Slashdot Log In
Linux Can't Kill Windows
Posted by
Zonk
on Thu Apr 14, 2005 07:49 AM
from the that's-his-opinion-he-could-be-wrong dept.
from the that's-his-opinion-he-could-be-wrong dept.
nberardi writes "Infoworld is running an article in which the author claims 'Linux is established and has a niche that, as various pendulums swing, will grow and shrink. Show me charts and stats and benchmarks that prove Linux superior to Windows in every measure and I'll not argue with you. But no matter how much money and dedication is poured into Linux, it will never put a dent in Windows' mind share or market share because Linux is an operating system, a way -- and probably the best way -- to make system hardware do what it's told. But you can't turn Linux into a platform even if you brand it, box it, and put a pricey sticker on it.'"
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Excellent Article! (Score:5, Funny)
Seems like a well-thought out article that certainly wasn't created for the purpose of increasing impressions or generating clicks to advertisers on the site.
Re:Excellent Article! (Score:4, Informative)
Linux isn't scalable ? It runs on everything from ARM to huge supercomputer clusters.
Consistent ? I will give it to him that across distributions linux is not consistent but businesses use RHEL or Novell against which all major applications like Oracle are certified.Within these distributions things are largely consistent.
Predictable ? What is unpredictable about Linux ?
What does self-contained mean ?
Doesn't this article give the feeling the author has no clue about what he is talking about and has just put together some buzzwords like scalable, self-contained to create a controversial article?
Parent
Re:Excellent Article! (Score:5, Funny)
They're all integral parts of the OS, after all.
Parent
Re:Excellent Article! (Score:4, Funny)
Parent
User interfaces are important, though (Score:5, Informative)
To most end users, a consistent look and feel, that works right out of the box, is really important. So it's a very good thing that Linux distributions are improving in this area (which the article conveniently forgets to mention).
For the same reason, I also think it's good to see Open Source applications adopting user interfaces that are more similar to their Windows counterparts. It may annoy some old-time Unix or Linux users to find "Options" under "Tools" rather than under "Edit" in the Firefox browser.
But for Windows users that are looking for a safer alternative to their present browser, the chance that they'll make the switch increases with every item that works as expected when they first try it out.
And it's only by convincing today's Windows users to switch, that Linux can avoid the fate that the article spells out.
Parent
Re:User interfaces are important, though (Score:5, Insightful)
Here is where you miss the point of "out of the box".
When your average idiot buys a computer from Dell, Gateway, HP, $RESELLER. He gets it home, opens the box, plugs it in, and lo and behold it WORKS. That is what the phrase means. All Joe Luser knows about Windows is that he buys a computer and turns it on and it WORKS.
Currently, you cannot do this with Linux. Mainly because almost no one sells preconfigured Linux boxes that you can just plug in and work. Lindows barely scratches the surface.
Installing Windows from scratch is a totally different story. It is, as you say, fraught with perils. But this is the same no matter what OS you try.
Joe Luser doesn't care about installing windows or any other OS. He wants a computer that he plugs in and it WORKS. He gets this from all major distributors.
Until a major reseller can offer a Linux PC that does the exact same thing CONSISTENTLY, Joe Luser will not use Linux.
Parent
Re:User interfaces are important, though (Score:5, Insightful)
> Open Excel and do some work?
I have a cheapo e-machine I bought to run Windows games on (at which it has done surprisingly well, I might add). It came with Windows Works, which is not unusual. Joe Luser gets home, plugs it in, and he's got a spreadsheet. Not a terribly good one, but Joe doesn't know the difference.
> Watch some DVD's?
It also came with PowerDVD 5, which is even more common than getting Works. Actually, it plays DVDs better than any of my Linux boxes, and did so right out of the box.
> Browse the internet risk free?
No, but Joe doesn't know this and can't see it. He double clicks on Internet Explorer, and it's teh Intarweb! Works right out of the box!
> No, he can't do any of those things "out of the box".
Actually, yes, as far as Joe can see, he *can* do all those things right out of the box He doesn't see how poorly or brokenly they may be done. All he sees is that he can't buy a Linux box that he can just plug in and have do these things with no requirement that he do things he doesn't understand.
Chris Mattern
Parent
Linux is the rebellion of the intellectuals (Score:5, Informative)
Linux changed that. Computer professionals are telling management that they will work with one standard OS. Their OS. Designing and building it themselves and distributing it freely is a brilliant strategy to counter management's claim that some other OS was cheaper.
All this happened concurrently with the widespread introduction of powerful inexpensive desktop computers into the workspace. Office computing adopted the Windows OS in order to maximize the productivity gains that could only be achieved by having the entire world adopt a single standard. An incredible stroke of luck for the company selling that standard. The price went to the company that was the most relentless and focused on forcing the world to adopt their standard. That company was also flexible and intelligent enough to integrate huge positive feedback loops into the process of getting the world to adopt its product. The astonishing success of the company in selling a product that the world was desperate to buy doesnt mean that they can do it again with another type of product.
The widespread introduction of powerful inexpensive desktop computers was predicated on the condition that the performance/price ratio of the PCs would double every few years.
The current problems that result from the conversion of all other Operating Systems to Linux are temporary. They are being addressed; they will be solved. The widespread introduction of powerful inexpensive desktop computers was predicated on the condition that the performance/price ratio of the PCs would double every few years. The entire next generation of desktop computers may find their doubling of power completely dedicated to transition from Windows to Linux. In other words, it may take a doubling of computer power to make Windows applications run on Linux with the same speed and efficiency that they currently run on Windows OS. This will be denounced as a complete waste by IT professionals. Theyre correct, but it will be a necessary step anyway.
Parent
Re:Excellent Article! (Score:4, Funny)
Mabye this is why I can't seem to get a submission accepted...I'm just not being inflammatory enough.
Parent
Re:Excellent Article! (Score:5, Funny)
Fun for the whole family.
Parent
Re:Excellent Article! (Score:5, Insightful)
False analogy.
Let me make it clearer to you by making the following two statements:
1. Linux revolves around the kernel. Every time you muck with the kernel to bring about yet another set of "gee whiz bang" features, dozens of things are broken.
2. Mac OS X and Windows revolve around the interface. On the library level, new interfaces are added, but older ones are still supported for a surprisingly long time (see Carbon / Classic Runtime Environment for Mac OS X, or Win9x Compatibility Mode / Application Compatibility Toolkit for Windows 2000 / XP). Certainly, support is eventually dropped, but the pace is normally quite slow for popular APIs.
On a visual interface level, both Apple and MS try to keep consistency in the interface. Sure, you'll see major changes in interface every 5-10 years (Windows 95, Windows XP, Mac OS X), but that's a pace most people can cope with, and they try not to change EVERYTHING in the process. Linux, on the other hand: for any random distro, you can't be assurred GUI consistency.
Tell me, how many people really know if there were major kernel revisions between all the Mac OX X releases? I imagine not many, because programmers don't have to care. That's the beauty of revolving around interfaces.
Until Linux stops revolving around the kernel, it will never break out of the server niche.
Parent
Long term impact (Score:4, Insightful)
But history has shown that the short term impact of most new things tend to be over-estimated, whereas the long term impact tends to be under-estimated.
Who knows where Linux will be in 20 years? I sure as hell don't, but I have a rather optimistic view.
You should be optimisitic (Score:5, Interesting)
Now having said that, what I see more off are peacock articles. All fluff and very little facts because the three operating systems are TOO similar. Compare it to cars. These days all of the cars are good enough! They will last four years without too many problems. So then how do you distinguish yourself? Write articles like a peacock struts its feathers, all emotional.
The easiest way to illustrate this peacock argument is to take a bushman from the jungle and get them to figure out what a computer does. Without helping them. My guess is that the bushman will have a hard time figuring out what the mouse is for. Most likely they will use the mouse as a slingshot and head back into the jungle. I am not saying that bushmen are dumb. I am saying that computers require some upfront learning time regardless of the OS used.
Parent
Re:You should be optimisitic (Score:5, Informative)
3. there's an option in Keyboard & Mouse->Keyboard shortcuts to "turn on full access". tried that?
4. tried clicking the down arrow to the right of the save name?
Parent
Re:Long term impact (Score:5, Funny)
Parent
Re:Long term impact (Score:5, Funny)
Parent
Re:Long term impact (Score:5, Funny)
It will be replaced by mentats.
Parent
Mindset (Score:5, Insightful)
For example, my dad is a Windows person, and his SO has a Mac with OS X. He can't seem to understand how OS X works, so he dissmisses it and claims that Windows is better (on the fact that he knows how to use Windows).
It's not that Windows is "special", it's just that that's all most people know. And half those people don't know much, if anything, about Windows anyway, so it's no wonder Linux has a difficult time trying to enter the mainstream market.
Re:Mindset (Score:5, Insightful)
Excellent point. Any OS is "difficult to learn" to a complete newbie. Someone familiar with only one OS will think that OS is the greatest and everything else is "subpar". While those users who know two or more OSes well can more easily transition from one to another, even to a totally new and unfamiliar OS. Therefore, in order for Linux or OSX to really make a major dent in the desktop arena, users need to be exposed and educated about them. That, of course, requires that the in-fighting between the various Linux distro fanboys needs to be put aside and join forces to make this happen. And that is a huge hurdle to overcome.
Parent
Re:Mindset (Score:5, Interesting)
Second point is i got my wife using tools for her everyday tasks that exist on both OS's. She isnt a power user either, most of what she does is her mommies gorups, emails, web pages, gaim and little photo editing etc etc. All of which she used open source packages to do on windows. I decided to rebuild her downstairs PC with Gentoo. Took her a day to get used to KDE, and where to find her programs. Now she just does what she used too. She doesnt miss Win XP and couldnt care less that she is using Gentoo.
Kinda sad that I'm the tech guy and I'm the only XP user left in the house. Damn EQ2 and its inability to run on Linux.. hehe.
Parent
I'm sorry (Score:5, Interesting)
My long term projection would be, that Linux will push Windows into a third of the market, something like 1/3 linux, 1/3 windows and 1/3 else.
This article is -1 flamebait (Score:5, Insightful)
Simon.
Re:This article is -1 flamebait (Score:4, Funny)
Parent
I disagree (Score:4, Insightful)
What does branding it, boxing it and putting on a price tag, have to do with a tool doing a job?
Re:I disagree (Score:5, Funny)
Who? The editor?
Parent
Vaguest article... (Score:5, Insightful)
what
Very poor indeed.
I disagree (Score:5, Interesting)
Re:I disagree (Score:5, Interesting)
All I had displayed was the fluxbox window manager with firefox, gvim, and a matplotlib window from a python session.
I had to switch vterms to convince him, as I was running Linux, as he also assumed Linux was all CLI.
He should've known better too: He wasn't some PHB, but someone who used X11 and fink under OS X! If those who are as technically literate as this don't get Linux, how will the "average consumer" ever get it?
Parent
Re:I disagree (Score:5, Funny)
Announcer: We're here at the Vigneswara Call Processing Center in Bangalore, India, where we've secretly replaced the customer service reps' Windows XP with Linux. Let's listen in.
Operator: Wow! That went completely smoothly.
Announcer: Did you know we replaced your Windows with Linux?
Operator: Impossible! Where's the bitter CLI taste?
Announcer (tapping the keyboard a few times): Right here!
Operator: Amazing! Can I work a third consecutive shift please?
Supervisor (shocked): they never ask for a third shift with Windows!
Parent
Fight network effects (Score:5, Insightful)
The web is a first step.
XUL and other technologies like thsi is one step is the right direction.
Open and RF standards are also a key in this process.
If you want to RTFA, but give no ad click bonus... (Score:5, Informative)
One fundamental difference guarantees that Windows will continue to dominate
By Tom Yager
April 13, 2005
You can quit proclaiming Linux the Windows killer.
ADVERTISEMENT
Linux is established and has a niche that, as various pendulums swing, will grow and shrink. Show me charts and stats and benchmarks that prove Linux superior to Windows in every measure and I'll not argue with you. But no matter how much money and dedication is poured into Linux, it will never put a dent in Windows' mind share or market share because Linux is an operating system, a way -- and probably the best way -- to make system hardware do what it's told. But you can't turn Linux into a platform even if you brand it, box it, and put a pricey sticker on it.
Businesses and organizations of all sizes need consistent, predictable, scalable, self-contained platforms for server solutions. Windows wins. Linux doesn't lose, because it can continue the legacy of another nonplatform, namely Unix, that needs to be refreshed and extended.
The practical need to keep Unix around isn't rooted in nostalgia or misguided conviction. There may be times when you're convinced that the solution you need doesn't exist as a whole. The total solutions that exist might be too confining or expensive, or -- as is sometimes the showstopper for me -- simply closed. Open source Unix, in which category I place Linux, BSD, and Darwin (the OS layer of Apple's OS X), is a 500,000-piece bag of Legos that comes with some drawings and a few models you can use, build on, or tap into as references for your own creations. On paper, an OS is an ideal place to start building, because you get to choose everything that sits above it and presumably you know just what belongs in each of those gaps between your hardware and your application. You see, while developers can write to an operating system's default API, they'll spend most of their time encapsulating and abstracting low-level system calls to create what is, in effect, an application platform.
No one is so foolish as to make what can be acquired cheaply or free; it's wiser to pick one from among hundreds of platforms and modules that fill in the holes between open source Unix and your applications.
In contrast, Windows fills in all the blocks between the hardware and your apps. It does it in ways that you can't alter, but which you can use in different ways. You can code with the tools of your choice and in the programming language of your choice, and unless you stray too far from the rule book, everything you create will interoperate with everything others write for Windows. An operating system is a rack into which device drivers and APIs are inserted. A platform is a rack into which applications are inserted.
Linux and Windows don't compete. Sun Microsystems (Profile, Products, Articles) sees this as an opportunity and has struggled mightily to position the combination of Solaris and Java as a platform. It almost makes it. I'd choose J2EE and Solaris over Linux for nonuser-facing server applications in shops that have expert administrators. But, similar to Linux and other flavors of Unix, Solaris is a nonstarter on clients, and that's enough to hurt its capability of competing with Windows. There is only one platform that can stand toe-to-toe with Windows, and that's the combination of OS X and Java.
Stay tuned; I'll tell you all about it.
it's not about killng (Score:5, Interesting)
There's too many people who are interested with tinkering.. with having something being totally customizable if they take their time. With being free and able to run their computer the way they want. Is this the majority of people? Not even close! But it's enough that Linux will sustain itself in spite of any FUD MS and crew would throw at it.
Who cares if Linux never overtakes Windows? I know before I discovered it in '98, I thought I was doomed to the endless update/virus/adware world that everyone else was in (except those crazy mac people.. which now due to the mac mini I am one as well.. side tracking....)
Anyway, the point being.. Linux is strong due to it's following, and has great potential to do quite a few things Windows has troubles with. The choice is there for anyone to pick up that option if they so choose. What's the big deal?
Wrong wrong wrong (Score:5, Interesting)
From the article:
Businesses and organizations of all sizes need consistent, predictable, scalable, self-contained platforms for server solutions.
I thought Windows was winning on the desktop? Isn't that what we're always hearing?
Linux and Windows don't compete.
Ok, so the whole "Get The Facts" campaign was done just for grins?
Open source Unix, in which category I place Linux, BSD, and Darwin (the OS layer of Apple's OS X), is a 500,000-piece bag of Legos that comes with some drawings and a few models you can use, build on, or tap into as references for your own creations.
Also wrong. There are distros that are like that, but there are distros that aren't. Linux offers choice, and not just the "bag of Legos" kind.
And, just in case the article author reads this...ever hear of Wine? As soon as Wine gets DCOM working correctly and Installshield working right, it won't matter to Joe User if the OS is Linux or Windows, just so long as he can install TurboTax and Doom3. Check back in a few years, and we'll see if you're singing a different tune.
True, but for the wrong reasons (Score:5, Insightful)
Even the concept of "competition" is a straw man.
Linux represents a total, brutal, and unstoppable commoditization of technology that follows the same rules which drive "Moore's Law". When you remove the costs of improving a technology, its marginal cost will fall to zero as people compete to be the key suppliers.
Software is basically becoming free, and this is what will kill Windows, whether or not it's something called "Linux" that takes over.
Most likely, "Linux" will never become more than a niche OS, excellent for servers but rare for desktops. But what it represents - unlimited and perfect software at no cost - will, inevitably, rule the desktop as it will rule every single computing platform, for the simple reason that no amount of lock-in or marketing is going to get people to keep paying more than the going rate for a commodity.
Apple's strategy - where the OS and a bunch of software is basically thrown in for free - is the trend of the future.
I hate to say it, because I truly love using Microsoft's well-engineered products, but between the commoditization of their core markets and the parasites eating their way in from the internets, they are dead, Linux or no Linux.
Before everyone flames him (Score:5, Interesting)
It sounds to me like what he's really saying is (Score:5, Interesting)
So let's imagine some company, we'll call them Red Hat, to pull a bogus name out of thin air, and let's say they were to take this Linux thing, and make a nice standardized platform out of it. People ship you an application, you take your server, we'll call it a "Red Hat Enterprise Server" or something like that, and you can simply load the app on it and run it. They wouldn't say their app runs on Linux. They'd say their app runs on Red Hat.
To him, _that_ would be a platform, and that would have a chance at taking on Windows. It would be Linux behind the scenes, but it's more that just Linux.
Too bad nobody's ever going to do something like that.
-JDF
[0] Thankfully, even if you generally only see one of these, you can still have the other behind the scenes and run stuff intended for either...
RE: Linux Can't Kill Windows (Score:5, Funny)
I see someone didn't try to dual boot Fedora Core 2
So what? (Score:5, Insightful)
A little history... (Score:5, Insightful)
1996 : Linux? So it makes a simple web server. It'll never scale as an enterprise server.
2001 : Linux? Yeah, it's nice for my enterprise servers, but it'll never give end-users any satisfaction.
2005 : Linux? So hackers have pretty desktop. Didja see the effort they had to go to make it work? It'll never be easy enough for our secretary Jane Typist.
Nope, Linux will never compete. Not even that Novell Linux Desktop that has proliferated our workplace and made every desktop look the same (but secure). It'll never happen.
Re:I think he's right (Score:5, Interesting)
Debian would be a platform, or Novell/SUSE, or RedHat - if they finally committed themselves to being one.
A platform is a platform only if its stable, and I don't mean "stable" as in "does not crash". I mean "stable" as in "does not change significantly every 6 months". So Debian would be an ideal choice.
However, Debian itself has zero commercial drive. I wonder what drives Debian at all, and other people wonder, too, given the admirable rise of Ubuntu.
But people want pretty software, and Debian stable features GNOME and a stone-aged KDE. And while GNOME on Debian seems to be more advanced than KDE, forgive me, I would not chose it for fancy software. It looks so painfully dull
KDE on the other hand looks nice and lively, but is it a platform? I wonder.
Obviously there has to be a balance between the drive forward, the wish to leave behind all that old cruft (fsck compatibility!) and the conservative approach to not chance anything to not break compatibility.
Still, what drives the PC market is cool software, not cool licensing.
Just to give you an example: mplayer is technically cool. But its complexity scares people away. It's only cool because it's free. You won't be able to sell it to anybody, because as a software _product_ it sucks. badly. Even with gmplayer.
Or take GIMP. It's cool 'cause it's free. But it's just an aggregation of image manipulation tools. It's not a _product_.
There is this small gap between a program and a product that Open Source software seems to be unable to bridge, this final, annoying, painful step of really _finishing_ it so that it _could_ be sold.
My conclusion: Linux needs commercial(-grade) software. Firefox is not enough. Instead of scaring commercial software vendors away with stupid fundamentalism we should be fair with them.
Parent
Re:I think he's right (Score:5, Insightful)
Yeah, a moving platform. With countless widget sets, multiple clipboards, different directory structures, an infinite number of combinations and permutations of shared libraries, and just as many sources of outdated, incorrect, misleading or utterly superb documentation, and crap vendors like Redhat which drop version support in a third the time of Microsoft.
One place where GNU/Linux is relatively stable is in POSIX and a vague semblence of commonly accepted extensions to the standard. That makes it a nice platform for server software, but does nothing on the desktop.
Windows was never an OS. It contains an OS, they changed OSes in the product lifetime, but the product has always been a desktop environment and a consistent, well documented, and long-supported API.
Parent
Re:I think he's right (Score:5, Insightful)
To take the points in order:
countless widget sets
A few major widget sets. If you're going to include every kit, you might as well include the buttons here in Opera, which are completely non-standard as far as Windows is concerned.
multiple clipboards
Yes, annoying and stupid.
different directory structures, an infinite number of combinations and permutations of shared libraries
Uusually well managed by your distribution. A cross-distro way to create a standalone installer would be nice though, LSB doesn't quite cut it.
just as many sources of outdated, incorrect, misleading or utterly superb documentation
Most projects have a homepage. That is the source of the most up-to-date information. Though most of the time, the docs in the package is enough.
crap vendors like Redhat which drop version support in a third the time of Microsoft
And Debian gets scolded each time they're mentioned for actually supporting something for a while.
Windows was never an OS. It contains an OS, they changed OSes in the product lifetime, but the product has always been a desktop environment and a consistent, well documented, and long-supported API.
Linux does that. But you should really mention a long-supported ABI. Linux does definately not have that.
In short, I see all of this as signs that Linux is moving too fast for people to consolidate and work out standards. Being more bazaar than cathedral, that is natural. But that is like a brake on a streamroller already in motion.
We're in a transition period where people are held back by old systems, but seek cross-platform compatibility on new systems. It's like watching pressure build for a switchover. Just because there's been no mass exodus you still see them untangle themselves from Windows strangleholds.
Parent
Re:I think he's right (Score:5, Insightful)
Many of us call that CHOICE .
I can pick the Linux distribution that best fits my needs, be they toolkit-driven, tool-driven, UI driven or otherwise.
With Windows, you get... well, Windows. You have to shim other things onto it to get it to be useful. For example, I don't use icons, toolbars, window frames or titlebars. Show me how I can configure Windows to provide that interface, in an easy way... you can't. Not without 10 different third-party products.
Its all about choice.
Parent
It's sort of the other way around (Score:5, Insightful)
I understand the point you're trying to make, but the analogy was the wrong choice.
Parent
Re:I think he's right (Score:5, Insightful)
Parent
Re:I think he's right (Score:5, Insightful)
Parent
Re:I think he's right (Score:5, Interesting)
Flash forward to now: I have worked my way through the following distros, by doing a full wipe-and-reinstall each time:
Mandrake 10 (as mentioned);
Mandrake 10.1;
Gentoo(lol)
Each time, as soon as the nVidia binary driver was installed, UT04 would start and run without a single tweak being made to the UT04 install.
The lesson to learn is this: although the majority of open-source Linux software is not self-contained (and this is by conscious design) and has dependencies that need to be tracked-down and installed first, there is no reason at all why a company can't just package up everything it needs in one big self-contained lump, eliminating the need for dependencies or the need to run on a specific distro entirely. As for the comment that you need to recompile for different hardware: I have no idea what you're talking about. Clearly, if you have a x86 app, it will need to be re-compiled to run at full speed on a PPC system - a difficulty not encountered in the Windows world for the sole reason that Windows is only capable of running on x86, and similary for MacOSX.
I suspect I've just been feeding a troll, but oh well - who cares? :)
Parent
Re:I think he's right (Score:5, Insightful)
General purpose programs are different. Look at the standard libraries on OS X or Windows. You have a complete windowing toolkit or two (Win32 / Avalon, Carbon / Cocoa), a media plaing framework (DirectShow, QuickTime), an HTML rendering engine (MSHTML, WebKit) and a whole host of other things which a guaranteed to be there. You can build your app expecting them to be there.
On Linux (or *BSD for that matter), alternatives to most of these things exist. In some cases, several alternatives exist. The problem is that you can't guarantee that they will be there. You can statically link everything, but then you have to update your entire app whenever small updates to dependant libraries are released. Alternatively you can just release the app dynamically linked, and hope that people have all of the required libraries (where you expect to find them), and hope that the distribution will package your app in such a way that it will work. The only way to really make sure it will work it to package it complete with dependencies for every distribution you plan on supporting, which generally limits things to Red Hat and maybe SuSE, even though the code would work with no modifications on a large number of other platforms.
Parent
Re:I think he's right (Score:5, Interesting)
DLL hell is not unique to either Linux or Windows. It has been a serious problem from the moment the first person updated a shared library on the first system to support them.
The logical symbols mechanism in VMS allowed many problems to be avoided, instead of loading a filename the program would load a logical symbol. It was easy to run a system with different versions of the same shared library in use simultaneously. UNIX and Windows never used symbols in quite the same way.
I think that folk need to look at what they get from a shared library. At one time it made great sense to save memory by having multiple processes share an in memor image of an executable. It makes particular sense if you are running something like Apache where child processes are being spawned off from the parent and share resources with it. I don't think it makes a lot of sense as a general approach when memory cost $50 per gigabyte.
We could probably do much better than shared memory if we went back to static libraries and instead used more intelligent linker technology. When I link to stdio I pull in maybe 500K of code and use at most 40% of the code paths. When I link to more recent libraries the library is much larger and the fraction I use much, much less. A shared library is an all or nothing affair, every part of the library has to be loaded in case another image might need it. Even if the code page is never touched the memory has to be allocated.
As for the question of user interfaces, I think that the way they are designed today is worse than sub-optimal. I would prefer to go back to an architecture similar to the one that the NextStation had. Instead of having the program implement the user interface as code it should send a description of the user interface to the windows manager and have it perform all the necessary animation.
This approach is similar to what we did in the early HTML days but the idea is to take the approach much further. I really dislike the fact that most programs are single threaded and the UI goes to sleep every time it is asked to do anything computationally intensive of requiring the network. The architecture I just described allows the window manager to keep the user interface alive even though the program logic is 'thinking' and the programmer does not have to do any work to achieve this.
The other advantage of this approach is a bit more controvertial, it limits the scope of the UI designer. This is a bad thing if you really, really love to foist a bizaro UI onto the user. On the other hand it means that every application can be skinned so implementing the bizarro UI is simply a matter of telling the program manager how to do it for every program on the machine.
Parent
Re:Wow (Score:5, Insightful)
A set of APIs and an ABI for writing graphical programs which is still supported now in spite of being over 10 years old and can be guaranteed to be available on 100% of Windows systems?
Parent