OSNews

OSX on any x86 hardware, and i may be game. but if i can only get it on apple approved hardware, forget it.

Me too!

Apple, do you read these comments? there are all over the web!!!

You said it. I wouldn't want to see Linux, BSD, Solaris or whatever taking 95% of the marketshare.

You can argue that these OS are free software, but you get the point.

Having at least 3 majors OS to force interoperability would result in free as in speech standards.

The goal is to make Windows easy enough that no one should need support. We aren't there yet, but if we just charged for support then anyone else could come along and support the software too. And any improvements in usability would have a negative impact on the bottom line.

The Windows software has value, and the value has increased over time as features are added. The price has also decreased over time due to inflation (the numerical value of the price has not changed much since the beginning).

Its true that Windows does have good value.

However this modular way of thinking is not new for Microsoft. In Windows 98 Second Edition you could chose components to be installed. And making things optional again should not come at the extra price and hassle of time based subscriptions.

That is why a more frequent release schedule for example Microsoft office seems better to me.

I'm not bashing Windows (With the exception of Windows ME). I just think that to further fragment Windows will be painful for us all.

Microsoft cannot release Windows more often; if it were to do so, they would have one hell of a time getting businesses to upgrade.

Consumers would also have trouble with this. The products would be much less polished (Vista was likely rushed).

As for price, they could use a reduction there, but I do not know what drives its price.

What is wrong with Unix(-like) again ?

I'm not saying unix-alike is bad, but I do feel it will be unfortunate if the only "man" standing at the end of the OS wars is unix-alike. I'd like to see various other OS implementations appear. I don't want to see the next 20 years be all *nix. A nice varied ecosystem is a good thing, as we've all seen the prevalence of windows systems leading to a hellstorm of viruses/malware. I'm interested in seeing what other vendors/groups can come up with that's neither linux/mac/windows. No system is perfect yet. I'd like to see what people come up with given a chance to explore other models.

Edited 2008-04-06 14:40 UTC

I would hazard a guess that if Windows were to decline by 50% over the next several years, the largest benefit would be to Apple simply because most people still think you have to pay for an OS, but Linux would also benefit enough that the split might be something like 46-36-13-5 (Windows-Mac-Linux-Other).

The mere fact that Apple is gaining mindshare at Windows' expense helps people to realize that a computer doesn't have to run Windows. And, the fact that Windows can easily be installed on current Apple hardware, makes it much easier for many people to understand that the major difference between Apple and Windows is a difference in software rather than hardware (Windows boxes can be designed well too).

This opens the door for people to ask if there are any other options. The obvious alternative to Windows and Mac is Linux, but many would argue that the BSDs and Solaris are very good alternatives as well. Of course, all of those fall under the category of being Unix or Unix-like, but I say that it would also make people more open to trying other systems.

If that were the result of Microsoft trying to modularize Windows for profit, I would be all for it. It would pave the way not only for an increase in *nix usage, but it would make the development of more alternatives increasingly likely. Competition would truly be king.

the problem is that apple is unlikely to license out their os to competitors.

its that one could shop around for the best offer in hardware, and still use ones software that have made the computer a home item.

still, there are some benefits for the home user in having a os locked to the hardware. and i have often found myself saying that the average user (if such exists) do not want a general computer, he wants a typewriter, media player/recorder and web portal device. at least thats the use i have seen over and over again, with the odd game thrown in.

so in many ways, said user would be better of with a combo of boxes hooked up to a single input and output. each box providing one of the features listed.

Well look at it this way. Look at when the Unix architecture came out and look at its prevalence today. That should tell you something right there. look at the scope in usage of *nix platform, look at the fact that MS themselves have licensed parts of the *nix platform to use in their own OS, look at the fact that MS is trying to do the same thing *nix has done for over 2 decades with windows 7. Modularity isn't a new concept, *nix has had this from the start. I'm not saying *nix is perfect or that every OS should be just like it but the proof is in the pudding and the *nix paradigm has been proven over and over again for more than 30 years. It has taken MS decades to come to the same technical conclusions that *nix made in the 70's.

There's so much more to Linux and especially Mac OS X that's not really Unix of yore that I think you are doing a disservice to these OSes by calling them holdovers from the 70s. Yes, the Mac OS X kernel is Unix-based, but even that is a lot different from the Unix you'll remember from the 70s. Almost the entire userland is not Unix, it is very modern and very different from the traditional Unix userland. That's where it counts these days anyways. The Linux userland is a more unix-y, but even so, it is still quite a bit more modern than you seem to be making it out to be.

I think at this point, saying that "all that will be left standing is Unix" is akin to saying "all that will be left standing is virtual memory based kernels". Yes, all mainstream kernels use virtual memory with paging. So? That's the solution that the the industry has adopted to solve the problem. Same with other Unix-y stuff. Even Windows got a lot more Unix-y with NT and later. A lot of things we take for granted as "how a system should work" were developed first with Multics and Unix.

Forgive me saying it out loud, but some of you people just need to get a cluebat and start hitting the wall at least. For far too many years has the consumer OS land been dominated by a single OS, and hey, a real minor number of people kept repeating that a varied OS landscape would be better. Now, that suddenly some people have woken up, and we start to see some light at the end of the tunnel - and we're still pretty far off to having a varied OS ecosystem (availability in itself means nothing) -, people pop up and start complaining that reducing the lead of Windows as an OS (reducing, not eliminating) will bring us doom, and we should have something, anything just not unix/linux-based OSes. It's just funny.

As regarding the subscription-based plans for Windows... my feeling about that are just skeptical. You see, even MS acknowledges that a very high percentage of the OS-income comes not from individual OS-buyers, but from OEMs and companies. For home users this subscription-based Windows will probably be a major pain in their proverbial behinds, and even more so if we consider that bugs and glitches in the respective services will take your patience by its neck, twist it around, spit on it and bust its face with a twenty inch steel-toe. For companies, well, I don't think a sane cio/cto will go into such a mess. I might be wrong, we'll see (unfortunately).

The problem with your argument is that you rely up CIO/CTO's to be sane.

That's funny stuff, right there.

I've worked for CIOs and CTOs of huge industries. I've worked with CIOs in several state governments.

I have yet to work with one who makes decisions based upon the precept that they're role is to solve problems with the least-complex, most effective, and cheapest technical solution. Using these measures as the means to selection would not only be sane, but would be rational, responsible, and likely lead to highly successful projects.

Instead, they read Dr. Dobbs, bow at the alter of Gartner, or do any combination of hair-brained, moronic, absolutely stupid, backwards, and self-defeating methods of choosing and selecting "solutions".

Honestly, these people are rarely sane, intelligent, knowledgeable when it comes to technology, and normally promoted out of the way at company X, where they use that high-ranking position to attain an equal position at Company Y, who realizes their inept, and promotes them out of the way, until they move on to Company Z.... and so goes the cycle.

I wish I were kidding, but I've seen it way too many times.

Haiku OS?

Amiga OS4.1 sounds like the answer to yer prayers!

While of course I disagree with your analysis of Unix(-like) operating systems (I have seen them improve, become a lot more user-friendly since the year 2000, getting (more) proprietary software, dramatically increase what you can do with them, with the possible exception of games...), I have no interest in seeing Windows disappear. But can't they conceive something completely different? I mean, for instance, an OS without the Registry and the DLL Hell.
Which takes us back to Unix. Maybe they shouldn't create yet another Unix based OS, but taking a few ideas from Unix wouldn't be such a bad thing.

In an ideal world we would be using the superior (technologically) operating system; Plan9 would become the new UNIX, people would clone/duplicate it and make it better. The sad reality is that pragmatism, not technology, decide the direction (along with success) of an operating system.

With that being said, UNIX isn't the greatest thing in the world, but when compared with the mess that is Windows NT, we could be doing alot worse.

If pragmatism decided the direction, then management would have to be thinking pragmatically.

That doesn't happen in > 90% of the businesses out there.

I call your bluff.

Most of the time, what happens is you find a manager with a hard-on for company / technology X. They won't accept anything other than company / technology X. Why did they get such a boner for this decision? It could be pragmatism, but that would imply rational, conservative problem-solving being applied to a problem set.

Normally they read it under a headline written by some industry analyst blow-hard, or got a few nice steak dinners on the vendors dime.

What the f*ck are you going on about "I call your bluff" - I made no such thing! idiot. I also said NOTHING about acquisition of information technology for companies - may I suggest that you're not so cocky next time.

The point I was making is that engineers within companies like Microsoft, Sun, IBM and so forth have to make pragmatic decisions; they have limited budgets, limited time, and ever demanding customers wanting more for less. It is up to the engineers to be pragmatic on what they can do given the constraints which are placed on them.

Again, read the posts before you start running off to the create reply button.

If modularization is fragmenting enough it might be just as cheap for businesses to retrain their workforce on something else.

Vista has shown that businesses aren't willing to swallow anything that Microsoft labels as "Windows" hook, line and sinker.

i suspect that both apple and microsoft is heading towards google, while google is heading towards the other two.

as in, both are heading towards a service and portal like place (.mac, live) and creating specialized boxes and solution that can tie into that to provide additional services with little effort from the user (zune, ipod, appletv, xbox360).

google on the other hand is a service and portal, and is now producing the first hardware related products (android). it will not surprise me if android can be scaled up to do PVR, online video rental or game console.

and now that one have the net, the formats and similar thats being used there acts to level the playing field. unlike earlier when each computer brand had a silo like market share.

these days one have to be net compatible or one is dead in a market sense. and so it becomes very simple for the user to move his data between vertical platforms.

the apps may be different, but as long as they can read the same data, will the user care?

That's technically true, but as I said in my previous post, I don't think they will ever really go away. There's no reason for Microsoft to completely disappear from the IT scene even if they do move away from operating system development. For one thing, there's a hell of a lot of talent there, albeit with less than stellar management. They have the resources to make a graceful shift away from OS work and towards service and support software. Why can't Office, or its future incarnation, be the flagship of the company ten years from now? Imagine your computer running a free OS (Linux, Haiku or whatever is out there then) and on top of the OS is Microsoft's productivity suite. I know, not everyone wants to see something like that, myself included, but it's one of the many possibilities when we go down the road the article suggests.

As I said before I don't see it playing out like this; I have a feeling Microsoft will be the gorilla in the room for many years to come. Eventually though, they will have to innovate on their own and stop borrowing ten year old ideas. That will be something to see.

vista anyone?

You read my mind perfectly. As far as I'm concerned, I can do the majority of my "work" on Linux. To me, Windows has become nothing more than a glorified gaming platform. I'd gladly take a version of Windows for gaming - why bother with all that stupid eye candy when you aren't going to be seeing it for more time than it takes for Spore or Crysis to load?

I'm not sure you're right that IIS6 required kernel changes to be upgraded in server 2k3. It's just a program, like apache or any other web-server. Perhaps there were some kernel enhancements that made it run a little faster, (I can think of one that might have affected IIS performance), but that does not imply that any part of IIS is attached to the core kernel.

I can't find the source right now, but I know I was reading about the fact that IIS6 being monolithical couldn't have been seriously upgraded without updating the files it has dependencies in. IIS is not just a program, http.sys is actually a kernel-mode driver, so it had dependencies in kernel code.

I think I was reading on IIS7 being modular and that was to be a benefit over IIS6 where it could be upgraded, like you say, like just a program.

That's a little different. The IIS7 modularity has to do with the fact that it processes web requests in stages in a pipeline. Each part of the pipeline is implemented as a module, and so can be filtered or modified. The older versions had larger chunks of processing being done in the engine itself, so you couldn't, for instance, change the way that SSL processing was being done without modifying the core engine (not a real example.. just for illustrative purposes). The new IIS would handle connection establishment as part of a module that could be switched out without changing the rest of the serving pipeline. But all of this added modularity is in IIS itself, and has nothing to do with its interaction with the OS. It's just a program like any other.

HTTP.SYS is just a normal driver that helps IIS. It can be replaced or modified independently of the OS Kernel. I think the interface to http.sys is also public (search MSDN for HTTP API), so even apache or lighttpd could use it to serve requests. It's pretty unrelated to the Core OS.

In a perfect world, I would agree that this would be the way to go. A la carte Windows could work, but since MS is more about making money than caring about product and customer satisfaction (the result of owning 95% of a market) a modular subscription based Windows would be a disaster in terms of end-user experience and implementation.

IMO, Windows is modularized enough as it stands now.

This would be a nice utopy for "windows" users, but sadly, this concept seems to open up new problems. Things that users might expect (e. g. have seen it in the office or at the neighbor's PC) arent present in fact and require manual installation afterwards. "But I want this working!" could be a common phrase. The additional installation of software that is needed to extend the product could develop into a problem for users who are fine with the concept of taking software as it is (first) and then start complaining that something isn't possible out of the box (later on). This is due to the tendency of seeing the PC as an allround device, so, following your example, as an office PC for games, file sharing and video editing.

The selection of what is minimal and what you can use as extensions is a difficult task because someone could miss something - missing something is the beginning of complaining, and complaining is a step closer to abandon a product and stopping using it.



And for the nice numbers on the price tag. :-)



This would require their modularized infrastructure to be well designed. Some users blame MICROS~1 for putting more work into the eye candy and the advertising than in testing their software before selling it. I'm not sure if MICROS~1 can reach this goal, it would imply that they break with some of the concepts they're familiar with for years...

Dude, it's more work to type MICROS~1 than to type Microsoft or MS. Why would you do more work just to be impolite?

Dude,^WIt's more work to start your sentence with "Dude" than to just leave it out. Why would you do more work just to be impolite? ;-)

Edited 2008-04-06 19:11 UTC

The dude abides.