(trying desperately to remember the quote from Ghost In The Shell)
It's not Microsoft, specifically. The problem is monoculture. No matter what the dominant OS - Windows, Linux, Mac OS, BeOS - the number one guy gets picked on the most, and exploited the most. That creates weakness all the "trustworthy computing" in the world can't fix.
What I fear is some kind of mathematical "reduction" of the problem. "OK," they'll say, "we'll mandate that 30% of stuff move to Linux". OK, great idea: which 30%? "Hmm, you're right. We'll say 10% of web servers, 10% of desktops, and 10% of back-end (DB, etc) stuff." Getting warmer: which 10% of the web servers? Which 10% of the DB servers? Can you get rid of some of your MSSQL on W2k and replace it with Sybase on Linux (easily, with not serious cost and porting problems)? Etcetera, etcetera. I call that "going nowhere fast".
I guess what I'm trying to say here is, I don't really see how to undo the monoculture, when it is backed by 1)such amazing industry power and 2)such entrenched mindset. Figure out how to get people to seriously believe they can run Linux, or Mac, or whatever, and you've gone a long way to solving the problem; but isn't that what people like Microsoft are working just as hard to undo?
I'm sure that there's considerable pressure to weave everything into Microsoft technologies, rather than pure research. Rather that looking at the goal, and evaluating all the technologies out there, they seem to be pretty much constrained to Microsoft technologies and tools.
In other words, if it won't sell, it doesn't happen. There's no pure research. There's no "trickle down" effect: the embedded Linux watch was supremely dumb, but it might have had a positive effect on slimming down the kernel, and illuminated problem areas that the other IBM kernel team (and the world kernel group at large) can look into. Result, better product, maybe. If not, oh well. Look at this neat watch!
Microsoft's focus on profit, profit, profit is laudable but the R&D people need to focus on innovation, dreaming up random crap for the sheer joy of it.
How can a 3rd party fix kernel problems? If they, the owners of the source, no longer support it, and no one else can support kernel-level bugs, and I have to buy the new OS, then I hardly see how it's simple for me.
Re:If developing for windows makes you a sharecrop
on
Don't Be a Sharecropper
·
· Score: 3, Insightful
If I code a game made to work in windows 98, Microsoft can not (at this point) block your game from being run at the OS level (aka "taking away land") but really only through suing you to stop the game from being distributed.
What they can do is put out a service pack (or in the probable case of Longhorn, an entire OS release) that breaks your game. Ideally, you release a patch; the problem is the worst case, where you (the developer) have to go out and get an entire new toolchain (new copy of Visual Studio, etc). Even though update prices are usually modest, you may not want to keep lots of VMWare images on your hard drive, multiple toolchains, etc.
So far, the effect has been minimal: people knew from the start that NT4 wasn't W9x, and things acted differently. However their latest moves are much more bold - Longhorn may be radically different from what we see today.
Unix is as much a collection of behaviors as it is lines of code. Moving from a.out to ELF meant patching and recompiling, sure, but the only investment is time, and in many cases you could do it at your leisure. Commercial software can get EOL'd and you have no choice but to plan your migration (witness the many companies happy with NT4, who are now forced to migrate to W2k or XP).
X11 runs on OS X. GNOME and KDE both run on OS X. So does Evolution, and the GIMP, and tons and tons of other things. Very little "Linux software" refuses to compile on OS X.
Linux definately wins on the "watches and toasters" front, though.
If memory serves correct, Microsoft has been involved in near-continual litigation with various parties including the US government since some time in the 90s.
Quoth Ballmer:Linux itself is a clone of an operating system that is 20-plus years old. That's what it is. That is what you can get today, a clone of a 20-year-old system.
Hmmm... Perhaps he should read this article, wherein the similarities between NT and VMS are brought to light. VMS 1.0 came out in 1977, according to the article.
>User-centered doesn't mean I am forbidden by all means of booting into OS 9 when I need to
Please. The sooner OS9 is forgotten, the better. I understand that there are devices, and applications that haven't been brought forward - Quark for example, and many scanners and printers. That doesn't mean Apple should have to maintain one modern OS, and one legacy OS.
Microsoft doesn't support Win3.1 for a reason, you know. Granted MS did a better job of making printers and scanners work, but the point remains.
Many apps now won't run on OS9, with good reason; either they're hybrid Carbon apps (meaning they use elements of Carbon, but also take advantage of OS X-isms) or Cocoa apps (for which no runtime exists on OS9). However, a quick glance at my Dock (and Fruitmenu) shows way too many apps that still run on OS9 via Carbonlib.
OS X is coming up on it's 4th major release, and still has plenty of warts, but EVERY TIME a developer keeps their "works on OS X and OS8.6+ w/ Carbonlib" alive, they're forcing Apple and the Apple userbase to deal with a world we generally don't want.
It's kind of the same with WINE. I used Linux exclusively for years. Some users I knew, they liked a couple Windows apps here and there. Forte's Agent, for example, was a "must have" for many of them. I'd say, Pan is really good, and uses Agent as inpiration. For years, it was never quite right for them; Agent was "done" and did everything they wanted. Rather than help out Pan, they ran WINE to run Agent (and a few other apps).
Pan lost out on a number of smart, capable users, potentially hindering its growth. Windows, on the other hand, retained users (even if it was via compatibility DLLs and whatnot).
You might say, "It's only a couple of users; Linux can survive that". Maybe, maybe not. I think that when users cling to applications like that, it hurts Linux. Just like people who start Classic on login hurts adoption of Cocoa.
it has been my understanding that there's a really great program-loader for Windows applications. It has native window management support, vendor device support, and lots more!
It's called Windows.
I always thought that WINE was a stopgap, a thing to tide you over until your users were comfortable with OpenOffice or whatever. Now we can run tomorrow's Windows apps today. I can't seem to shake the idea that by running Windows apps on Linux waters down the latter and strengthens the former.
although the original essay talks about Unix and the LISP machines, it just keeps being true. Linus talks about the "charming oddities", well there you go: worse is better. Try for perfection, and the real world will eat you alive.
I also think he's right about the masses being what matter; I think Intel is still thinking about the data centre, not Joe Sixpack, with Itanium.
>You know, some people really shouldn't be using linux. Honest, not trying to be funny or sarcastic; linux is not for everyone.
I know! I mean, someone came to our local LUG meeting using an iBook running Mac OS X! As we were beating the shit out of him, we faintly heard that he uses Linux on his servers, but I think he was just a Commie plant, sent to weaken our resolve.
If we can't keep out the riff-raff, then what good is running Linux? Not trying to be funny or sarcastic, but what if the Dell kid bought a new Dell running Linux? That would be horrible. What if some random CS student somewhere used gcc instead of the free "Introductory" copy of Visual Studio that came in their textbook? Horrible. That person might not really "get it" and instead just be a user.
I'm so glad that Linux is done, and there is no longer any further need for attention on matters like usability. If regular users got a hold of it.. whew. They might be able to use it. We can't have that.
He's one of the implementors of the original Netscape on Unix/X11, part of the Mozilla team (IIRC he actually named it Mozilla), and part of the original XEmacs team.
Not to mention xscreensaver (which, I think, ships with every Linux distro out there), and a few other cool hacks.
A former roommate of mine got in on the ground floor. One night he cancels all social events, due to some guild event.
Later he walks into the kitchen and starts making dinner. He then consumes dinner, in the living room, watching TV. It occurs to me that he cancelled all social events, including an evening with his girlfriend. "Oh, did the meet get cancelled or postponed?"
"No, we're fighting right now. I just put it on autoattack. It'll take another 20 minutes or so to win, so I'll check on it later."
Um, OK.
I remember when some class finally got the power it was promised when the game was released; but it cost too much and didn't work right, anyway. Roommate was trying to show me how cool it was, but it made litle sense and was anticlimactic (you did something like 10 extra points of damage every how and then).
I remember trains, and chatty people, and all that. I remember my roommate developing a whole new vocabulary - bubs and zoning and whatever else. I remember patches with reputations worse than NT service packs, and useless tech support, and dropped connections, and lost characters due to server-side upgrades, and more.
This was all before the first expansion pack!
I say, yes, the game sucks you in with a feedback loop you can't escape from (I think it's called a Skinner Box?). I knew all this before the first expansion came out. While I agree with the entire screed (from an observer POV, anyway) I can't understand how it took you that long to realize all this. Most people jumped off a while ago, and the rest are getting exactly what they deserve.
Man, this thread is getting harsh. I mean, nearly ever sane retailer has SOME means to let employees know when someone is the store. yeah, the dinger SUCKS, but I'd rather know when someone is coming in.
The CueCat was a stupid venture, to be sure; but at some point in their lifespan, every retailer does something stupid at some point.
Yes, Radio Shack has morphed from hobbyist products and radio gear to basically a smaller, less well stocked Best Buy. But can you blame them? While many slashdotters may in fact still need diodes, Joe Average doesn't. The death of the electronics hobbyist almost killed the company. They are trying to stay alive in the face of serious competition while retaining what used to make the Shack a place to buy stuff. If you have a better idea, a way to make the company really stand out, get a job there and tell the boss. They might even listen.
And no one is really commenting on the fact that a high-profile retailer like the Shack taking a step like this may, in fact, influence others to drop their mailing. I can't buy anything anymore without a request for zip code, or some other deal.
Also Note: the Shack has one of the most tolerant, liberal intellectual property waivers ever. Unless you invent a new point-of-sale system, and do so on the job (or using work-provided materials), they don't give a crap. I mean, Best Buy would probably try to make you turn over your latest patch to BitchX but the Shack doesn't care.
Yes, nearly every store is still a commission-based store. The smallest stores are paid a regular rate, plus a bonus for exceeding sales targets.
Let's look at the trends.
on
Is Mac OS X Slow?
·
· Score: 5, Insightful
10.0.0 Public Beta was barely usable, in every way. It was beyond slow. It was almost a toy. The genie effect took forever.
10.0.0 release was slow. It was a pain.
10.1.0 was improved; my machines are quite old, and it showed.
10.1.5 was improved; as the last of the 10.1 branch, it showed improvement.
10.2 brought a noticeable improvement. I wasn't spurting my shorts but I could not recommend it to others without hesitation, with the exception of the guys that buy a new CPU every time AMD or Intel comes out with one, because the old was one "just too slow". Whatever.
Is everyone seeing the trend? Getting better all the time. I forgot who did the presentation, but the quote was along the line of, "We have to improve in software because we can't trust Motorola to speed up the hardware". Each new release boosts performance on the same hardware with no noticeable new bugs or problems (other than what Apple introduces on purpose, like breaking LiteSwitch w/ 10.2).
In short: it's sad that the unacceptable performance of older versions, esp. betas, has tainted a great OS with the moniker "slow".
Slashdot Mirror
(trying desperately to remember the quote from Ghost In The Shell)
It's not Microsoft, specifically. The problem is monoculture. No matter what the dominant OS - Windows, Linux, Mac OS, BeOS - the number one guy gets picked on the most, and exploited the most. That creates weakness all the "trustworthy computing" in the world can't fix.
What I fear is some kind of mathematical "reduction" of the problem. "OK," they'll say, "we'll mandate that 30% of stuff move to Linux". OK, great idea: which 30%? "Hmm, you're right. We'll say 10% of web servers, 10% of desktops, and 10% of back-end (DB, etc) stuff." Getting warmer: which 10% of the web servers? Which 10% of the DB servers? Can you get rid of some of your MSSQL on W2k and replace it with Sybase on Linux (easily, with not serious cost and porting problems)? Etcetera, etcetera. I call that "going nowhere fast".
I guess what I'm trying to say here is, I don't really see how to undo the monoculture, when it is backed by 1)such amazing industry power and 2)such entrenched mindset. Figure out how to get people to seriously believe they can run Linux, or Mac, or whatever, and you've gone a long way to solving the problem; but isn't that what people like Microsoft are working just as hard to undo?
After all, The First Rule of Bulk Club is, You do not talk about Bulk Club.
I'm sure that there's considerable pressure to weave everything into Microsoft technologies, rather than pure research. Rather that looking at the goal, and evaluating all the technologies out there, they seem to be pretty much constrained to Microsoft technologies and tools.
In other words, if it won't sell, it doesn't happen. There's no pure research. There's no "trickle down" effect: the embedded Linux watch was supremely dumb, but it might have had a positive effect on slimming down the kernel, and illuminated problem areas that the other IBM kernel team (and the world kernel group at large) can look into. Result, better product, maybe. If not, oh well. Look at this neat watch!
Microsoft's focus on profit, profit, profit is laudable but the R&D people need to focus on innovation, dreaming up random crap for the sheer joy of it.
That's "subscriber-dotted" now.
How can a 3rd party fix kernel problems? If they, the owners of the source, no longer support it, and no one else can support kernel-level bugs, and I have to buy the new OS, then I hardly see how it's simple for me.
What they can do is put out a service pack (or in the probable case of Longhorn, an entire OS release) that breaks your game. Ideally, you release a patch; the problem is the worst case, where you (the developer) have to go out and get an entire new toolchain (new copy of Visual Studio, etc). Even though update prices are usually modest, you may not want to keep lots of VMWare images on your hard drive, multiple toolchains, etc.
So far, the effect has been minimal: people knew from the start that NT4 wasn't W9x, and things acted differently. However their latest moves are much more bold - Longhorn may be radically different from what we see today.
Unix is as much a collection of behaviors as it is lines of code. Moving from a.out to ELF meant patching and recompiling, sure, but the only investment is time, and in many cases you could do it at your leisure. Commercial software can get EOL'd and you have no choice but to plan your migration (witness the many companies happy with NT4, who are now forced to migrate to W2k or XP).
What's "Linux Software"?
X11 runs on OS X. GNOME and KDE both run on OS X. So does Evolution, and the GIMP, and tons and tons of other things. Very little "Linux software" refuses to compile on OS X.
Linux definately wins on the "watches and toasters" front, though.
If memory serves correct, Microsoft has been involved in near-continual litigation with various parties including the US government since some time in the 90s.
Quoth Ballmer:Linux itself is a clone of an operating system that is 20-plus years old. That's what it is. That is what you can get today, a clone of a 20-year-old system.
Hmmm... Perhaps he should read this article, wherein the similarities between NT and VMS are brought to light. VMS 1.0 came out in 1977, according to the article.
In summary, glass houses, stones, etc.
Touche, but, the point still stands. Apple's borked "next gen OS" strategy was to blame for OS9 overstaying its welcome.
>User-centered doesn't mean I am forbidden by all means of booting into OS 9 when I need to
Please. The sooner OS9 is forgotten, the better. I understand that there are devices, and applications that haven't been brought forward - Quark for example, and many scanners and printers. That doesn't mean Apple should have to maintain one modern OS, and one legacy OS.
Microsoft doesn't support Win3.1 for a reason, you know. Granted MS did a better job of making printers and scanners work, but the point remains.
Carbon shouldn't be dropped - Office v.X is a Carbon app, for example - but the "run on OS9" stuff should go.
There's a different analogy with OS X.
Many apps now won't run on OS9, with good reason; either they're hybrid Carbon apps (meaning they use elements of Carbon, but also take advantage of OS X-isms) or Cocoa apps (for which no runtime exists on OS9). However, a quick glance at my Dock (and Fruitmenu) shows way too many apps that still run on OS9 via Carbonlib.
OS X is coming up on it's 4th major release, and still has plenty of warts, but EVERY TIME a developer keeps their "works on OS X and OS8.6+ w/ Carbonlib" alive, they're forcing Apple and the Apple userbase to deal with a world we generally don't want.
It's kind of the same with WINE. I used Linux exclusively for years. Some users I knew, they liked a couple Windows apps here and there. Forte's Agent, for example, was a "must have" for many of them. I'd say, Pan is really good, and uses Agent as inpiration. For years, it was never quite right for them; Agent was "done" and did everything they wanted. Rather than help out Pan, they ran WINE to run Agent (and a few other apps).
Pan lost out on a number of smart, capable users, potentially hindering its growth. Windows, on the other hand, retained users (even if it was via compatibility DLLs and whatnot).
You might say, "It's only a couple of users; Linux can survive that". Maybe, maybe not. I think that when users cling to applications like that, it hurts Linux. Just like people who start Classic on login hurts adoption of Cocoa.
it has been my understanding that there's a really great program-loader for Windows applications. It has native window management support, vendor device support, and lots more!
It's called Windows.
I always thought that WINE was a stopgap, a thing to tide you over until your users were comfortable with OpenOffice or whatever. Now we can run tomorrow's Windows apps today. I can't seem to shake the idea that by running Windows apps on Linux waters down the latter and strengthens the former.
Worse is better
although the original essay talks about Unix and the LISP machines, it just keeps being true. Linus talks about the "charming oddities", well there you go: worse is better. Try for perfection, and the real world will eat you alive.
I also think he's right about the masses being what matter; I think Intel is still thinking about the data centre, not Joe Sixpack, with Itanium.
Not open source?
I wonder what that source download is.
Or mod_rendezvous for Apache.
Or the CVS access.
>You know, some people really shouldn't be using linux. Honest, not trying to be funny or sarcastic; linux is not for everyone.
I know! I mean, someone came to our local LUG meeting using an iBook running Mac OS X! As we were beating the shit out of him, we faintly heard that he uses Linux on his servers, but I think he was just a Commie plant, sent to weaken our resolve.
If we can't keep out the riff-raff, then what good is running Linux? Not trying to be funny or sarcastic, but what if the Dell kid bought a new Dell running Linux? That would be horrible. What if some random CS student somewhere used gcc instead of the free "Introductory" copy of Visual Studio that came in their textbook? Horrible. That person might not really "get it" and instead just be a user.
I'm so glad that Linux is done, and there is no longer any further need for attention on matters like usability. If regular users got a hold of it.. whew. They might be able to use it. We can't have that.
He's one of the implementors of the original Netscape on Unix/X11, part of the Mozilla team (IIRC he actually named it Mozilla), and part of the original XEmacs team.
Not to mention xscreensaver (which, I think, ships with every Linux distro out there), and a few other cool hacks.
>JWZ should learn how to program
Implementing Netscape on Unix, and working on Mozilla, doesn't count?
Sir, it seems the entire 3rd Division is feeling a little anxious.
You will also note that most of the Marines in the 187th Expeditionary Force are also nervous about their impending landing.
Lastly sir, the pilots have... the heebee jeebies.
Impending War/Risk of painful death == anxiety
A former roommate of mine got in on the ground floor. One night he cancels all social events, due to some guild event.
Later he walks into the kitchen and starts making dinner. He then consumes dinner, in the living room, watching TV. It occurs to me that he cancelled all social events, including an evening with his girlfriend. "Oh, did the meet get cancelled or postponed?"
"No, we're fighting right now. I just put it on autoattack. It'll take another 20 minutes or so to win, so I'll check on it later."
Um, OK.
I remember when some class finally got the power it was promised when the game was released; but it cost too much and didn't work right, anyway. Roommate was trying to show me how cool it was, but it made litle sense and was anticlimactic (you did something like 10 extra points of damage every how and then).
I remember trains, and chatty people, and all that. I remember my roommate developing a whole new vocabulary - bubs and zoning and whatever else. I remember patches with reputations worse than NT service packs, and useless tech support, and dropped connections, and lost characters due to server-side upgrades, and more.
This was all before the first expansion pack!
I say, yes, the game sucks you in with a feedback loop you can't escape from (I think it's called a Skinner Box?). I knew all this before the first expansion came out. While I agree with the entire screed (from an observer POV, anyway) I can't understand how it took you that long to realize all this. Most people jumped off a while ago, and the rest are getting exactly what they deserve.
Man, this thread is getting harsh. I mean, nearly ever sane retailer has SOME means to let employees know when someone is the store. yeah, the dinger SUCKS, but I'd rather know when someone is coming in.
The CueCat was a stupid venture, to be sure; but at some point in their lifespan, every retailer does something stupid at some point.
Yes, Radio Shack has morphed from hobbyist products and radio gear to basically a smaller, less well stocked Best Buy. But can you blame them? While many slashdotters may in fact still need diodes, Joe Average doesn't. The death of the electronics hobbyist almost killed the company. They are trying to stay alive in the face of serious competition while retaining what used to make the Shack a place to buy stuff. If you have a better idea, a way to make the company really stand out, get a job there and tell the boss. They might even listen.
And no one is really commenting on the fact that a high-profile retailer like the Shack taking a step like this may, in fact, influence others to drop their mailing. I can't buy anything anymore without a request for zip code, or some other deal.
Also Note: the Shack has one of the most tolerant, liberal intellectual property waivers ever. Unless you invent a new point-of-sale system, and do so on the job (or using work-provided materials), they don't give a crap. I mean, Best Buy would probably try to make you turn over your latest patch to BitchX but the Shack doesn't care.
Yes, nearly every store is still a commission-based store. The smallest stores are paid a regular rate, plus a bonus for exceeding sales targets.
10.0.0 Public Beta was barely usable, in every way. It was beyond slow. It was almost a toy. The genie effect took forever.
10.0.0 release was slow. It was a pain.
10.1.0 was improved; my machines are quite old, and it showed.
10.1.5 was improved; as the last of the 10.1 branch, it showed improvement.
10.2 brought a noticeable improvement. I wasn't spurting my shorts but I could not recommend it to others without hesitation, with the exception of the guys that buy a new CPU every time AMD or Intel comes out with one, because the old was one "just too slow". Whatever.
Is everyone seeing the trend? Getting better all the time. I forgot who did the presentation, but the quote was along the line of, "We have to improve in software because we can't trust Motorola to speed up the hardware". Each new release boosts performance on the same hardware with no noticeable new bugs or problems (other than what Apple introduces on purpose, like breaking LiteSwitch w/ 10.2).
In short: it's sad that the unacceptable performance of older versions, esp. betas, has tainted a great OS with the moniker "slow".
JWZ's Law Of Software Envelopment states that all programs expand until they can read mail.
I posit a corollary, the Law Of GUI Envelopment: all windows will eventually be round.
Aqua has rounded corners. Now XP does. What, do they think they sharp edges might hurt someone?