Actually, I am one of those ex-Mac users that decided that what I really wanted was Emacs, FVWM, and a bunch of xterms. Screw GUIs (unless I am drawing, of course).
A Macintosh is nothing more than a giant hardware key. Apple doesn't have to worry about installing pirated copies of Mac OS simply because you can't get Mac compatible hardware unless you buy it from them (with a Mac OS license, of course). I am surprised that they don't give the new software away. After all, what better way to sell new Apple hardware than to write software that is to bloated to use on your old hardware. That's why you can guarantee that new versions of OS X aren't going to run any better on G3s than they do now.
Apple is a hardware company, the only reason they write software is to give their hardware something to do.
As far as porting to x96 is concerned, Apple knows that the x86 OS market is getting a little crowded as it is. Between Linux, FreeBSD, and a handful of versions of Windows there is something for everyone. Linux and FreeBSD are having a hard time getting a run and they are available free of charge, come with source, and are actually getting pretty darn useable. Mac OS X wouldn't have a chance. Sure the interface is slick, but it won't run existing Windows software (and it wouldn't run existing Mac software either), it wouldn't support very much hardware (probably less than Linux), and you wouldn't be able to get it preinstalled. Bill and Co. would make sure of that.
Apple has a fairly large market of users who are more than happy to pay premium prices for PC hardware, and they are happy with that. If Apple really wanted to see a Mac OS rennaisance what they need to do is price their hardware so that it is price competitive. Otherwise, no dice.
I used to feel the same way, but now I have a hard time reading physical books. I set up the Weasel reader (formerly GutenPalm) on my Handspring so that it scrolls automatically (screen wrap style) and now I can read for hours without having to move a muscle. The backlight means I can read in bed without waking my wife, and I can carry a wide selection of books around with me all of the time. I can even do keyword searches (try that with a book). The real problem so far has been getting reading material that was worth my time, but between baen.com and the Gutenberg project I have managed to keep myself pretty well stocked.
I am a fairly new convert to the Palm platform; I have only owned my Handspring for a few weeks, and I used to swear that reading on a handheld would never take off. I didn't think that anyone would pay several hundred dollars for a gizmo that would allow them to read e-books. Nowadays, however, you can get an 8M Handspring for less than $100, and being able to read in the dark without waking my wife is priceless.
The butterfly ballots are easy enough to use. The Democrats just realized that they were going to lose (by an amazingly small number of votes) and they hoped to "rectify" things. In cases like that it is always possible to find folks that will come forward and testify that the system was "too hard."
The one thing I like about a punch card system is that it is easy for the voter to verify that that their vote was cast correctly. They simply take the ballot out of the machine and make sure that the holes are in the right places. With computerized terminals it would be too easy for someone to write a one line Perl script that changed every third vote for their oponent into a vote for them. Now, if the computerized touch screen system also generated a physical ticket that the voter could verify (and which would be used for hand counts) then I am all for it. I am all for simplifying the system, but I want to be able to verify that no one is diddling the results.
Then perhaps they'll require all the comments in the Open Source to be available in German, and then start requiring a certain percentage of all the software government uses to be of German origin.
Do we want to encourage that sort of thing?
Probably not, but then again, I don't see how we can stop it. For some applications (say German tax applications) I don't even really see the harm. I think that English will remain the Lingua Franca for Free Software hackers for some time yet, but a certain amount of regionalization is inevitable. I participate in several mailing lists where there is a significant non-English speaking contingent, but English is still what gets posted to the lists (and put in the code). If you want to create a truly global project you don't really have much choice of the language to use. However, that could easily change.
This case is a bit different because the German government is even more of a monopoly than Microsoft. Nearly everyone does business with their government, and if the government asks for documents in StarWriter format, you don't send them a Word document.
That's why Microsoft is up in arms, they know that if the German government starts using (and mandating) StarOffice then they are likely to see sales of the German edition of MS Office plummet. Even if relatively few machines in the Bundestag are converted it could have a totally devastating effect due to the millions of folks that are required to share documents with the Bundestag.
The fact that MS Office is somewhat more featureful than StarOffice won't matter one bit, because if you are dealing with the government the most important feature is using the same Office suite that they are using. The fact that StarOffice is a good product that is available for free (and that was originally made in Germany) doesn't hurt either. If the government were to switch, it probably wouldn't take too long for StarOffice to become the official German Office suite. The only reason that Microsoft keeps mentioning Linux as their number one threat is that they don't want to tip people off about StarOffice. Changing your operating system is hard, and requires you to learn a whole new set of programs. Replacing MS Office with StarOffice is much easier, and it carries almost all of the same financial benefits of a full migration to Linux (Windows is included in the price of nearly all machines MS Office isn't).
We are not talking about maintenance jobs. Linux, properly deployed, should actually lower maintenance costs. We are talking about programming jobs. Linux would allow the local software market to flourish instead of being tied to a handful of companies in the United States. Why should the Germans spend their money on U.S. software when they can get Linux solutions from a local German vendor (at a lower price).
For far less than a Microsoft solution the Bundestag could get a custom solution based on Linux and other open standards. This money they spend will even stay in their own economy (meaning that they will get the bulk of it back in taxes).
Let's face it, the reason that this appears to have actual supporters in the Bundestag is because it makes good political sense. Not only would using Linux lower the cost of running the government, but it would support the German economy.
I don't think that many government workers are likely to lose their jobs. There will still be plenty to do, the difference will be that much of their work will no longer be wasted caring for fragile Windows PCs. However, if I worked in IT for the Bundestag I would brush up on my Linux skills right away just in case.
Right this moment, you would probably have to use MS Word format (in the real world). With the newest beta versions of OpenOffice I wouldn't even be too sad about that. I haven't had any trouble with Word documents for some time.
Fairly soon in Germany, however, the format of choice could very well be StarOffice format, and that's why Microsoft is so worried. They know that if the government starts using StarOffice, then companies doing business with the government (ie. nearly everyone) will soon be using StarOffice if for no other reason than to be compatible with the folks at the Bundestag. The only barrier to loading StarOffice on a computer is disk space. StarOffice itself is very inexpensive (and it can be installed on multiple machines), and StarOffice has a GPLed cousin OpenOffice which is Free. Both use the same document formats.
For a while most folks will probably have copies of both MS Word and StarOffice, but come upgrade time the cost of Microsoft Office will almost certainly seem like a big chunk of change. After all, MS Office is good, but if your clients and customers (and the government) are reasonably likely to have a copy of StarOffice, then paying for MS Office is just one more added expense.
The fact that StarOffice was originally written in Germany probably isn't hurting its case either.
If StarOffice takes off, then Microsoft is in big trouble. Not only does StarOffice completely destroy the market for Microsoft's Office cash cow, but it paves the way for alternative OSes like Linux or Solaris. No doubt that's why Sun is giving StarOffice away for free.
Yes, SOAP is slower than Corba. However, we aren't talking about RPC. Or at least I wasn't. I certainly wasn't talking about SOAP vs. Corba. The beauty of Mono is that it will allow you to import packages written in another language but compiled to run on the same virtual machine. It's sort of like compiling Class files written in Jython, except it's available to a host of different languages. No RPC required. I hope that clears things up.
It strikes me as funny that someone that uses the monniker "AnalogBoy" would assume that I am some untrained Linux fanboy.
It also strikes me as funny that someone who uses big-iron Unix can't see the writing on the wall. Linux is being used in mission-critical applications by some very large corporations. The reason that they are using Linux has nothing to do with RMS, the FSF, or Mr. Torvalds and has everything to do with names like IBM, HP, and Compaq (and now Oracle). Big-iron Unix users are finding out that they can migrate their applications from commercial Unix to Linux on commodity Intel-based hardware and "save a bundle."
My advice to you is that if you don't like the Linux zealots on/., hang out with the Linux zealots (wearing suits and ties) at IBM.
As for replacing desktops with StarOffice, I would personally love to replace my Windows desktops with X terminals connected to a Linux server, as this would make my life much easier, but Linux isn't quite there yet.
Linux is perfectly capable of replacing my expensive Unix servers, however, which is what we were talking about.
Larry Ellison's thin client was received less than warmly because it didn't run hardly any software. A couple of years ago the only useful X Windows application was Netscape. That is no longer the case. OpenOffice (the new beta) is a very credible MS Office replacement, Evolution is a credible email client, Mozilla is finally shaping into something useable, and the list goes on and on.
Trust me, the better the Linux desktop gets the more folks are going to decide that it is "good enough" and they are going to switch. The potential cost savings are simply too big to ignore. The fact that Microsoft is now charging companies whether they upgrade or not even removes the one other alternative that corporations had for cost-saving (not upgrading).
I am not saying that it is going to happen overnight, but StarOffice and Linux use are going to grow in the same manner that MS Office replaced WordPerfect and Lotus 1-2-3 and Windows NT replaced Netware.
The virtual machine should be much faster than CORBA, after all, your code will be compiled down to bytecodes. My guess is that it will be slower than C, probably considerably slower, in much the same way that Java (even with the best JIT) is slower for most real-life problems. It will almost certainly be slower than the wealth of well-worn C that makes up the bulk of free software.
The problem with Mono is that when folks think of.NET they immediately think of the negative bits like Passport and Hailstorm. Mono isn't even concerned with those bits. I personally am pretty happy with a combination of Python and C (with XML-RPC for remote procedure calls and Zope as an application server), but I can at least see the potential of Mono.
That's why I said legacy Windows applications. You would have to switch your users to StarOffice and another email package (Bynari has a nice one with all of the features of Exchange).
The Citrix server would be for those documents that you might receive that don't work with Staroffice, or for old Windows applications that don't have a Linux replacement. Outlook, Word, Excel, and PowerPoint would be replaced.
Which is why, right now, switching to Linux on the desktop is a tricky deal. However, I honestly think that the industry is going to move in this direction. Microsoft currently is planning on charging their larger customers for new versions of their software whether they upgrade or not, and StarOffice is getting pretty darn good.
If the economy doesn't pick up soon, companies are going to seriously consider this stuff, and sooner rather than later.
For now I am doing the same sorts of things that you are doing. I am using Linux as as server OS only. File and print with Samba, new development with Python and Zope, PostgreSQL for the database. All of these products interact well with MS Office while still paving the way for future movements to client side Free Software.
Yes, Microsoft constantly breaks SMB, but they can't break it too badly, otherwise their own products can't talk to each other. That makes their customers angry.
Which is why SAMBA mostly works, and it is also why Novell gave up trying to write their own Windows client and now uses SMB. I remember using Novell's client (when they used to have one) and Microsoft broke nearly every time you installed a piece of MS Software.
In other words while SMB is bad, the fact of the matter is that it is ubiquitous. That's why you see lots of Unix servers with Samba installed and almost no Windows PCs with an NFS client installed. NFS may be better, stronger, and faster, but it is not available on the client PCs.
The choice is simple. Free Software hackers can create software that speaks.NET (and gives them access to the millions of Windows clients with.NET installed) or we can go home and cry. Even now most Linux servers spend all of their time serving up content to Windows clients. We can pretend that we are following "standards" but if our pages don't render correctly in IE, then they get changed.
For now, Microsoft controls the client, and the client speaks.NET.
It's funny you should ask that question, as it is one that Miguel has answered over and over again:
Later on we decided to use better methods for encapsulating our APIs, and we started to use CORBA to define interfaces to components. We complemented it with policy and a set of standard GNOME interfaces for easily creating reusable, language independent components, controls and compound documents. This technology is known as Bonobo. Interfaces to Bonobo exist for C, Perl, Python, and Java.
CORBA is good when you define coarse interfaces, and most Bonobo interfaces are coarse. The only problem is that Bonobo/CORBA interfaces are not good for small interfaces. For example, an XML parsing Bonobo/CORBA component would be inefficient compared to a C API.
The beauty of Mono is that it would allow you to call the Perl module that you really like from Python (and vice versa). You could write an XML parser (for example) in C# (or Python or Perl or C, etc.) and call it from whatever language you like because all of these languages would rely on the CLR (common language runtime).
More importantly Miguel knows that the client calls the shots. That's why you see far more Unix boxes running Samba than Windows PCs running an nfs client. As crappy as SMB is, you are at least guaranteed that it is available on every Windows PC. Even the folks at Novell are finally giving up on trying to maintain their own protocol. Nowadays they serve up files using SMB just like the rest of us.
Mono allows Free Software folks to leverage millions of installed clients. Sure, Microsoft will try and break compatibility, but if they try too hard they are likely to upset their customers.
Tell that to Oracle. They are replacing their big Sparc boxes with commodity Intel-based servers running Linux.
If Oracle can do it, you can bet that you can too. In fact, you can bet that if you don't use a combination of commodity Intel-based servers and commodity Free Software to lower your costs your competitor will, and his overhead will be that much lower than yours.
The days where big-iron Unix users could afford to laugh at Linux are over.
Ha, you could probably already do that by simply shutting down the main file server. Your PC would still work, but all your files would be gone.
With the money you would save you could probably afford two Linux application servers. Set up your clients so that half of them point to one server and the other half point to the second server, and configure them to roll over to the alternate on failure. Your clients might lose 15 minutes of work if one of your server class machines failed, but this sort of a setup would almost certainly be more robust than relying on a pile of commodity desktop PCs running Windows.
That's why companies shouldn't be looking at replacing Windows desktops with Linux desktops. They should take a look at replacing Windows desktops with thin clients that connect to a Linux server.
Key Largo has hundreds of these whatsits connected to one commodity Intel-based server. Instead of administering hundreds of PCs, they administer one server, and have a pile of disposable clients. Not only does this drastically reduce client license costs, but it also drastically reduces maintenance costs.
Want to upgrade your client OS? You only have to touch one machine. Want to install a new application? You only have to touch one machine. Want to increase performance for a CPU intensive application? You guessed it, you only have to touch one machine.
Throw in a Citrix box for legacy Windows applications, and you are set.
RedHat has always been smart about understating their position. They didn't even talk about replacing commercial Unix until it became obvious that the industry was headed in that direction. Bob Young was simply pointing out that RedHat isn't dependent on Linux taking off on the desktop.
It would appear the Solaris on Intel has come to the end of the road. Here is the quote from Sun's website.
Please note: Sun is deferring the productization and release of the Solaris 9 OE for Intel IA-32. We will continue to sell and support Intel versions of the Solaris 8 OE. Per normal lifecycle policy it will remain available and supported under normal terms until mid CY2004, and supported under contract for up to five years beyond that date.
At this time, we have discontinued Solaris 8 OE for Intel downloads. While we have discontinued the download program, we have also slashed the price of Solaris 8 OE for Intel media kits by 40% to $45 US (plus S/H).
However, even if Sun wasn't end-of-lifing Solaris for Intel, there are obvious reasons why Oracle can't base their future business plans on the availability of a low-cost version of Solaris on Intel based hardware. The most obvious of these reasons is that Oracle doesn't own Solaris. If Oracle were to start suggesting to their customers that they run Oracle on free copies of Solaris for Intel (instead of Sun's Sparc hardware) then Sun is bound to notice, and they would almost certainly change the license for their Intel version. After all, they can't really let their free Intel version of Solaris cannabalize sales of their Sparc hardware.
Linux, on the other hand, is safe because Oracle has as much control over it as they need. Since Oracle has access to the source code they can easily customize Linux to their particular needs.
Larry is right, it sucks to be in the operating systems business right now. Especially if you are trying to sell a Unix-like operating system (although Microsoft is feeling the pinch as well). Linux on commodity hardware continues to improve at a remarkable pace, and you can't beat the price.
Actually, they will probably only support one or two major distributions of Linux, and they will probably subcontract out to the Linux vendor some of the Linux problems.
This actually makes a lot of sense for Oracle. After all, they want you to spend as little on your hardware and operating system as possible. After all, they are selling a database and applications, not Solaris licenses. If they can cut Sun out of the loop that is billions more in potential profits for them. Their solutions become less expensive (and more competitive) without any loss of profit margin.
The fact of the matter is that the operating system is quickly becoming a commodity. In a few years even Microsoft won't be selling their operating system (that's why they are so desperate to move to a service and support type business).
I am sure that Intel's compiler is better than gcc at crunching numbers, in fact, it's probably signficantly better. But my guess is that for most folks gcc is "good enough" at a much lower price.
The commercial compiler companies have been trying to rid the world of gcc for quite some time now, and yet, come the close of the day it's the commercial companies that are getting closed down or bought out, and it's gcc that continues to gain users.
Now, if Intel were to allow you to use their compiler for free (even without source code) for commercial purposes, then perhaps I would start to worry about the future of gcc. But that isn't going to happen, and gcc will continue to chug right along.
The Itanium is a flop because it isn't what consumers want. Intel's Itanium is basically an expensive 64 bit chip that runs no popular software. Furthermore, the unpopular software that it does run it generally runs slower than if you were to just go and buy some Pentium chips at your local WalMart. Who in their right mind is going to pay a premium price for a chip that only runs beta versions of Windows and Linux? Not only that, but it runs both of those operating systems slowly.
In fact, if Itanium were to take off I would take it as proof positive that the free market system is broken. If Intel's clout, money, and marketing were all that mattered then Itanium would be all the rage, but it isn't. Nor is it likely to be all the rage anytime soon.
Your innovation remarks are another point entirely. Sometimes the market rewards innovators, but only if the innovation is something that people will pay money for. For example, the inventor of the "innovative" new MicroHat (it's a Microwave and fashionable headgear all in one) isn't likely to make billions. Likewise, the Itanium might have an innovative design, but current implementations are almost completely useless. You can run Windows 2000 advanced server on it (slowly), with almost no applications, or you can run Linux on it (also slowly), with a respectable amount of Free Software. Of course, if you are running Linux you have your pick of platforms, and Itanium probably won't be at the top of the list.
Once again, if the free market system were broken, then it wouldn't matter. Intel could simply force us all to migrate to Itanium.
Hopefully Timothy you have more of a platform that Free Software. If that is the case, then tailor your message to the audience. When posting to/. talk about Free Software, when talking to businessmen use phrases like "fiscal responsibility," "lower taxes," or "increased services." When pushed for details then you can mention specifics (like Key Largo and their IT infrastructure). When talking to soccer moms, stress the importance of public education.
There is no way to make Free Software sound bite friendly. Half the of the words that we use have at least somewhat negative connotations with most normal folk. In fact, even the phrase "Free Software" is problematic unless you are talking to an audience that knows what Emacs is. Whatever you do, don't mention the word "hacker."
Slashdot Mirror
Actually, I am one of those ex-Mac users that decided that what I really wanted was Emacs, FVWM, and a bunch of xterms. Screw GUIs (unless I am drawing, of course).
A Macintosh is nothing more than a giant hardware key. Apple doesn't have to worry about installing pirated copies of Mac OS simply because you can't get Mac compatible hardware unless you buy it from them (with a Mac OS license, of course). I am surprised that they don't give the new software away. After all, what better way to sell new Apple hardware than to write software that is to bloated to use on your old hardware. That's why you can guarantee that new versions of OS X aren't going to run any better on G3s than they do now.
Apple is a hardware company, the only reason they write software is to give their hardware something to do.
As far as porting to x96 is concerned, Apple knows that the x86 OS market is getting a little crowded as it is. Between Linux, FreeBSD, and a handful of versions of Windows there is something for everyone. Linux and FreeBSD are having a hard time getting a run and they are available free of charge, come with source, and are actually getting pretty darn useable. Mac OS X wouldn't have a chance. Sure the interface is slick, but it won't run existing Windows software (and it wouldn't run existing Mac software either), it wouldn't support very much hardware (probably less than Linux), and you wouldn't be able to get it preinstalled. Bill and Co. would make sure of that.
Apple has a fairly large market of users who are more than happy to pay premium prices for PC hardware, and they are happy with that. If Apple really wanted to see a Mac OS rennaisance what they need to do is price their hardware so that it is price competitive. Otherwise, no dice.
I used to feel the same way, but now I have a hard time reading physical books. I set up the Weasel reader (formerly GutenPalm) on my Handspring so that it scrolls automatically (screen wrap style) and now I can read for hours without having to move a muscle. The backlight means I can read in bed without waking my wife, and I can carry a wide selection of books around with me all of the time. I can even do keyword searches (try that with a book). The real problem so far has been getting reading material that was worth my time, but between baen.com and the Gutenberg project I have managed to keep myself pretty well stocked.
I am a fairly new convert to the Palm platform; I have only owned my Handspring for a few weeks, and I used to swear that reading on a handheld would never take off. I didn't think that anyone would pay several hundred dollars for a gizmo that would allow them to read e-books. Nowadays, however, you can get an 8M Handspring for less than $100, and being able to read in the dark without waking my wife is priceless.
The butterfly ballots are easy enough to use. The Democrats just realized that they were going to lose (by an amazingly small number of votes) and they hoped to "rectify" things. In cases like that it is always possible to find folks that will come forward and testify that the system was "too hard."
The one thing I like about a punch card system is that it is easy for the voter to verify that that their vote was cast correctly. They simply take the ballot out of the machine and make sure that the holes are in the right places. With computerized terminals it would be too easy for someone to write a one line Perl script that changed every third vote for their oponent into a vote for them. Now, if the computerized touch screen system also generated a physical ticket that the voter could verify (and which would be used for hand counts) then I am all for it. I am all for simplifying the system, but I want to be able to verify that no one is diddling the results.
Probably not, but then again, I don't see how we can stop it. For some applications (say German tax applications) I don't even really see the harm. I think that English will remain the Lingua Franca for Free Software hackers for some time yet, but a certain amount of regionalization is inevitable. I participate in several mailing lists where there is a significant non-English speaking contingent, but English is still what gets posted to the lists (and put in the code). If you want to create a truly global project you don't really have much choice of the language to use. However, that could easily change.
This case is a bit different because the German government is even more of a monopoly than Microsoft. Nearly everyone does business with their government, and if the government asks for documents in StarWriter format, you don't send them a Word document.
That's why Microsoft is up in arms, they know that if the German government starts using (and mandating) StarOffice then they are likely to see sales of the German edition of MS Office plummet. Even if relatively few machines in the Bundestag are converted it could have a totally devastating effect due to the millions of folks that are required to share documents with the Bundestag.
The fact that MS Office is somewhat more featureful than StarOffice won't matter one bit, because if you are dealing with the government the most important feature is using the same Office suite that they are using. The fact that StarOffice is a good product that is available for free (and that was originally made in Germany) doesn't hurt either. If the government were to switch, it probably wouldn't take too long for StarOffice to become the official German Office suite. The only reason that Microsoft keeps mentioning Linux as their number one threat is that they don't want to tip people off about StarOffice. Changing your operating system is hard, and requires you to learn a whole new set of programs. Replacing MS Office with StarOffice is much easier, and it carries almost all of the same financial benefits of a full migration to Linux (Windows is included in the price of nearly all machines MS Office isn't).
We are not talking about maintenance jobs. Linux, properly deployed, should actually lower maintenance costs. We are talking about programming jobs. Linux would allow the local software market to flourish instead of being tied to a handful of companies in the United States. Why should the Germans spend their money on U.S. software when they can get Linux solutions from a local German vendor (at a lower price).
For far less than a Microsoft solution the Bundestag could get a custom solution based on Linux and other open standards. This money they spend will even stay in their own economy (meaning that they will get the bulk of it back in taxes).
Let's face it, the reason that this appears to have actual supporters in the Bundestag is because it makes good political sense. Not only would using Linux lower the cost of running the government, but it would support the German economy.
I don't think that many government workers are likely to lose their jobs. There will still be plenty to do, the difference will be that much of their work will no longer be wasted caring for fragile Windows PCs. However, if I worked in IT for the Bundestag I would brush up on my Linux skills right away just in case.
Right this moment, you would probably have to use MS Word format (in the real world). With the newest beta versions of OpenOffice I wouldn't even be too sad about that. I haven't had any trouble with Word documents for some time.
Fairly soon in Germany, however, the format of choice could very well be StarOffice format, and that's why Microsoft is so worried. They know that if the government starts using StarOffice, then companies doing business with the government (ie. nearly everyone) will soon be using StarOffice if for no other reason than to be compatible with the folks at the Bundestag. The only barrier to loading StarOffice on a computer is disk space. StarOffice itself is very inexpensive (and it can be installed on multiple machines), and StarOffice has a GPLed cousin OpenOffice which is Free. Both use the same document formats.
For a while most folks will probably have copies of both MS Word and StarOffice, but come upgrade time the cost of Microsoft Office will almost certainly seem like a big chunk of change. After all, MS Office is good, but if your clients and customers (and the government) are reasonably likely to have a copy of StarOffice, then paying for MS Office is just one more added expense.
The fact that StarOffice was originally written in Germany probably isn't hurting its case either.
If StarOffice takes off, then Microsoft is in big trouble. Not only does StarOffice completely destroy the market for Microsoft's Office cash cow, but it paves the way for alternative OSes like Linux or Solaris. No doubt that's why Sun is giving StarOffice away for free.
Yes, SOAP is slower than Corba. However, we aren't talking about RPC. Or at least I wasn't. I certainly wasn't talking about SOAP vs. Corba. The beauty of Mono is that it will allow you to import packages written in another language but compiled to run on the same virtual machine. It's sort of like compiling Class files written in Jython, except it's available to a host of different languages. No RPC required. I hope that clears things up.
It strikes me as funny that someone that uses the monniker "AnalogBoy" would assume that I am some untrained Linux fanboy.
It also strikes me as funny that someone who uses big-iron Unix can't see the writing on the wall. Linux is being used in mission-critical applications by some very large corporations. The reason that they are using Linux has nothing to do with RMS, the FSF, or Mr. Torvalds and has everything to do with names like IBM, HP, and Compaq (and now Oracle). Big-iron Unix users are finding out that they can migrate their applications from commercial Unix to Linux on commodity Intel-based hardware and "save a bundle."
My advice to you is that if you don't like the Linux zealots on /., hang out with the Linux zealots (wearing suits and ties) at IBM.
As for replacing desktops with StarOffice, I would personally love to replace my Windows desktops with X terminals connected to a Linux server, as this would make my life much easier, but Linux isn't quite there yet.
Linux is perfectly capable of replacing my expensive Unix servers, however, which is what we were talking about.
Larry Ellison's thin client was received less than warmly because it didn't run hardly any software. A couple of years ago the only useful X Windows application was Netscape. That is no longer the case. OpenOffice (the new beta) is a very credible MS Office replacement, Evolution is a credible email client, Mozilla is finally shaping into something useable, and the list goes on and on.
Trust me, the better the Linux desktop gets the more folks are going to decide that it is "good enough" and they are going to switch. The potential cost savings are simply too big to ignore. The fact that Microsoft is now charging companies whether they upgrade or not even removes the one other alternative that corporations had for cost-saving (not upgrading).
I am not saying that it is going to happen overnight, but StarOffice and Linux use are going to grow in the same manner that MS Office replaced WordPerfect and Lotus 1-2-3 and Windows NT replaced Netware.
The virtual machine should be much faster than CORBA, after all, your code will be compiled down to bytecodes. My guess is that it will be slower than C, probably considerably slower, in much the same way that Java (even with the best JIT) is slower for most real-life problems. It will almost certainly be slower than the wealth of well-worn C that makes up the bulk of free software.
The problem with Mono is that when folks think of .NET they immediately think of the negative bits like Passport and Hailstorm. Mono isn't even concerned with those bits. I personally am pretty happy with a combination of Python and C (with XML-RPC for remote procedure calls and Zope as an application server), but I can at least see the potential of Mono.
That's why I said legacy Windows applications. You would have to switch your users to StarOffice and another email package (Bynari has a nice one with all of the features of Exchange).
The Citrix server would be for those documents that you might receive that don't work with Staroffice, or for old Windows applications that don't have a Linux replacement. Outlook, Word, Excel, and PowerPoint would be replaced.
Which is why, right now, switching to Linux on the desktop is a tricky deal. However, I honestly think that the industry is going to move in this direction. Microsoft currently is planning on charging their larger customers for new versions of their software whether they upgrade or not, and StarOffice is getting pretty darn good.
If the economy doesn't pick up soon, companies are going to seriously consider this stuff, and sooner rather than later.
For now I am doing the same sorts of things that you are doing. I am using Linux as as server OS only. File and print with Samba, new development with Python and Zope, PostgreSQL for the database. All of these products interact well with MS Office while still paving the way for future movements to client side Free Software.
Yes, Microsoft constantly breaks SMB, but they can't break it too badly, otherwise their own products can't talk to each other. That makes their customers angry.
Which is why SAMBA mostly works, and it is also why Novell gave up trying to write their own Windows client and now uses SMB. I remember using Novell's client (when they used to have one) and Microsoft broke nearly every time you installed a piece of MS Software.
In other words while SMB is bad, the fact of the matter is that it is ubiquitous. That's why you see lots of Unix servers with Samba installed and almost no Windows PCs with an NFS client installed. NFS may be better, stronger, and faster, but it is not available on the client PCs.
The choice is simple. Free Software hackers can create software that speaks .NET (and gives them access to the millions of Windows clients with .NET installed) or we can go home and cry. Even now most Linux servers spend all of their time serving up content to Windows clients. We can pretend that we are following "standards" but if our pages don't render correctly in IE, then they get changed.
For now, Microsoft controls the client, and the client speaks .NET.
It's funny you should ask that question, as it is one that Miguel has answered over and over again:
The beauty of Mono is that it would allow you to call the Perl module that you really like from Python (and vice versa). You could write an XML parser (for example) in C# (or Python or Perl or C, etc.) and call it from whatever language you like because all of these languages would rely on the CLR (common language runtime).
More importantly Miguel knows that the client calls the shots. That's why you see far more Unix boxes running Samba than Windows PCs running an nfs client. As crappy as SMB is, you are at least guaranteed that it is available on every Windows PC. Even the folks at Novell are finally giving up on trying to maintain their own protocol. Nowadays they serve up files using SMB just like the rest of us.
Mono allows Free Software folks to leverage millions of installed clients. Sure, Microsoft will try and break compatibility, but if they try too hard they are likely to upset their customers.
Tell that to Oracle. They are replacing their big Sparc boxes with commodity Intel-based servers running Linux.
If Oracle can do it, you can bet that you can too. In fact, you can bet that if you don't use a combination of commodity Intel-based servers and commodity Free Software to lower your costs your competitor will, and his overhead will be that much lower than yours.
The days where big-iron Unix users could afford to laugh at Linux are over.
Err, oops. Yeah, I knew that :).
Ha, you could probably already do that by simply shutting down the main file server. Your PC would still work, but all your files would be gone.
With the money you would save you could probably afford two Linux application servers. Set up your clients so that half of them point to one server and the other half point to the second server, and configure them to roll over to the alternate on failure. Your clients might lose 15 minutes of work if one of your server class machines failed, but this sort of a setup would almost certainly be more robust than relying on a pile of commodity desktop PCs running Windows.
That's why companies shouldn't be looking at replacing Windows desktops with Linux desktops. They should take a look at replacing Windows desktops with thin clients that connect to a Linux server.
Key Largo has hundreds of these whatsits connected to one commodity Intel-based server. Instead of administering hundreds of PCs, they administer one server, and have a pile of disposable clients. Not only does this drastically reduce client license costs, but it also drastically reduces maintenance costs.
Want to upgrade your client OS? You only have to touch one machine. Want to install a new application? You only have to touch one machine. Want to increase performance for a CPU intensive application? You guessed it, you only have to touch one machine.
Throw in a Citrix box for legacy Windows applications, and you are set.
RedHat has always been smart about understating their position. They didn't even talk about replacing commercial Unix until it became obvious that the industry was headed in that direction. Bob Young was simply pointing out that RedHat isn't dependent on Linux taking off on the desktop.
It would appear the Solaris on Intel has come to the end of the road. Here is the quote from Sun's website.
However, even if Sun wasn't end-of-lifing Solaris for Intel, there are obvious reasons why Oracle can't base their future business plans on the availability of a low-cost version of Solaris on Intel based hardware. The most obvious of these reasons is that Oracle doesn't own Solaris. If Oracle were to start suggesting to their customers that they run Oracle on free copies of Solaris for Intel (instead of Sun's Sparc hardware) then Sun is bound to notice, and they would almost certainly change the license for their Intel version. After all, they can't really let their free Intel version of Solaris cannabalize sales of their Sparc hardware.
Linux, on the other hand, is safe because Oracle has as much control over it as they need. Since Oracle has access to the source code they can easily customize Linux to their particular needs.
Larry is right, it sucks to be in the operating systems business right now. Especially if you are trying to sell a Unix-like operating system (although Microsoft is feeling the pinch as well). Linux on commodity hardware continues to improve at a remarkable pace, and you can't beat the price.
Actually, they will probably only support one or two major distributions of Linux, and they will probably subcontract out to the Linux vendor some of the Linux problems.
This actually makes a lot of sense for Oracle. After all, they want you to spend as little on your hardware and operating system as possible. After all, they are selling a database and applications, not Solaris licenses. If they can cut Sun out of the loop that is billions more in potential profits for them. Their solutions become less expensive (and more competitive) without any loss of profit margin.
The fact of the matter is that the operating system is quickly becoming a commodity. In a few years even Microsoft won't be selling their operating system (that's why they are so desperate to move to a service and support type business).
I am sure that Intel's compiler is better than gcc at crunching numbers, in fact, it's probably signficantly better. But my guess is that for most folks gcc is "good enough" at a much lower price.
The commercial compiler companies have been trying to rid the world of gcc for quite some time now, and yet, come the close of the day it's the commercial companies that are getting closed down or bought out, and it's gcc that continues to gain users.
Now, if Intel were to allow you to use their compiler for free (even without source code) for commercial purposes, then perhaps I would start to worry about the future of gcc. But that isn't going to happen, and gcc will continue to chug right along.
The Itanium is a flop because it isn't what consumers want. Intel's Itanium is basically an expensive 64 bit chip that runs no popular software. Furthermore, the unpopular software that it does run it generally runs slower than if you were to just go and buy some Pentium chips at your local WalMart. Who in their right mind is going to pay a premium price for a chip that only runs beta versions of Windows and Linux? Not only that, but it runs both of those operating systems slowly.
In fact, if Itanium were to take off I would take it as proof positive that the free market system is broken. If Intel's clout, money, and marketing were all that mattered then Itanium would be all the rage, but it isn't. Nor is it likely to be all the rage anytime soon.
Your innovation remarks are another point entirely. Sometimes the market rewards innovators, but only if the innovation is something that people will pay money for. For example, the inventor of the "innovative" new MicroHat (it's a Microwave and fashionable headgear all in one) isn't likely to make billions. Likewise, the Itanium might have an innovative design, but current implementations are almost completely useless. You can run Windows 2000 advanced server on it (slowly), with almost no applications, or you can run Linux on it (also slowly), with a respectable amount of Free Software. Of course, if you are running Linux you have your pick of platforms, and Itanium probably won't be at the top of the list.
Once again, if the free market system were broken, then it wouldn't matter. Intel could simply force us all to migrate to Itanium.
Hopefully Timothy you have more of a platform that Free Software. If that is the case, then tailor your message to the audience. When posting to /. talk about Free Software, when talking to businessmen use phrases like "fiscal responsibility," "lower taxes," or "increased services." When pushed for details then you can mention specifics (like Key Largo and their IT infrastructure). When talking to soccer moms, stress the importance of public education.
There is no way to make Free Software sound bite friendly. Half the of the words that we use have at least somewhat negative connotations with most normal folk. In fact, even the phrase "Free Software" is problematic unless you are talking to an audience that knows what Emacs is. Whatever you do, don't mention the word "hacker."
Good Luck.
Did anyone think to use crayons?