But let's be completely honest: GPL-ed software is not literally and wholly free. It is "mostly" free (yes, I can distinguish between free beer and free speech). I can't do just anything I wish with it.
What are you complaining about? When you get GPL'ed software, you get good software, its source code, and a limited redistribution license. That's a great deal better than you get with most commercial software; Microsoft, for example, doesn't even let you redistribute their software, let alone modify its source, even though you paid them for it.
Maybe the GPL license doesn't fit your needs. In that case, you can exercise your own choice: don't use the GPL'ed code. Nobody is forcing you. If you like, you can even create your own, proprietary implementation of a GPL'ed library, an option you generally don't have with closed source software.
This is not a free market. My competitors do not compete on equal terms. They have salaries coming from elsewhere (another company, government, unemployed, studying) and can afford to dump the prices without risking their financial situation.
"Dumping" is carried out by companies and countries that want to temporarily undercut prices so that they can later have a monopoly position in the market and charge more for their own products. Free software may end up taking over a market by being cheaper, but by its nature, nobody can charge monopoly prices for it--it is and remains free. Furthermore, the users of free software pay for it, directly and efficiently, in their own contributions, avoiding the overhead of commercial software development and corporations. We call that "competition", not "dumping".
Microsoft, on the other hand, has engaged in something one might call "dumping": they have temporarily undercut competitors, swallowed the losses temporarily, and later (effectively) raised prices on their products.
It is impossible to compete with people on those terms, driving all commerical vendors out of a given market.
And that's the way it should be. Once the development costs of a piece of software have been amortized, it costs nothing to make an additional copy. In an efficient market, the price of software should therefore go to zero. Open source software is simply one of several means by which that happens.
The fact that Microsoft and a few other players continue to make big bucks with old technology is an indication that either they aren't selling software (maybe they are selling services or something less tangible like membership in a "user community"), or that they are engaging in monopolistic practices.
Then everyone has to rely on freely developed software, without support or someone interested in the "customers".
Providing support for free software costs money, and that's why it isn't free. That's also why it is a great opportunity for consulting and for making money.
The enemies of the free market are people like you, not free software. You have unreasonable expectations of the big bucks you can make with software development, and you expect the government to protect you from cheap competition. Well, things fortunately don't work that way. Get used to it, and maybe find a more profitable market niche.
Yes, that $1700 for the Macintosh is overpriced. You can get 1GHz+ Athlons with similar hardware plus DVD and better graphics for around $1000. They won't be as well made, but they will work quite well, and they will run a lot more software.
Why is there not one workingd GUI/graphic engine on Unix that is not X11
Because X11 works and because it works well. X11 even allows applications from different desktops and toolkits to work together reasonably well.
Lack of responsiveness of some GUIs on Linux has nothing to do with X11, it's a consequence of toolkit design. Talking to developers of several toolkits for X11, I found that many of them have disdain for the X11 APIs and just don't expect it to do well. When their toolkits perform poorly, they don't bother to look to their code, they just assume it's X11's fault. That way, performance problems in toolkits never get fixed.
The fact that "that other OS" can be sluggish also suggests that it isn't X11's design. X11's design may not be all that pretty after nearly two decades of evolution, but it gives a lot of functionality that other window systems lack, and the mechanisms X11 uses for local communications (unix domain sockets, shared memory) are reasonably efficient.
Re:you've fallen for MS strategy
on
Mozilla 0.9 Out
·
· Score: 2
Actually, I was just thinking of HTML, DOM, and style sheets. XML, XML Schema, XSL, SOAP, etc. are yet to come, but they are probably already consuming lots of time for other browser developers.
Microsoft generally encourages complexity in these standards, because they know they have more resources than their competitors to implement it and don't need a focussed standard. It's hard enough to keep committee-designed standards reasonably small to begin with, and with a behemoth like Microsoft weighing in on the side of complexity, it's basically impossible: everybody wants their pet features in every standard.
In order to access your data, created with Windows applications, you need Mirosoft Windows. These kinds of licensing schemes are holding customers hostage: if you want to be able to get at your data beyond three years, you better pay up.
I think sooner or later, people are going to figure out that this is a bad deal. And that's good for Linux.
There are two kinds of emulations possible. The first is complete emulation, in which off-the-shelf Windows games just run on Linux. That is really bad, I think: the vendor doesn't know how large their Linux user community is, and they may easily end up using Windows APIs that trip up Wine. The second is one in which the vendor uses Linux versions of Windows APIs and only makes minor changes to packaging and the installer. That avoids many of the problems. In fact, where Windows's APIs are good, that may even be a good idea and encourage native versions of those APIs.
But in many cases, I think the Linux APIs are preferable, and a full Linux port seems altogether better. Ports to native Linux will drive the Linux APIs to improve further and get hardened, and that's important: without that kind of real-world usage, Linux multimedia and 3D APIs will just get stale. And native ports will likely have better performance: Windows APIs make all sorts of assumptions about the underlying OS and kernel that just aren't true in Linux.
But why buy commercial games at all? Have twice the fun: write your own. Yes, getting commercial game quality graphics and sound is hard, but we can make up for that with smarter, more fun games. To me, games like nethack still have better game play than any of the Windows equivalents, which have nice graphics but are much more simplistic.
you've fallen for MS strategy
on
Mozilla 0.9 Out
·
· Score: 2
IE has pulled so far ahead because Microsoft has been pushing for a complex mess of standards, as part of the W3C and outside, while implementing these features in parallel in IE. They have been cheered along in this effort by academics and startups who want their overly complex and mostly useless features added to the web standards. Microsoft has said as much that this was going to be their strategy with respect to open source, and it is similar to what they have used against their previous competitors.
It's no wonder that any other browser has a hard time catching up with the current W3C standards. The needs of web users would be better served with a much simpler set of standards. And previous multimedia and hypertext systems have shown that much simpler systems are possible.
Konqueror is good, and I am using it as my standard browser, but I think it still has occasional problems. First, it doesn't seem to work with some of the more complex JavaScript out there. There seem to be some problems with style sheets. (And does anybody know how to display a JavaScript console in it?) Still, Konqueror is very good for most web sites.
I think the Konqueror license presents a problem. Its license will limit its adoption in commercial products, but a browser succeeds by being adopted widely. With Mozilla getting better in terms of features, stability, and performance, I would expect Mozilla to predominate ultimately.
The US has been trying to dictate US-style business practices for a long time. In many areas that is actually good, but when it comes to privacy, US laws and practices are unacceptably poor. Rampant identity theft and theft of large numbers of credit card numbers and other customer information (kept around by web sites long after an order has been fulfilled) in the US are examples of that. It is good that Europe is putting their foot down on this matter.
And Europe certainly has the clout and experience to do so. B2C E-commerce has existed in Europe about a decade longer than in the US, and Europe itself is a multicultural economy comparable in size to the US and with a significantly larger population.
But Windows would have to drop features in order to become "UNIX" in the original sense of the word: the spirit of UNIX is simplicity and elimination of extraneous features. In that sense, Linux and BSD are also straying ever further from the original UNIX ideals.
BTW, I believe the NT kernel would easily support hard links, just like it easily supports UNIX mount points, and a lot of other features. Microsoft just doesn't want those features in there.
There have been numerous attempts at reproducing color with b/w emulsions. This is one of them. Several others used patterned filters, not unlike the color filters found in today's CCD cameras. All of them were difficult to reproduce and required precise alignment. That's why, ultimately, color emulsions won out.
MS is powerful not because of ignorance, but because their software is Good Enough.
But their software isn't "good enough". Among friends who aren't computer experts, many paid thousands of dollars for the Windows PCs and they end up gathering dust in a corner somewhere because they don't work (or, worse, computer expert friends get called in to fix the Microsoft messes).
People buy Windows because marketing and market share tell them it's the thing to buy. Even though it, too, has problems, MacOS X probably would be a better choice, but people don't buy it because it costs extra and the benefit is not clearly obvious before the purchase.
In fact, I think none of the desktop OSes is really "good enough" for non-expert use yet. But nobody is really interested in trying to create such a system because there is no money in it: either it doesn't catch on, or it gets cloned by Microsoft. And the open source software community hasn't been able to produce anything other than look-alikes of Windows and MacOS for their desktops either, copying most of the misfeatures of those systems.
In the Star Wars films, George Lucas makes lavish use of computer- generated characters and scenes, but they never overwhelm the intriguing characters at the center of the saga.
There were "intriguing characters" in StarWars? Where? I must have missed them.
As long as Windows programs don't use UNIX/POSIX APIs as their primary development APIs, Windows isn't UNIX. Since Microsoft sees the Win32 APIs as a tool for customer retention, they aren't going to change any time soon.
In order for NT to become UNIX, they would have to drop a lot of the junk they added. Many of the "features" in NT (features Microsoft likes to tout as innovative) were considered for UNIX many years ago and rejected. Cutler didn't get the UNIX philosophy of simplicity with VMS, and he still doesn't get it with NT. (However, Linux and *BSD would do well to think about simplicity as well: their kernels are becoming Rube Goldberg contraptions and dumping grounds for every feature anybody can think of as well.)
Apple, on the other hand, is pushing UNIX APIs (and libraries built on top of them) for new applications development. Of course, there is a business reason there as well: Apple is behind in terms of software and programmer community, so it makes sense for them to go with something standard.
I think it's better to get your lifetime email address from somewhere like computer.org, ieee.org, acm.org, or your preferred professional society.
I would recommend against using those E-mail addresses: they are tied to continued membership in those organizations, and you may at some point decide to leave them. After all, membership is expensive, benefits are minimal, and the organizations may take political positions that you disagree with.
The IEEE was particularly bad: when I renewed late one year, they immediately reassigned my E-mail address to someone else and didn't give it back. Any well-run organization should at least have a non-trivial exclusionary period during which an address can't be reassigned; anything else is a security problem. I also found IEEE customer support in general pretty slow and unsatisfactory, and the E-mail forwarding was unreliable anyway when I first got it (maybe they have fixed it by now).
So far MOSIX was developed 7 times, for different versions of UNIX and architectures. It has been used as a production system for many years. The first PC version was developed for BSD/OS. The latest version is for Linux on X86/Pentium/AMD platforms.
Yes, they did start out basing their system on proprietary kernels, then they moved to BSD, then to Linux. The current work is not about the basic idea anymore, moving processes around somehow, but about things like distributed virtual memory, distributed file systems, and migration strategies.
This isn't "playing catch-up", it is cutting edge research by the people who did the original work moving to the BSD and Linux platforms because they are more widely available, are better supported, are easier to license and share, and have more software available for them.
Does that mean they are implementing something like crowds?
I just hope they do it right, because making anonymity work is a bit more complex than just shipping stuff through a bunch of intermediaries.
It doesn't matter how you divide the airwaves in terms of time or frequency, the total bandwidth is constant. Transmissions in the time domain may look like they utilize bandwidth better, but that's an illusion. Time domain transmissions simply spread their data through currently "unused" frequency bands, but those bands are allocated and unused for a reason, and it isn't acceptable to sell and deploy large numbers of transmitters that deliberately add noise to those bands. It's like the bank employee that gets rich by stealing a penny from every account. Existing spread spectrum technologies, on the other hand, aim for taking maximum advantage of the bands they have been allocated without stepping on other allocations.
If time domain methods caught on widely, we'd have to partition the space of sequences into different "bands" so that particular users are guaranteed a certain amount of bandwidth. In the end, we'd have replaced our frequency based system with a time domain based system and we'd greatly increase the cost and complexity of even the simplest transmitters. That's a nice deal for shareholders of companies pushing that technology, but it isn't for everybody else.
Maybe eventually, it does make sense to move to time domain systems. But let's make that step deliberately and without believing in the existence of some magical, hitherto undiscovered bandwidth.
The problem is that the things were ever referred to as database systems.
Well, given that it was priced and marketed like a high-end, enterprise-grade database system, that kind of seems reasonable.
What we should have done from day one was to sell persistence for C++.
Indeed. But a persistence library for C++ might cost a few hundred dollars per developer and have no or minimal per copy runtime costs. ObjectStore was priced out of that market by orders of magnitude.
In addition to the license costs itself, there are training costs, retooling, and the cost associated with the risk of picking a single vendor solution. Even if you had given ObjectStore away for free it would have been difficult to displace RDBMSs.
The best chance for success I see these days would be to have a simple, reasonably good open source OODBMS and make money on management tools and high performance versions. Still not the stuff of billion dollar companies, but a decent living.
That strategy was necessary in some ways, because we were venture-funded, and the VCs weren't going to be happy with a small niche. They wanted something that would get into every insurance company and bank. However, by aiming high and failing (by VC standards), we abandoned our natural market too soon and avoided becoming a small success in that market.
It's unfortunate that good technology like ObjectStore failed, but ultimately the choice was yours when you accepted the money and the business model.
I think the main reasons people aren't even considering OODBMSs is the price, limited language bindings, and the lack of standardization. RDBMSs, in contrast, are standardized and accessible from a large number of languages with a single API, and there are numerous excellent free implementations.
In fact, none of the open source implementations mentioned even comes close. Ozone, XL2, and Zope are quite language specific and don't even have C++ bindings. (There is a free OODBMs for C++, the Texas Persistent Store, but I think it also has many limitations.) FramerD isn't even an OODBMS because it doesn't attempt to bind language objects to database objects but introduces its own data model.
More generally, I think the traditional OODBMS approach turns out to be too rigid and too low level for many applications. If someone going to keep data around for a long time, they don't want it abstracted and encapsulated, they want concrete, exposed, well-defined representations with powerful operations for manipulating it. RDMBSs provide that. I think among the non-RDBMs systems, FramerD comes closest to that.
So, there are practical reasons why people don't use OODBMSs. But I think there are also some fundamental theoretical issues having to do with how data is manipulated and data models evolve. Still, OODBMSs have many attractive features, so one can hope that they will evolve to meet people's needs more. For some applications, they are already preferable.
Thanks to the simplicity of XML it got the widespread usage that SGML never managed,
XML's popularity probably has little to do with its design. SGML was marketed as a text markup language. XML is being marketed as a universal data representation. The market for the latter is several orders of magnitude larger than for the former, and that accounts for XML's popularity. In terms of its design, XML is neither particularly simple (compared to alternatives), nor particularly clean.
First, for XML itself. What is XML? A standard way to store and describe data in a manner that is readily addressable by virtually any computing platform. [...] What else offers that?
There are many existing textual representations that are equivalent in power to XML but a lot simpler. The simplest example would be Lisp's textual representation. Lisp's textual representation is a lot easier to define and parse than XML. In fact, any collection of functions and type constructors in a programming language, together with the syntax of that programming language, define such a representation. A Schema corresponds to a type system in such a representation.
I don't think XML has been very well thought out. It's a standard for data representation, but it's based, through historical accident, on a standard for text markup, and that causes all sorts of problems. Still, despite its failings and shortcomings, at least XML gets the industry away from junk like OLE structured storage, Bento, or ad-hoc binary formats. For that, I'm willing to live with XML's messy syntax and semantics.
intellectual property rights not violated
on
MS VP Speech Online
·
· Score: 2
Mundie keeps saying that open source violates intellectual property rights. Of course, it doesn't. People give away their source code freely, and that is entirely within their intellectual property rights.
What is going on is really that market mechanisms force the "incremental price" of delivering a piece of software to an additional customer to zero, reflecting its zero incremental cost. That's what should happen in an efficient market, and open source happens to be the mechanism that makes it happen.
Microsoft is scared of this. They like to paint the free market mechanisms that are undercutting their business model as some kind of "un-American socialism". But open source is doing what the free market should be doing: driving profits to zero. What is "un-American" and anti-free-market is the artificial maintenance of disequillibrium wages.
Microsoft already enjoys enormous advantages over any competitor, through their installed base, network effects, business relationships, marketing muscle, brand name recognition, and willingness to engage in unfair business practices. Yet, open source already threatens their products. Imagine what would happen if Microsoft actually were willing to compete equally and fairly in a free market.
Slashdot Mirror
What are you complaining about? When you get GPL'ed software, you get good software, its source code, and a limited redistribution license. That's a great deal better than you get with most commercial software; Microsoft, for example, doesn't even let you redistribute their software, let alone modify its source, even though you paid them for it.
Maybe the GPL license doesn't fit your needs. In that case, you can exercise your own choice: don't use the GPL'ed code. Nobody is forcing you. If you like, you can even create your own, proprietary implementation of a GPL'ed library, an option you generally don't have with closed source software.
"Dumping" is carried out by companies and countries that want to temporarily undercut prices so that they can later have a monopoly position in the market and charge more for their own products. Free software may end up taking over a market by being cheaper, but by its nature, nobody can charge monopoly prices for it--it is and remains free. Furthermore, the users of free software pay for it, directly and efficiently, in their own contributions, avoiding the overhead of commercial software development and corporations. We call that "competition", not "dumping".
Microsoft, on the other hand, has engaged in something one might call "dumping": they have temporarily undercut competitors, swallowed the losses temporarily, and later (effectively) raised prices on their products.
It is impossible to compete with people on those terms, driving all commerical vendors out of a given market.
And that's the way it should be. Once the development costs of a piece of software have been amortized, it costs nothing to make an additional copy. In an efficient market, the price of software should therefore go to zero. Open source software is simply one of several means by which that happens.
The fact that Microsoft and a few other players continue to make big bucks with old technology is an indication that either they aren't selling software (maybe they are selling services or something less tangible like membership in a "user community"), or that they are engaging in monopolistic practices.
Then everyone has to rely on freely developed software, without support or someone interested in the "customers".
Providing support for free software costs money, and that's why it isn't free. That's also why it is a great opportunity for consulting and for making money.
The enemies of the free market are people like you, not free software. You have unreasonable expectations of the big bucks you can make with software development, and you expect the government to protect you from cheap competition. Well, things fortunately don't work that way. Get used to it, and maybe find a more profitable market niche.
Yes, that $1700 for the Macintosh is overpriced. You can get 1GHz+ Athlons with similar hardware plus DVD and better graphics for around $1000. They won't be as well made, but they will work quite well, and they will run a lot more software.
Because X11 works and because it works well. X11 even allows applications from different desktops and toolkits to work together reasonably well.
Lack of responsiveness of some GUIs on Linux has nothing to do with X11, it's a consequence of toolkit design. Talking to developers of several toolkits for X11, I found that many of them have disdain for the X11 APIs and just don't expect it to do well. When their toolkits perform poorly, they don't bother to look to their code, they just assume it's X11's fault. That way, performance problems in toolkits never get fixed.
The fact that "that other OS" can be sluggish also suggests that it isn't X11's design. X11's design may not be all that pretty after nearly two decades of evolution, but it gives a lot of functionality that other window systems lack, and the mechanisms X11 uses for local communications (unix domain sockets, shared memory) are reasonably efficient.
Microsoft generally encourages complexity in these standards, because they know they have more resources than their competitors to implement it and don't need a focussed standard. It's hard enough to keep committee-designed standards reasonably small to begin with, and with a behemoth like Microsoft weighing in on the side of complexity, it's basically impossible: everybody wants their pet features in every standard.
I think sooner or later, people are going to figure out that this is a bad deal. And that's good for Linux.
But in many cases, I think the Linux APIs are preferable, and a full Linux port seems altogether better. Ports to native Linux will drive the Linux APIs to improve further and get hardened, and that's important: without that kind of real-world usage, Linux multimedia and 3D APIs will just get stale. And native ports will likely have better performance: Windows APIs make all sorts of assumptions about the underlying OS and kernel that just aren't true in Linux.
But why buy commercial games at all? Have twice the fun: write your own. Yes, getting commercial game quality graphics and sound is hard, but we can make up for that with smarter, more fun games. To me, games like nethack still have better game play than any of the Windows equivalents, which have nice graphics but are much more simplistic.
It's no wonder that any other browser has a hard time catching up with the current W3C standards. The needs of web users would be better served with a much simpler set of standards. And previous multimedia and hypertext systems have shown that much simpler systems are possible.
I think the Konqueror license presents a problem. Its license will limit its adoption in commercial products, but a browser succeeds by being adopted widely. With Mozilla getting better in terms of features, stability, and performance, I would expect Mozilla to predominate ultimately.
And Europe certainly has the clout and experience to do so. B2C E-commerce has existed in Europe about a decade longer than in the US, and Europe itself is a multicultural economy comparable in size to the US and with a significantly larger population.
Hello??? Didn't I just say that Windows can support these?
The issue is more like "hey, what should all these _millions_ of apps do when they see a symlink when they've never worried about them before ?"
Microsoft changes Windows in much more fundamental ways with every release. Links are transparent to most applications anyway.
If you want a singly rooted fs, symlinks, inetd, and all those on NT/w2k , you can get them today, via Services for Unix.
Which brings us back to my original point: the problem with Windows is not that it doesn't have enough functionality, it is that it has too much.
BTW, I believe the NT kernel would easily support hard links, just like it easily supports UNIX mount points, and a lot of other features. Microsoft just doesn't want those features in there.
There have been numerous attempts at reproducing color with b/w emulsions. This is one of them. Several others used patterned filters, not unlike the color filters found in today's CCD cameras. All of them were difficult to reproduce and required precise alignment. That's why, ultimately, color emulsions won out.
But their software isn't "good enough". Among friends who aren't computer experts, many paid thousands of dollars for the Windows PCs and they end up gathering dust in a corner somewhere because they don't work (or, worse, computer expert friends get called in to fix the Microsoft messes).
People buy Windows because marketing and market share tell them it's the thing to buy. Even though it, too, has problems, MacOS X probably would be a better choice, but people don't buy it because it costs extra and the benefit is not clearly obvious before the purchase.
In fact, I think none of the desktop OSes is really "good enough" for non-expert use yet. But nobody is really interested in trying to create such a system because there is no money in it: either it doesn't catch on, or it gets cloned by Microsoft. And the open source software community hasn't been able to produce anything other than look-alikes of Windows and MacOS for their desktops either, copying most of the misfeatures of those systems.
There were "intriguing characters" in StarWars? Where? I must have missed them.
In order for NT to become UNIX, they would have to drop a lot of the junk they added. Many of the "features" in NT (features Microsoft likes to tout as innovative) were considered for UNIX many years ago and rejected. Cutler didn't get the UNIX philosophy of simplicity with VMS, and he still doesn't get it with NT. (However, Linux and *BSD would do well to think about simplicity as well: their kernels are becoming Rube Goldberg contraptions and dumping grounds for every feature anybody can think of as well.)
Apple, on the other hand, is pushing UNIX APIs (and libraries built on top of them) for new applications development. Of course, there is a business reason there as well: Apple is behind in terms of software and programmer community, so it makes sense for them to go with something standard.
I would recommend against using those E-mail addresses: they are tied to continued membership in those organizations, and you may at some point decide to leave them. After all, membership is expensive, benefits are minimal, and the organizations may take political positions that you disagree with.
The IEEE was particularly bad: when I renewed late one year, they immediately reassigned my E-mail address to someone else and didn't give it back. Any well-run organization should at least have a non-trivial exclusionary period during which an address can't be reassigned; anything else is a security problem. I also found IEEE customer support in general pretty slow and unsatisfactory, and the E-mail forwarding was unreliable anyway when I first got it (maybe they have fixed it by now).
Yes, they did start out basing their system on proprietary kernels, then they moved to BSD, then to Linux. The current work is not about the basic idea anymore, moving processes around somehow, but about things like distributed virtual memory, distributed file systems, and migration strategies.
This isn't "playing catch-up", it is cutting edge research by the people who did the original work moving to the BSD and Linux platforms because they are more widely available, are better supported, are easier to license and share, and have more software available for them.
Does that mean they are implementing something like crowds? I just hope they do it right, because making anonymity work is a bit more complex than just shipping stuff through a bunch of intermediaries.
If time domain methods caught on widely, we'd have to partition the space of sequences into different "bands" so that particular users are guaranteed a certain amount of bandwidth. In the end, we'd have replaced our frequency based system with a time domain based system and we'd greatly increase the cost and complexity of even the simplest transmitters. That's a nice deal for shareholders of companies pushing that technology, but it isn't for everybody else.
Maybe eventually, it does make sense to move to time domain systems. But let's make that step deliberately and without believing in the existence of some magical, hitherto undiscovered bandwidth.
Well, given that it was priced and marketed like a high-end, enterprise-grade database system, that kind of seems reasonable.
What we should have done from day one was to sell persistence for C++.
Indeed. But a persistence library for C++ might cost a few hundred dollars per developer and have no or minimal per copy runtime costs. ObjectStore was priced out of that market by orders of magnitude.
In addition to the license costs itself, there are training costs, retooling, and the cost associated with the risk of picking a single vendor solution. Even if you had given ObjectStore away for free it would have been difficult to displace RDBMSs.
The best chance for success I see these days would be to have a simple, reasonably good open source OODBMS and make money on management tools and high performance versions. Still not the stuff of billion dollar companies, but a decent living.
That strategy was necessary in some ways, because we were venture-funded, and the VCs weren't going to be happy with a small niche. They wanted something that would get into every insurance company and bank. However, by aiming high and failing (by VC standards), we abandoned our natural market too soon and avoided becoming a small success in that market.
It's unfortunate that good technology like ObjectStore failed, but ultimately the choice was yours when you accepted the money and the business model.
In fact, none of the open source implementations mentioned even comes close. Ozone, XL2, and Zope are quite language specific and don't even have C++ bindings. (There is a free OODBMs for C++, the Texas Persistent Store, but I think it also has many limitations.) FramerD isn't even an OODBMS because it doesn't attempt to bind language objects to database objects but introduces its own data model.
More generally, I think the traditional OODBMS approach turns out to be too rigid and too low level for many applications. If someone going to keep data around for a long time, they don't want it abstracted and encapsulated, they want concrete, exposed, well-defined representations with powerful operations for manipulating it. RDMBSs provide that. I think among the non-RDBMs systems, FramerD comes closest to that.
So, there are practical reasons why people don't use OODBMSs. But I think there are also some fundamental theoretical issues having to do with how data is manipulated and data models evolve. Still, OODBMSs have many attractive features, so one can hope that they will evolve to meet people's needs more. For some applications, they are already preferable.
XML's popularity probably has little to do with its design. SGML was marketed as a text markup language. XML is being marketed as a universal data representation. The market for the latter is several orders of magnitude larger than for the former, and that accounts for XML's popularity. In terms of its design, XML is neither particularly simple (compared to alternatives), nor particularly clean.
There are many existing textual representations that are equivalent in power to XML but a lot simpler. The simplest example would be Lisp's textual representation. Lisp's textual representation is a lot easier to define and parse than XML. In fact, any collection of functions and type constructors in a programming language, together with the syntax of that programming language, define such a representation. A Schema corresponds to a type system in such a representation.
I don't think XML has been very well thought out. It's a standard for data representation, but it's based, through historical accident, on a standard for text markup, and that causes all sorts of problems. Still, despite its failings and shortcomings, at least XML gets the industry away from junk like OLE structured storage, Bento, or ad-hoc binary formats. For that, I'm willing to live with XML's messy syntax and semantics.
What is going on is really that market mechanisms force the "incremental price" of delivering a piece of software to an additional customer to zero, reflecting its zero incremental cost. That's what should happen in an efficient market, and open source happens to be the mechanism that makes it happen.
Microsoft is scared of this. They like to paint the free market mechanisms that are undercutting their business model as some kind of "un-American socialism". But open source is doing what the free market should be doing: driving profits to zero. What is "un-American" and anti-free-market is the artificial maintenance of disequillibrium wages.
Microsoft already enjoys enormous advantages over any competitor, through their installed base, network effects, business relationships, marketing muscle, brand name recognition, and willingness to engage in unfair business practices. Yet, open source already threatens their products. Imagine what would happen if Microsoft actually were willing to compete equally and fairly in a free market.