Firstly, you're assuming the average Google user realises that whatever they search for will be saved, and can be personally linked to them for 18-24 months, and perhaps longer depending on how effective the 'anonymising' is. This is not a reasonable assumption.
Secondly, in the EU at any rate, we have this idea that there are certain rights that cannot be given up. One cannot give up basic rights to freedom by signing a contract agreeing to become someone else's slave, for example. This is the same sort of issue, just on a more modest level. Agreeing to use Google's search engine cannot take away the right to privacy.
RMS didn't split the OSS community, the "open-sourcers" split the Free Software community.
Open source software, including examples such as BSD, was around before Free Software, but without any ideological opposition to its use in proprietary software. Whether or not Stallman's ideological stance is more productive than the one taken by open source developers before him, and those who didn't join the Free Software movement, is an open question, and depends on the goals.
If the goal is the best software, then the GPL is largely irrelevant. If the goal is the best open source (or Free) software, then the need for something like the GPL ultimately depends on whether or not the open source development model is more effective than the closed model. If it is, there isn't much of a need for preventing proprietary additions, because any temporary advantage from closed development will be negated over time, by parallel open development. If closed source is actually a more efficient development model, then something like the GPL becomes much more important. I don't know which is the case.
I'm not ideological, so I'm glad to be able to benefit from the incorporation of BSD sockets into proprietary software. If Windows 3.x/9x, SunOS, Mac OS and others hadn't been able to take advantage of BSD sockets code, it almost certainly would have prevented BSD sockets becoming a de facto standard. I don't know which networking API most of us would be using, but my guess would be something designed by Microsoft. Thanks to BSD sockets, all of these systems can interoperate well, and open source systems like Linux are actually useful as network servers, serving mostly Windows and Mac OS clients.
At least as important, I think, is if Dell actually work to ensure there are Linux drivers for the hardware they're selling. Linux hardware support on laptops is often so bad that it's effectively unusable. If Dell restrict their Linux offerings to a few specific hardware configurations that already have good hardware support, I don't think it will make much of a difference to the market.
Their thinking is probably that anyone who wants Linux will look for it, and anyone who doesn't know doesn't want it, but might not be able to understand the difference from a brief description. If users expecting Windows started buying PCs with Ubuntu because it's cheaper, it could lead to all sorts of negative customer service and publicity issues for Dell.
Ubuntu can run well on cheaper hardware than Vista (mainly RAM and video, if you want Aero). So comparing the same hardware means one OS will run better than the other.
I don't think Vista Home Basic includes Aero, which is why it has lower hardware requirements, including half the RAM and lower video card requirements, than the more expensive versions of Vista.
3.0 was unusable. When 3.1 came out I saw people playing solitaire in lots of places, and this was before 1990. Course, this was in California, ymmv someplace else.
Windows 3.0 was released in 1990, and Windows 3.1 was released in 1992, so it's actually impossible for you to have seen anyone using Windows 3.1 in the 1980s. Mind you, that's beside the point anyway, since the OS line Microsoft marketed as Windows 3.0 and Windows 3.1 is now extinct. The last release was Windows Me, which shipped seven years ago.
The first version of the OS that's today known as Windows was Windows NT 3.1, released in 1993. Since it was actually a completely new OS, it was really a '1.0' relese, but Microsoft used the same 'look and feel' as Windows 3.1, and offered a similar API (although it was 32-bit rather than 16-bit), so cleverly marketed it with the same 'Windows' brand and '3.1' version number. A subset of the Win32 API was later ported from Windows NT to DOS/Windows, resulting in Windows 95.
"On the whole, from a code perspective, Windows and Linux are roughly similar in age"
Not from what I saw. There were California computer companies that went under because they didn't do windows in the 80s and linux took almost a decade to enter the mainstream. Just check usenet postings to see what people were doing with them back then. That's "google groups" for you young 'uns.
I'm sorry, but that's simply impossible. As I pointed out above, the '1.0' release of the operating system currently known as Windows was Windows NT 3.1, which shipped in September of 1993. The '1.0' release of Linux came six months later, March of 1994.
Without commenting on the relevance of your points, your facts aren't quite right. Microsoft as a brand has been around since 1975, not 1981. The original Windows system was announced in 1983, and released in 1985, so that brand can perhaps be said to be 22-24 years old, although it didn't gain a wide following until the early 1990s, with Windows 3.x. In terms of the code, the Windows NT development group formed in 1988, when Dave Cutler and his team left DEC and came to Microsoft. The first release was in 1993, so the modern Windows OS itself (not the brand) is either 14 or 19 years old, depending on how you define the starting point.
On the Linux side, the GNU Project to create an open source clone of the Unix operating system was announced by Richard Stallman in 1983, coincidentally the same year Microsoft announced Windows. Some of the GNU codebase dates from 1984, but it's more difficult to indentify the equivalent of a 'release date' for open source software than for closed. Linus Torvalds started his computer science studies in 1988, and announced Linux in 1991, but the '1.0' release didn't come until 1994. Moreover, Linux made extensive use of existing GNU software.
On the whole, from a code perspective, Windows and Linux are roughly similar in age. From a brand perspective, Windows has an age advantage over Linux, but not GNU. Linux could have negated this by using the existing the 'GNU' brand (as Richard Stallman insists it ought to do), in the same way that NT used the existing 'Windows' brand, and thus have had effective parity with Windows, in terms of the brand's age. The fact that this was not done largely reflects Stallman's extremely poor choice of a name ('GNU'), which led to virtually no brand value being developed from 1983 to 1991, when the Linux brand came along. It was actually better to start over than continue to be saddled with such a poor brand name.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 1
This can't be, at least not (necessarily) in the case of software. A large scale producer may enjoy certain economies of scale and their marketing will be more pervasive. However, producing and distributing free software is more efficient than the proprietary way. This is most evident from the consumers' point of view. On the development side, the ability of users to contribute more in the process can have a big impact on testing and developing software. All this, even though it has a lower market share.
On the contrary, this is precisly why network effects are important. If the utility of the software to users is a function of both features and market share, a high-market-share producer can provide the same set of features as a low-market-share producer, but users will get more utility from the high-market-share good. Once market share has got past a certain point, you could conceivably have a sitaution where the added utility from the higher market share is greater than the total cost of development/production, meaning the low-market-share competitor, even with zero costs, would actually need a negative price to be competitive. (There's also the fact that most of the key Linux developers, for example, are paid rather than volunteers, but that's rather peripheral to the point.)
Another important issue is that Microsoft does more than just write the software. A lot of computer standards have been developed by dominant firms, e.g. the ACPI power management standard was developed by Microsoft and Intel, together with Toshiba, I believe. This is one of the reasons other PC operating systems like Linux are so far behind in power management support: they're playing catch-up. As long as Microsoft and Intel are defining the standards for the PC platform, they'll have an advantage in being able to provide products that take advantage of those standards first, giving them at least a window of time in which they can offer better products.
The market does tend to monolopy though. What I've been nitpicking you on is that in my view, if MS had 70% rather than 95% the market would demand interoperability, the current state of affairs wouldn't fly.
I suppose we'll just have to agree to disagree on this point. From my perspective, if Windows had 70% market share, with the other 30% divided amongst others, e.g. 10% for Mac OS X, 10% for Linux and 10% for all the rest, I'm not at all convinced this would have any impact on Microsoft's dominant position: is it worth duplicating OS-related cost structures to go from reaching 70% of customers to reaching 80%? Would the burden of interoperability fall on the users of the platform with 70% market share, or the ones that each have 10% or less?
If the market were a duopoly, with 70% for Windows and 30% for Linux (or perhaps it would have to be a specific distribution of Linux), that might be different, but with even a 70/20/10 split for Windows/Linux/others, it's still far from certain that the dominant-firm dynamics would change. With a lot of competitors, a firm can even hold a dominant position with a minority market share, as Microsoft arguably did in the late 1980s.
Your calculations are correct for a producer in a one shot game, one who wasn't concerned about the impact of pricing in this round has on the next one.
No, not at all. My assumption is rather that a market share of 70% for Windows, instead of a 90-95%, would not change the dominant-firm dynamics of the industry, including Microsoft's ability to set prices for Windows and benefit from network effects. If the dynamics did change, then of course the situation would be quite different.
The real monopoly is in the windows installed base, once that starts dropping the monopolists' game is up. They can still lock people in, but the tide will start to move against them (more than it currently is).
Well, the high market share of Windows leads to pressure in the opp
Interesting, I didn't know about that. All I meant is it doesn't use the old BSD distinction of ports below 1024 being reserved for privileged users, with 1024 and above being open to unprivileged users. I suppose you could effectively set it up that way using port filtering.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 1
The situation is far more complex than either earning monopoly rents or earning a normal economic profit. As a dominant firm with a large but clearly non-monopoly OS market share, Microsoft could still set prices to a large degree, and benefit from scale and network effects, allowing it to earn a very large supernormal profit without being considered a monopoly (which it arguably isn't in the pure sense now anyway).
As I've pointed out before, Microsoft's market share in the PC operating systems market was less than 50% in 1990, and only about 58% in 1992, yet practices such as tying and various forms of price discrimination were used very successfully: that's one of the reasons Microsoft's market share in the same market was above 90% by the end of the decade. Ironically, Microsoft probably ought to have stopped competing aggressively once it reached 70-80% market share, leaving an 'AMD', as it were.
If Microsoft had 70% paid market share, for example, it would still have a substantial advantage over its competitors, particularly if several were sharing the other 30%. If you start defining a 'monopoly' from the regulators' perspective as including firms with 70% market share, you're on a very slippery slope: Intel, Apple (iPod) and Cisco already have higher market shares than that, and Google could very well reach that level within the next few years, as could Windows Server (which is not currently subject to the regulations applied to the Windows desktop OS). Are they all monopolies?
The key, really, is that if the market naturally tends towards monopoly, then a producer with a higher market share can produce more efficiently than producers with lower market shares. If there are a number of such related markets, a firm which is dominant in several of them, earning supernormal profits, could very easily earn a larger overall economic profit that a firm with a monopoly (in the non-technical sense of a very high market share, e.g. above 90%) in one of the markets.
Dominant firms with very high market shares, whether Apple, Google, Intel, Cisco, Microsoft or others, must take care with their actions, but regulation of dominant firms is far more controversial than regulation of monopolies. If Microsoft Windows had held 70% of the desktop OS in the late 1990s, rather than more than 90%, it might very well have escaped classification as a monopoly. Microsoft might still have faced some penalties for anti-competitive practices, but being ruled a 'monopoly' had far-reaching implications.
On the whole, assuming the global 35% piracy rate holds for Microsoft Windows, and a 95% market share, this means 61.75% of users in the desktop OS market are paying Windows customers. Even if Microsoft were to capture only 10% of that 35% through a successful anti-piracy scheme, with the other 25% going to other operating systems, the percentage of paying Windows customers would rise to 71.25%. It's exceedingly unlikely that a ca. 70% market share for Microsoft Windows would be below the tipping point, hence it would have virtually no impact on Microsoft's ability to set prices. The increase in the number of paying customers would thus increase Microsoft's revenue from Windows by ca. 15%, and have virtually no impact on costs. On top of that, it would potentially reduce regulatory pressure, allowing more scope for Microsoft to leverage its dominance when competing in related markets.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 1
If restrictions on pricing were the major cost of regulation, I might agree, but I'd argue that the most important cost by far is the imposition of restrictions on activity, such as prohibiting tying (in fact, I don't think regulators actually interfere with Microsoft's pricing). The reason Microsoft is forbidden from tying other products to Windows, or using certain discriminatory pricing, is because desktop Windows is considered a monopoly OS by regulators. If a substantial percentage of the people pirating Windows were using other operating systems instead, Windows might have a low enough market share that it would be untenable to treat it as a monopoly OS. Microsoft would then be allowed to tie other products to Windows, add client functionality that only works with Microsoft servers, offer better prices to vendors who agree to push other Microsoft products (e.g. Windows Server), etc. In short, Microsoft could leverage its high market power to gain competitive advantages in related markets.
I'm not sure about the 'techie' group either, but you have a good point there. It's not my field, but it seems to me that a lot of computer science education is already Linux-centric, because the source code is available to be studied, and it has the best hardware support of any open source OS, so students can easily modify it and test their modifications on common hardware. On the other hand, almost every other field is extremely Windows-centric, so once the technical students get out into the real world, they're forced to use Windows, because it's what the buyers of their services demand. I suppose the critical issue is how many semi-technical students there are, meaning those who build their own PCs, but wouldn't be comfortable using Linux, so would be willing to pay something for Windows, assuming they couldn't pirate it. That doesn't necessarily mean the current retail price, which is set with the assumption of high levels of piracy.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 1
Monopolies don't require a market with network effects
No, I hope I didn't imply otherwise. If I did, it was unintentional.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 1
How many ppl would be irate if they had to chose an OS b4 driving away a new toyota?
If you could get a better car for half the price, a lot of people probably wouldn't mind it so much. That's more or less what happened with PC hardware, especially after Windows took off in the 1990s, which provided standard interfaces to devices, based on common drivers. Before that, developers had often been forced to write to specific devices, e.g. this or that graphics card, sound card, printer, etc., so the change was huge. That's one of the major reasons all of the non-PC platforms of the 80s fell away (except for the Mac, which gradually evolved into being more or less a PC with a more restricted set of hardware and a different OS), and Microsoft shot from less than half of the desktop OS market in 1990 to a near monopoly by the end of the decade.
Re:Piracy is marker of immature market
on
Piracy Economics
·
· Score: 4, Insightful
A network economy is different to an addictive drug. The reason Microsoft might want to allow piracy is because the more people use Windows, or Office, etc., the higher the utility to each user. In contrast, an addictive drug is addictive even if you're the only one taking it.
For a company like Microsoft, there are at least three or four different phases, and the implications of piracy are different in each.
1. Minor producer: if you're a minor producer with low market share, piracy may be good for gaining market share, as long as revenue from paying customers remains high enough to cover costs.
2. Dominant producer: if you're the dominant producer in your market, but perhaps still with only a minority share of the market, piracy is good, because most people pirating will be pirating the dominant product, This will spur a network effect, and any revenue implications are likely to be less important than for smaller producers.
3. (Near) monopoly, without regulation: if you've got a near monopoly, you'll gain the benefits of network effects. The network gains from piracy, and the extent to which it keeps out competitors, are both gains. Without viable alternatives, however, there is the potential for higher revenue from those who are pirating, but would pay if they had to. The network effect and the exclusion of new entrants might be worth more than the lost revenue.
4. (Near) monopoly, with regulation: if regulatory restrictions are imposed on a firm with a near monopoly, that means the gains of network effects and the prevention of new entrants are offset by both the lost revenue and the costs of the restrictions (e.g. no bundling, limitations on pricing strategies, etc.). I this case, the more onerous the restrictions, the less value there is from piracy. It may be worthwhile to give up unpaid market share, in exchange for higher revenue, especially if this leads to a reduction in regulation.
In the late 1980s and early 1990s, Microsoft was in category 2, with a dominant position, but a market share near the middle: it crossed the 50% mark in 1990. During this time, piracy was arguably good for Microsoft. By the middle of the 1990s, however, Microsoft had moved to category 3, and so whilst piracy was no longer as clear a benefit, it was still arguably less bad than good.
With the monopoly ruling against Microsoft in 2000, it moved into 4, although the level of regulation has varied. With the regulatory costs offsetting some of the network gains, piracy arguably became less valuable to Microsoft, and this may in part explain the increase in anti-piracy measures in Microsoft's software since then. Giving up some non-paying customers to competitors, in exchange for converting some non-paying to paying customers, is arguably a good strategy, especially if it reduces the regulatory pressure.
An interesting point is whether people who pirate Windows, and would switch to Linux or something else if they couldn't pirate it, are willing to pay for other software. The expected answer is no, so Microsoft could arguably give up these low-value customers without losing the benefits of being the dominant platform for commercial software development. Producers of commercial software would have little interest in developing for Linux if Linux users wouldn't buy their software anyway, so a higher market share for Linux would have little impact on the network effect there.
From the above, the risk of giving up some market share comes from network effects other than those relating to commercial application development. For example, people who won't pay for software may still pay for products and services bought over the web, etc., in which case they'd be targeted by website developers. They might also still be willing to buy relatively expensive hardware, which could reduce the network effect regarding device driver development.
The Network Service account on Windows has similar privileges to a normal user, which means it can't access files owned by other users, but can of course read some files owned by the system. The notion of reserved ports doesn't exist on Windows, so no software makes security assumptions based on whether or not a port is below 1024, and the ability to open port 80 doesn't imply any higher privileges than the ability to open any other port.
At any rate, running in a chroot jail is arguably better in some ways than just running as an unprivileged user. Vista has some sandboxing features, using 'integrity levels' and redirecting various file and registry accesses to a 'virtual store', but I'm not really familiar with them, except for the basics, and I don't know if IIS uses them anyway.
Windows doesn't really have an OS/2 heart. The whole point of 'new technology' OS/2 (which became Windows NT) was to create a new 'heart', i.e. a completely new and portable kernel, for the OS/2 API, with the POSIX and MS-DOS APIs included for good measure. This is because the 'old technology' OS/2 was tied to either the 286 (OS/2 1.x) or 386 (OS/2 2.x), and most people in the late 80s and early 90s thought RISC was going to kill off the x86.
If you include NT OS/2, which was started in 1988, in Bill Gates's 1987 prediction about OS/2, then I suppose you can say it was, in a way, right.
Well, Microsoft had less than half the OS market as late as 1990, and in 1987, its market share would have been even smaller than that, so nothing close to a monopoly. On the other hand, NT certainly wouldn't have become as dominant as it is today if it hadn't been compatible with the Windows 3.x line, which is what drove Microsoft from being just the largest desktop OS vendor to being virtually a monopoly (especially after Windows 95). That's of course why Microsoft wanted to add a Windows-like API to NT OS/2, which in turn led to the dissolution of the IBM/Microsoft partnership.
The problem is, Gates made most (if not all) of these comments in order to push efforts that Microsoft was working on at the time. As a CEO of a major software company, part of his job was to make comments in public that would try to influence the industry to move in the direction that would align with what his own company was doing (or at least attempting) already.
Another way of looking at it is that perhaps Microsoft started those efforts because Bill Gates thought, or was convinced by colleagues at Microsoft, that they were going to take off. If Gates thought everyone would rather use a pen than a keyboard to interact with their computer, then it naturally follows that he wanted to invest resources in the idea. Remember, the reason Gates quit Harvard Business School and started developing software is because he correctly thought selling software was going to be a huge busineess in the future. Some people see that he got that one right, so think maybe he'll be right on other technology issues.
The reason I don't give a huge amount of weight to Bill Gates's technology predictions is because the context is completely different. In 1975, the personal computer industry was in its infancy: hardly anyone even knew about it, much less about selling software for microcomputers, if they even knew what software was at all. The combination of familiarity with software, knowledge of microcomputers and a grounding in business, which Gates had, was relatively rare. It wasn't altogether surprising that Bill Gates had a better ability to predict the rise of the software industry in the mid-1970s than most other people at the time: he had a relatively rare combination of information. Today, there is a massive number of people analysing the IT industry, and Microsoft itself, so the idea that Bill Gates is going to have a similarly rare set of information today isn't very plausible.
You could replace 'Bill Gates' with 'Tim Berners-Lee', who was right about the web (and even invented it), and come to much the same conclusion. The fact that Berners-Lee had a rare set of knowledge that allowed him to predict and invent the web doesn't necessarily mean he'll be better at predicting anything else.
You see you live in a different world than I do. You are worried about your system. I help support 15,000 users.
Yes, indeed I do. I expect it will be some time before my university upgrade from XP, since their perspective is, like yours, more utilitarian than mine: the computers are there to let students and staff get work done, and nothing else counts. Indeed, when I'm using university PCs, that's my perspective too. The fact that we're still on Office 2003 instead of 2007 is a nuisance, because it directly impacts the ability to get work done (2007 has a number of useful features, substantial improvements to existing ones, etc.), but the lack of Vista isn't.
For the owner of an individual PC, the situation is completely different, as it is with other personal possessions. If individual purchase decisions were purely utilitarian, we'd all dress like Soviet citizens of the 1950s, and those with cars and no children would be driving Fiat Puntos, or something like that. Why do people buy nice-looking clothes, flash cars and high-priced electronics, from mobiles to TVs to computers? It's all down to tastes and disposable income: non-utilitarian product features actually matter to most consumers, in most purchase decisions.
Honestly for 100% of current Windows end users XP IS GOOD ENOUGH. The only group of people that it isn't good enough for are Developers that are working on Vista software, support techs that must support users on Vista, and people writing books and blogs about Vista. There isn't any software that only runs on Vista so yes XP is good enough.
Again, this utilitarian view applies in an organisational setting, but not for individuals. I imagine a lot of organisations will put off upgrading unless they need specific features in Vista, just like they did with the 2000 to XP upgrade, or the NT 4.0 to 2000 upgrade before it. The key market in the short run is clearly individual consumers, not organisations, but the latter have always followed eventually, once the incremental improvements came to dominate the initial costs of unfamiliarity and more limited hardware support, both of which decline over time.
I constantly hear support techs saying "Oh no she is on Vista".
I'm not surprised. Their job is made much easier if everyone is running exactly the same software, on exactly the same hardware. That's one reason organisations have always been slow to upgrade.
I wonder if these deals are going to work out or have the same stink of AOL-Time Warner.
Well, I don't think you can really compare them. These deals are for cash not shares, are tiny relative to the overall businesses operated by Google and Microsoft, and in both cases the acquiring firms are large and very profitable.
The AOL takeover of Time Warner involved using ludicrously overpriced AOL shares to effectively buy Time Warner. The deal was only possible because of the.com bubble, and when the bubble burst, the major effect was that Time Warner shareholders had handed over most of the company to AOL shareholders, for very little in return. It's a completely different situation.
Yeah, well, *sorta* is the right word. Windows still can't delete a file that's open. I really wish for that behavior for file is removed from directory immediately and from disk when it's closed.
This isn't entirely true. Win32 uses 'sharing flags' when opening a file, to determine how subsequent access attempts will be treated. By default, Win32 denies delete access when opening a file or directory, but if the delete sharing flag is set when the file is opened, that file can still be deleted, even though it's open.
The Unix subsystem for Windows, which is what provides POSIX compliance (or at least did at one point), uses Unix semantics for files, so a file or directory opened by a Unix process can have its name removed from the file system tree immediately, and will be removed from the disk when the last handle is closed. However, for some reason there's an exception to this for executables mapped into a process. These can be renamed, but not deleted. It looks like a bug to me, but maybe there's a reason.
I agree with you, however, that I'd really like to see the Unix semantics offered for Win32, though some optional global setting. I'd like to see the bug with unlinking mapped executables under the Unix subsystem fixed too.
No, not at all. If both standards are adopted, the odds are that users will gravitate to one or the other over time, and that one will become the de facto standard, even if there are multiple de jure standards. Viable standards almost always come from successful implementations, and not the other way round, so the key to becoming the de facto standard is to have the more successful implementation.
If Microsoft believe their standard, OOXML, is better than ODF, or that they have an advantage in spurring its adoption, then there's no need to try and block ODF. All they have to do is make sure OOXML is on a level field with ODF, and wait for it to win. Alternatively, if they think ODF is as good as or better than OOXML, they can add support for both formats, and use their influence to guide the winning standard, whichever it is.
At the end of the day, if customers care more about open formats than rapid addition of features, Microsoft will shift towards open formats. The alternative would be a decrease in competitiveness vis-à-vis competitors using open formats, such as StarOffice/OpenOffice.
The only real risk of open formats to Microsoft's business comes from making it easier to migrate from MS Office to competing products. However, if Microsoft believe they can produce a better office suite than their competitors, this isn't really a threat at all. Mind you, I mean 'better' in the sense of providing value to users, which may include non-technical features, such as similarity to previous versions.
Except that about 90% of the people that us Vista turn off the UAC because it is too big of a pain to deal with.
I'd be surprised if that's the case. UAC popups are typically only encountered when installing new software and changing global system settings. In terms of everyday usage, these are relatively rare.
Hardware drivers.
Needless to say, it supports the hardware it's sold with. For pre-Vista hardware, I'd only recommend upgrading a machine to Vista if it's got the 'Vista Capable' logo, or a similar set of hardware to machines that have. New Windows releases always have more limited support for old hardware than the old versions, but also invariably have vastly better support for newer hardware than any non-Windows OS.
Stability. I have heard that NVidia drivers have issues.
I ran into stability issues with video drivers in the beta version, but the first set of drivers released after Vista was launched have been problem-free. In my case it's about as stable as XP, maybe marginally more so.
Cost. It is expensive.
Yes. This doesn't matter for new PCs, becaues OEM licences tend to be quite cheap, as with XP, but I'd imagine the high cost has stopped a lot of people upgrading to it. I've no plans to upgrade my older PC either.
Resource consumption. It requires more ram, processor, and GPU power than XP. That also means more power is used than XP.
The GPU-hungry features can be turned off, or set to automatically turn off when going from mains to battery power. Most people are quite impressed by them, howerver, so might rather have flash graphics than a little longer battery life. In any case, you still get all of the 'plumbing' and security improvements, some of which improve performance over what XP offers, if you have got the resources.
For what? I used W2K for years after XP came out. I only "upgraded" when I built a new system and used an MSDN copy of XP.
I listed a few of the improvements in my earlier post, and that's certainly not all of them. If you expect every OS release to be revolutionary, you're going to be disappointed. In fact, if you look at the Windows NT line, it's all been evolutionary, from Windows NT 3.1 all the way through Vista. It was only revolutionary for people switching from the Windows 9x to the Windows NT line.
In technical terms, going from Windows XP to Windows Vista is at least as big a step as NT 3.1 to 3.5/3.51, NT 4.0 to NT 5.0 (Windows 2000) or NT 5.0 to NT 5.1 (Windows XP). The only release that's really comparable was NT 4.0, and it's arguable which of the two releases was a bigger evolutionary step.
I still say that for 95% of the people on the planet Vista isn't a good value. XP really is just good enough.
95% is probably a bit high, but I'd agree most PC users would have to spend more on the upgrade than it would be worth to them to have Vista. Some will upgrade for specific features, but most will leave their existing PCs with XP, and upgrade to Vista when they replace them, just as was the case with XP. I'd have done that myself if I hadn't had a need to buy a new PC shortly before Vista was released, and I planned all along to upgrade it to Vista.
It's probably good public relations, and that's about all. Linus Torvalds is a genial, broadly respected representative of Linux developers, and people listen to what he says (or read what he writes). A rather blunt statement of this kind might also stop the more outlandish FOSS supporters (I'll refrain from mentioning any specific names) weighing in further with their, shall we say, counterproductive hyperbole.
It really depends on how you use it, and which features are enabled. The recommended RAM is 1GiB for the premium versions, and 512MiB for the basic version. I'd add at least another 512MiB, or 1GiB if you use Microsoft Office, but if you only run a few things at once, those recommended limits might be alright.
Slashdot Mirror
Secondly, in the EU at any rate, we have this idea that there are certain rights that cannot be given up. One cannot give up basic rights to freedom by signing a contract agreeing to become someone else's slave, for example. This is the same sort of issue, just on a more modest level. Agreeing to use Google's search engine cannot take away the right to privacy.
If the goal is the best software, then the GPL is largely irrelevant. If the goal is the best open source (or Free) software, then the need for something like the GPL ultimately depends on whether or not the open source development model is more effective than the closed model. If it is, there isn't much of a need for preventing proprietary additions, because any temporary advantage from closed development will be negated over time, by parallel open development. If closed source is actually a more efficient development model, then something like the GPL becomes much more important. I don't know which is the case.
I'm not ideological, so I'm glad to be able to benefit from the incorporation of BSD sockets into proprietary software. If Windows 3.x/9x, SunOS, Mac OS and others hadn't been able to take advantage of BSD sockets code, it almost certainly would have prevented BSD sockets becoming a de facto standard. I don't know which networking API most of us would be using, but my guess would be something designed by Microsoft. Thanks to BSD sockets, all of these systems can interoperate well, and open source systems like Linux are actually useful as network servers, serving mostly Windows and Mac OS clients.
At least as important, I think, is if Dell actually work to ensure there are Linux drivers for the hardware they're selling. Linux hardware support on laptops is often so bad that it's effectively unusable. If Dell restrict their Linux offerings to a few specific hardware configurations that already have good hardware support, I don't think it will make much of a difference to the market.
Their thinking is probably that anyone who wants Linux will look for it, and anyone who doesn't know doesn't want it, but might not be able to understand the difference from a brief description. If users expecting Windows started buying PCs with Ubuntu because it's cheaper, it could lead to all sorts of negative customer service and publicity issues for Dell.
The first version of the OS that's today known as Windows was Windows NT 3.1, released in 1993. Since it was actually a completely new OS, it was really a '1.0' relese, but Microsoft used the same 'look and feel' as Windows 3.1, and offered a similar API (although it was 32-bit rather than 16-bit), so cleverly marketed it with the same 'Windows' brand and '3.1' version number. A subset of the Win32 API was later ported from Windows NT to DOS/Windows, resulting in Windows 95.
I'm sorry, but that's simply impossible. As I pointed out above, the '1.0' release of the operating system currently known as Windows was Windows NT 3.1, which shipped in September of 1993. The '1.0' release of Linux came six months later, March of 1994.On the Linux side, the GNU Project to create an open source clone of the Unix operating system was announced by Richard Stallman in 1983, coincidentally the same year Microsoft announced Windows. Some of the GNU codebase dates from 1984, but it's more difficult to indentify the equivalent of a 'release date' for open source software than for closed. Linus Torvalds started his computer science studies in 1988, and announced Linux in 1991, but the '1.0' release didn't come until 1994. Moreover, Linux made extensive use of existing GNU software.
On the whole, from a code perspective, Windows and Linux are roughly similar in age. From a brand perspective, Windows has an age advantage over Linux, but not GNU. Linux could have negated this by using the existing the 'GNU' brand (as Richard Stallman insists it ought to do), in the same way that NT used the existing 'Windows' brand, and thus have had effective parity with Windows, in terms of the brand's age. The fact that this was not done largely reflects Stallman's extremely poor choice of a name ('GNU'), which led to virtually no brand value being developed from 1983 to 1991, when the Linux brand came along. It was actually better to start over than continue to be saddled with such a poor brand name.
On the contrary, this is precisly why network effects are important. If the utility of the software to users is a function of both features and market share, a high-market-share producer can provide the same set of features as a low-market-share producer, but users will get more utility from the high-market-share good. Once market share has got past a certain point, you could conceivably have a sitaution where the added utility from the higher market share is greater than the total cost of development/production, meaning the low-market-share competitor, even with zero costs, would actually need a negative price to be competitive. (There's also the fact that most of the key Linux developers, for example, are paid rather than volunteers, but that's rather peripheral to the point.)
Another important issue is that Microsoft does more than just write the software. A lot of computer standards have been developed by dominant firms, e.g. the ACPI power management standard was developed by Microsoft and Intel, together with Toshiba, I believe. This is one of the reasons other PC operating systems like Linux are so far behind in power management support: they're playing catch-up. As long as Microsoft and Intel are defining the standards for the PC platform, they'll have an advantage in being able to provide products that take advantage of those standards first, giving them at least a window of time in which they can offer better products.
I suppose we'll just have to agree to disagree on this point. From my perspective, if Windows had 70% market share, with the other 30% divided amongst others, e.g. 10% for Mac OS X, 10% for Linux and 10% for all the rest, I'm not at all convinced this would have any impact on Microsoft's dominant position: is it worth duplicating OS-related cost structures to go from reaching 70% of customers to reaching 80%? Would the burden of interoperability fall on the users of the platform with 70% market share, or the ones that each have 10% or less?
If the market were a duopoly, with 70% for Windows and 30% for Linux (or perhaps it would have to be a specific distribution of Linux), that might be different, but with even a 70/20/10 split for Windows/Linux/others, it's still far from certain that the dominant-firm dynamics would change. With a lot of competitors, a firm can even hold a dominant position with a minority market share, as Microsoft arguably did in the late 1980s.
No, not at all. My assumption is rather that a market share of 70% for Windows, instead of a 90-95%, would not change the dominant-firm dynamics of the industry, including Microsoft's ability to set prices for Windows and benefit from network effects. If the dynamics did change, then of course the situation would be quite different.
Well, the high market share of Windows leads to pressure in the opp
Interesting, I didn't know about that. All I meant is it doesn't use the old BSD distinction of ports below 1024 being reserved for privileged users, with 1024 and above being open to unprivileged users. I suppose you could effectively set it up that way using port filtering.
As I've pointed out before, Microsoft's market share in the PC operating systems market was less than 50% in 1990, and only about 58% in 1992, yet practices such as tying and various forms of price discrimination were used very successfully: that's one of the reasons Microsoft's market share in the same market was above 90% by the end of the decade. Ironically, Microsoft probably ought to have stopped competing aggressively once it reached 70-80% market share, leaving an 'AMD', as it were.
If Microsoft had 70% paid market share, for example, it would still have a substantial advantage over its competitors, particularly if several were sharing the other 30%. If you start defining a 'monopoly' from the regulators' perspective as including firms with 70% market share, you're on a very slippery slope: Intel, Apple (iPod) and Cisco already have higher market shares than that, and Google could very well reach that level within the next few years, as could Windows Server (which is not currently subject to the regulations applied to the Windows desktop OS). Are they all monopolies?
The key, really, is that if the market naturally tends towards monopoly, then a producer with a higher market share can produce more efficiently than producers with lower market shares. If there are a number of such related markets, a firm which is dominant in several of them, earning supernormal profits, could very easily earn a larger overall economic profit that a firm with a monopoly (in the non-technical sense of a very high market share, e.g. above 90%) in one of the markets.
Dominant firms with very high market shares, whether Apple, Google, Intel, Cisco, Microsoft or others, must take care with their actions, but regulation of dominant firms is far more controversial than regulation of monopolies. If Microsoft Windows had held 70% of the desktop OS in the late 1990s, rather than more than 90%, it might very well have escaped classification as a monopoly. Microsoft might still have faced some penalties for anti-competitive practices, but being ruled a 'monopoly' had far-reaching implications.
On the whole, assuming the global 35% piracy rate holds for Microsoft Windows, and a 95% market share, this means 61.75% of users in the desktop OS market are paying Windows customers. Even if Microsoft were to capture only 10% of that 35% through a successful anti-piracy scheme, with the other 25% going to other operating systems, the percentage of paying Windows customers would rise to 71.25%. It's exceedingly unlikely that a ca. 70% market share for Microsoft Windows would be below the tipping point, hence it would have virtually no impact on Microsoft's ability to set prices. The increase in the number of paying customers would thus increase Microsoft's revenue from Windows by ca. 15%, and have virtually no impact on costs. On top of that, it would potentially reduce regulatory pressure, allowing more scope for Microsoft to leverage its dominance when competing in related markets.
I'm not sure about the 'techie' group either, but you have a good point there. It's not my field, but it seems to me that a lot of computer science education is already Linux-centric, because the source code is available to be studied, and it has the best hardware support of any open source OS, so students can easily modify it and test their modifications on common hardware. On the other hand, almost every other field is extremely Windows-centric, so once the technical students get out into the real world, they're forced to use Windows, because it's what the buyers of their services demand. I suppose the critical issue is how many semi-technical students there are, meaning those who build their own PCs, but wouldn't be comfortable using Linux, so would be willing to pay something for Windows, assuming they couldn't pirate it. That doesn't necessarily mean the current retail price, which is set with the assumption of high levels of piracy.
For a company like Microsoft, there are at least three or four different phases, and the implications of piracy are different in each.
1. Minor producer: if you're a minor producer with low market share, piracy may be good for gaining market share, as long as revenue from paying customers remains high enough to cover costs.
2. Dominant producer: if you're the dominant producer in your market, but perhaps still with only a minority share of the market, piracy is good, because most people pirating will be pirating the dominant product, This will spur a network effect, and any revenue implications are likely to be less important than for smaller producers.
3. (Near) monopoly, without regulation: if you've got a near monopoly, you'll gain the benefits of network effects. The network gains from piracy, and the extent to which it keeps out competitors, are both gains. Without viable alternatives, however, there is the potential for higher revenue from those who are pirating, but would pay if they had to. The network effect and the exclusion of new entrants might be worth more than the lost revenue.
4. (Near) monopoly, with regulation: if regulatory restrictions are imposed on a firm with a near monopoly, that means the gains of network effects and the prevention of new entrants are offset by both the lost revenue and the costs of the restrictions (e.g. no bundling, limitations on pricing strategies, etc.). I this case, the more onerous the restrictions, the less value there is from piracy. It may be worthwhile to give up unpaid market share, in exchange for higher revenue, especially if this leads to a reduction in regulation.
In the late 1980s and early 1990s, Microsoft was in category 2, with a dominant position, but a market share near the middle: it crossed the 50% mark in 1990. During this time, piracy was arguably good for Microsoft. By the middle of the 1990s, however, Microsoft had moved to category 3, and so whilst piracy was no longer as clear a benefit, it was still arguably less bad than good.
With the monopoly ruling against Microsoft in 2000, it moved into 4, although the level of regulation has varied. With the regulatory costs offsetting some of the network gains, piracy arguably became less valuable to Microsoft, and this may in part explain the increase in anti-piracy measures in Microsoft's software since then. Giving up some non-paying customers to competitors, in exchange for converting some non-paying to paying customers, is arguably a good strategy, especially if it reduces the regulatory pressure.
An interesting point is whether people who pirate Windows, and would switch to Linux or something else if they couldn't pirate it, are willing to pay for other software. The expected answer is no, so Microsoft could arguably give up these low-value customers without losing the benefits of being the dominant platform for commercial software development. Producers of commercial software would have little interest in developing for Linux if Linux users wouldn't buy their software anyway, so a higher market share for Linux would have little impact on the network effect there.
From the above, the risk of giving up some market share comes from network effects other than those relating to commercial application development. For example, people who won't pay for software may still pay for products and services bought over the web, etc., in which case they'd be targeted by website developers. They might also still be willing to buy relatively expensive hardware, which could reduce the network effect regarding device driver development.
At any rate, running in a chroot jail is arguably better in some ways than just running as an unprivileged user. Vista has some sandboxing features, using 'integrity levels' and redirecting various file and registry accesses to a 'virtual store', but I'm not really familiar with them, except for the basics, and I don't know if IIS uses them anyway.
If you include NT OS/2, which was started in 1988, in Bill Gates's 1987 prediction about OS/2, then I suppose you can say it was, in a way, right.
Well, Microsoft had less than half the OS market as late as 1990, and in 1987, its market share would have been even smaller than that, so nothing close to a monopoly. On the other hand, NT certainly wouldn't have become as dominant as it is today if it hadn't been compatible with the Windows 3.x line, which is what drove Microsoft from being just the largest desktop OS vendor to being virtually a monopoly (especially after Windows 95). That's of course why Microsoft wanted to add a Windows-like API to NT OS/2, which in turn led to the dissolution of the IBM/Microsoft partnership.
The reason I don't give a huge amount of weight to Bill Gates's technology predictions is because the context is completely different. In 1975, the personal computer industry was in its infancy: hardly anyone even knew about it, much less about selling software for microcomputers, if they even knew what software was at all. The combination of familiarity with software, knowledge of microcomputers and a grounding in business, which Gates had, was relatively rare. It wasn't altogether surprising that Bill Gates had a better ability to predict the rise of the software industry in the mid-1970s than most other people at the time: he had a relatively rare combination of information. Today, there is a massive number of people analysing the IT industry, and Microsoft itself, so the idea that Bill Gates is going to have a similarly rare set of information today isn't very plausible.
You could replace 'Bill Gates' with 'Tim Berners-Lee', who was right about the web (and even invented it), and come to much the same conclusion. The fact that Berners-Lee had a rare set of knowledge that allowed him to predict and invent the web doesn't necessarily mean he'll be better at predicting anything else.
For the owner of an individual PC, the situation is completely different, as it is with other personal possessions. If individual purchase decisions were purely utilitarian, we'd all dress like Soviet citizens of the 1950s, and those with cars and no children would be driving Fiat Puntos, or something like that. Why do people buy nice-looking clothes, flash cars and high-priced electronics, from mobiles to TVs to computers? It's all down to tastes and disposable income: non-utilitarian product features actually matter to most consumers, in most purchase decisions.
Again, this utilitarian view applies in an organisational setting, but not for individuals. I imagine a lot of organisations will put off upgrading unless they need specific features in Vista, just like they did with the 2000 to XP upgrade, or the NT 4.0 to 2000 upgrade before it. The key market in the short run is clearly individual consumers, not organisations, but the latter have always followed eventually, once the incremental improvements came to dominate the initial costs of unfamiliarity and more limited hardware support, both of which decline over time. I'm not surprised. Their job is made much easier if everyone is running exactly the same software, on exactly the same hardware. That's one reason organisations have always been slow to upgrade.The AOL takeover of Time Warner involved using ludicrously overpriced AOL shares to effectively buy Time Warner. The deal was only possible because of the .com bubble, and when the bubble burst, the major effect was that Time Warner shareholders had handed over most of the company to AOL shareholders, for very little in return. It's a completely different situation.
The Unix subsystem for Windows, which is what provides POSIX compliance (or at least did at one point), uses Unix semantics for files, so a file or directory opened by a Unix process can have its name removed from the file system tree immediately, and will be removed from the disk when the last handle is closed. However, for some reason there's an exception to this for executables mapped into a process. These can be renamed, but not deleted. It looks like a bug to me, but maybe there's a reason.
I agree with you, however, that I'd really like to see the Unix semantics offered for Win32, though some optional global setting. I'd like to see the bug with unlinking mapped executables under the Unix subsystem fixed too.
If Microsoft believe their standard, OOXML, is better than ODF, or that they have an advantage in spurring its adoption, then there's no need to try and block ODF. All they have to do is make sure OOXML is on a level field with ODF, and wait for it to win. Alternatively, if they think ODF is as good as or better than OOXML, they can add support for both formats, and use their influence to guide the winning standard, whichever it is.
At the end of the day, if customers care more about open formats than rapid addition of features, Microsoft will shift towards open formats. The alternative would be a decrease in competitiveness vis-à-vis competitors using open formats, such as StarOffice/OpenOffice.
The only real risk of open formats to Microsoft's business comes from making it easier to migrate from MS Office to competing products. However, if Microsoft believe they can produce a better office suite than their competitors, this isn't really a threat at all. Mind you, I mean 'better' in the sense of providing value to users, which may include non-technical features, such as similarity to previous versions.
In technical terms, going from Windows XP to Windows Vista is at least as big a step as NT 3.1 to 3.5/3.51, NT 4.0 to NT 5.0 (Windows 2000) or NT 5.0 to NT 5.1 (Windows XP). The only release that's really comparable was NT 4.0, and it's arguable which of the two releases was a bigger evolutionary step.
95% is probably a bit high, but I'd agree most PC users would have to spend more on the upgrade than it would be worth to them to have Vista. Some will upgrade for specific features, but most will leave their existing PCs with XP, and upgrade to Vista when they replace them, just as was the case with XP. I'd have done that myself if I hadn't had a need to buy a new PC shortly before Vista was released, and I planned all along to upgrade it to Vista.It's probably good public relations, and that's about all. Linus Torvalds is a genial, broadly respected representative of Linux developers, and people listen to what he says (or read what he writes). A rather blunt statement of this kind might also stop the more outlandish FOSS supporters (I'll refrain from mentioning any specific names) weighing in further with their, shall we say, counterproductive hyperbole.