancy. Without some level of consistency you wouldn't be able to get anything done. A system that is designed as a whole rather than cobbled toegether from a variety of components has the potential advantages of enforced compliance and more comprehensive standards. The decentralized organic evolving "cobbled together" compenents of GNU/linux has other advantages but the more it can be standardised and so become "internally consistent" the better and more useful it will be.
but for the legions of DIY'ers out there, is this something to be worried about in an open-source OS?
That depends: Do you want it just for the sake of being a DIY'er or do you want it to be an effective tool? Do you want it to be an effective tool for other people to use it or is it just for yourself? If actual use is a secondary concern to the joy of doing it for yourself and you don't care if anyone else will use it then consistency is not so important. If on the other hand being a useful tool is important then internal consistency is very important.
A loophole in the OEM License... (Score:3, Interesting)
Of course since this is posted to
Windows, Unix, Mac OSX, BeOS, Amiga SDK, and QNX (Score:2, Interesting)
Ok, that said...let me talk about the features of an OS that are important...I'll take it from the lowest level to the highest.
1. Functionality -- how much stuff your OS can do...i.e., how many operations/manipulations of data, ways to do things, etc.
2. Performance -- when something is operating, how long does it take? How long is load time? Boot time? What about the memory footprint in RAM?
3. Size -- how large is it? Smaller for the same functionality is better. Obviously, smaller progs tend to load and run faster, so this ties into performance.
4. Stability -- this one's pretty obvious. Does it crash or doesn't it? How often does it crash, and how difficult or easy is it to crash it.
5. Security -- related but distinct from stability. How secure can an OS keep your files? i.e., encryptions, permissions, access levels, file sharing, etc.
6. User interface -- this one's composed of several categories. Its not just ease of use, as some Macphiles would have you believe. Ease of use is important. It should also be pretty, so long as the prettyness contributes to making it easier to use & understand (anything beyond that is wasteful). But furthermore, it should allow you to get things done fast. Power features, shortcuts, etc. This is where having a command line and being able to do everything from a keyboard comes in handy. Max OSX may be easy to use, but many tasks are repetitive, and people don't want to constantly have to use the mouse.
7. Compatability -- How much software/hardware/user support does your OS have? This is where M$ gets to kick everyone else in the nuts until their eyes pop out of their head.
8. Of course, their is availability. This is where Linux gets to kick everyone else in the nuts until their eyes pop out. Having something freely available and such that any can see the code is a great benefit. BeOS doesn't get hit as hard, b/c it has a limited version available free of cost (though no source code). M$ doesn't get hit at all -- no fault in their operating system hinders them or costs them money.
Linux and BSD (yes, I know these are DIFFERENT...don't go nuts). These OS' have a great concept behind them -- that the source code should be available for all to see and analyze, and modify on. This also happens to make them free
On to the great Satan, Microsoft Windows. This is an OS which is a prime example of mediocracy and slovenlyness. Most things are OK, some are terrible. MS is all about standards -- that's why its so successful. More simply put, MS is about "popularity". Every hardware vendor makes makes drivers for MS and every software company makes software for MS. As long as this continues, and no other OS' get this kind of support, MS will invariably dominate. The main reason ppl don't switch from MS is because: (1) They've spend hundreds of dollars on Windows games like Descent and Tomb Raider, and don't want to waste that; (2) They have lots of MS software, and don't want to waste that; (3) They want to be able to get all the latest, greatest, and best hardware, which they can always do with MS.
Now, onto the Max OSX. Its all about ease of use. Very easy to use (though annoying not having a right click, and little keyboard menu support). Though easy to use, it is slow -- things open slow, and getting things done is slow, b/c EVERYTHING has to be done with the mouse, or almost so. Very poor performance. Its BSD-core, so good security and stability, if you configure it so. Not too much functionality -- by this, I mean, you can't customize it to your choosing. Very little User control. Apple RAMS their UI down your throat and you better like it or else (cause if you don't, and try to offer programs for modifying MaxOSX's appearance/features on the net, Apple will sue you).
Now, onto three of my favorite proof-of-point OS' in terms of performance: BeOS, Amiga SDK, and QNX. Let me summarize the specialties of each before I treat them all as one cummulative OS. BeOS -- very fast, great for graphics, great file system, fast load-time, boot time, etc. Amiga SDK -- same story as BeOS, but crossplatform and offers interestingly fast VP Assembly code, w/c is crossplatform. Apparently, code runs at near-native speeds once loaded; also, progs written in VP Assembly (w/c is like Java in cross-platformedness) load faster, b/c there is "less" stuff to load from the hard drive, and more CPU transformation (dynamic compilation) of code...CPU much faster than HD, so as far as loading, better to load less and have to "dynamically compile" it than to have to load larger thing to start w/ but not transform it. QNX -- prime example of minimalism: truely, an Orwellian OS in terms of efficiency. No unneeded junk. Now, let me summarize the advantages of these OS: namely, performance performance performance. They boot up quicker than Windows, UNIX, or MacOSX (though QNX is a "UNIX"). Programs load faster on them, tasks are performed faster, and their memory footprint is smaller.
So, what is it the USER really needs?
(1) An OS w/ the PERFORMANCE of BeOS/Amiga/QNX. Fast boot time, fast run time, fast load time, small memory footprint. This comes down to fine tuning and revolutionary thinking in terms of file-systems, algorithms, etc etc. You also need cross-platform code like VP Assembly, w/c can run faster than native code, and w/c can load faster due to less "information" on the HD, w/c needs to be transformed into binary code by the CPU dynamically.
(2) An OS w/ the POWER, FUNCTIONALITY, and CUSTOMIZABILITY of the UNIXs. In Linux/BSD/IRIX/etc, you have enormous power. Everything is customizable. You can customize your browser to selectively ignore certain images on web-sites, etc. Vast array of commands to perform repetitive tasks quickly (such as replacing all instances of ": " in a file with a TAB.
(3) An OS with the EASE OF USE of MacOSX. "Prettyness" is a secondary concern. Prettyness is only something they add to it to make it look better to OEMs. The main concern is to make the interface very intuitive, as well as quick to use. MacOSX tends to be very intuitive, but not very quick to use...you have to drag your mouse to do everything.
(4) An OS with the SOFTWARE SUPPORT, HARDWARE SUPPORT, and general INTERCOMPATABILITY as Windows. As said before, all software companies support Windows, as do all Hardware companies. For software, solutions like Wine may easy to pain for games who already have hundreds of games. But for Hardware? You need to sell companies on that, or make the drivers yourself. How do you sell companies on it? Well, you convince them that b/c your OS is so mean and lean, their product will perform v. fast on it, w/c makes it look good...this only tends to work for gaming and 3D developing software companies, though. But for other companies, doesn't quite have the same effect -- so you have to make it yourself, until your OS becomes popular enough.
What apps, outside of games and 3D progs, do you need? Well, I'll tell you what progs I usually use every day. (1) E-mail prog; (2) Internet browser; (3) Word-processor; (4) Spreadsheet; (5) Database; (6) Drawing/graphics program; (7) Media-viewing program (something that can play ALL kinds of sounds, show ALL kinds of images, and play ALL kinds of videos); (8) Encoders; (9) FileSharing prog; (10) Antivirus; (11) Various scientific utilities. This comes to 11 -- ELEVEN -- programs that I use regularly.
Is itr really that hard for people to come up with 11 GOOD programs which accomidate people's everyday needs? I wouldn't think so.
So, hows all this to be accomplished? Well, I think we start out with the IDEA behind Linux/BSD: you need a free and openly available source code. This gives uers control, and insures a project is immortalized. Maybe you even start out with the BSD or Linux OS?
But, I think thats too difficult. Like BeOS, we need to start from scratch. Our aspirations need to be towards excellence and nothing less. Linux' file system -- while more efficient than Windows and MacOSX -- simply could not be worked to be made as efficient as BeOS'. Granted, Linux has a lot of good things -- OpenSource, and many many useful commands. We shouldn't abandon any of the many many UNIX commands. But we should abandon the Linux file system...in fact, we should abandon all file systems.
It needs to be a clean break -- sometimes, a house is so infested by termites that the only solution is to tear it down and build another house. It won't be easy, and it won't come fast. It certainly won't provide a viable solution for many years...but good things come over time. The pyradmids took lifetimes to build (well, one lifetime of a pharoh, many lifetimes of the avg. Egyptian citizen, since they lived shortly). A good opertaing system may take decades to build -- and that's just to get to the core OS.
But, if you want your efforts to be worthwhile, you have to bite the bullet on one thing -- cross compatability. You need to develop on top of a code which can be run unaltered on any platform, now and in the future. That means something like Amiga SDK's VP Assembly. This does mean a performance hit in terms of run-time once somethings open -- generalized code will never run as quick as a finely-optimized piece of Asm. But it will load faster -- as its basically stored as a smaller executable, which is then translated dynamically. So you optimize the "machine" as much as possible to speed up translation and then bite the bullet on that. This is the only way you'll ever have time to really work on some fundamentals of the file-system and OS, w/o falling vastly behind and finding out your OS can't run on the latest CPU.
Then, you take it one step at a time. First, you plan out the entire system...find new revolutionary ways to make code smaller, more efficient...to make the file system quicker, for example. Of course, to give the user maximal customizability, you need to try to make everything modular. This also makes your OS faster down the line, b/c it can call and load only functions w/c are needed.
Then you proceed logically, first building a solid foundation before building atop it. You don't add new an unnecessary features to a program until you've resolved stability/security issues, as well as performance issues; you also focus firstly on improving performance. Chances are, your prog has all the critical features. LimeWire, for example, doesn't need any more features: it needs to be streamlined. Finally, when adding features -- only add needed and useful features. Don't add features just to "impress people" or make it "look cooler". Add features which are really needed.
If you want an example, lets take MS Word. MS Word had all the features it *needed* in Word 98. Now, MS is just adding new features to impress OEMS. What they really should be doing is making the program smaller, making it run and load faster. Furthermore, they don't need to make it any "easier" to use. It had a simple help system, operated by indexes and contents -- that was great. And a decent menuing and button system. Why did they need to add those stupid office assistants? Only justification, promotion. Dumbing it down to the lowest common denominator. What Word really needs, from MY experience, is faster load times and faster run-times for operations. It also needs more power-shortcuts. Making legends or equations in MS Word is an excercize in "CTRL +"ing or "CTRL SHFIT +"ing...and that's if your an "expert".
As a final note, let me say that I rarely find programs sorely lacking in features. Most progs have plenty of features -- more than you need, in fact. What I do often find is progs that are bloated, huge, slow, and load slowly.
A few additions and corrections. (Score:3, Interesting)
> (although, to be fair, I have lost power on
> this machine and found that it booted back
> up in a normal time span without appearing
> to do anything special).
Mac OS X runs fsck on each and every boot, but because of the way the HFS+ file system is constructed, running fsck multiple times on an 80GB disk takes only a few seconds, so you don't notice it.
If you check a disk with Mac OS X's Disk Utility, it actually runs fsck, and you'll notice it is done in a blink. Same with formatting disks
> The [long] filenames were truncated with garbage
> characters when viewed in the Finder.
They're not actually random garbage characters
> I don't mind AppleScript. I wish the system
> were open to other languages, but
> AppleScript does a fine job, and is very powerful.
The system is open to other languages. What most people call "AppleScript" is actually called "Open Scripting Architecture (OSA)", and AppleScript is just the default language. You can already get a JavaScript plug-in for Mac OS X.
http://www.latenightsw.com/freeware/JavaScriptO
Once installed (drop it in
The Mac OS X Script menu also launches Perl and shell scripts in addition to OSA scripts.
> This is fairly minor, but it seems that some apps
> remember their window positions when closed
> and some do not. Mail.app and Internet Explorer
> do remember their exact size and position
> between runs, but Terminal and many
> others do not. This is another good candidate
> for consistency in the user experience.
Mac OS X can hosts apps with a number of different heritages, so it's definitely true that there is great inconsitency between apps than there was before. As time goes on this will probably get better, as the "Mac OS X way" emerges completely and developers are all familiar with it to some degree.