I agree with you, don't add more stuff, just add more flexibility and let people go where they will with it. The more you add, the more people there are who will have to tear that out or change it to do what the want to do with it (i.e. that thing you never considered it could be used for). I do want to disagree with the poster about ethernet. Don't add it to the product, make it so others can add it IF they need it and add exactly what they need. Slap in a couple of PCMCIA slots and a better hard drive interface and leave it alone otherwise.
>There are performance issues, however. But using Java for a front-end, and c++ for the back-end >can result in a quick application, with cross-platform, consistant UI.
The thing about a statement like this is that it is so often repeated that it has become something that people generally accept as fact without asking if it is actually true. Speaking as an experienced Java programmer I can tell you that it is not true. Java is not a slow language. It used to be slow due to design problems back in the 1.0 days and due to slow VMs, but it is not slow now. Some people will say that it is inherently slower than C++ due to the garbage collection and everything I've read says they're right. The real issue is how much slower. The answer: not much. If a program is written using a modern version of Java and runs on a reasonably up to date VM, it will not be slow due to the performance of the language itself. If it is slow, it is the fault of the programmer. One of the tradeoffs of Javas ease of use, is that it makes it disgustingly easy to write really lousy code. I see it all the time, sloppy programmers who don't really understand what the language and VM are actually doing with the code they write, will needlessly write their program to be an order of magnitude slower and more resource hungry. Beginner programmers assume that just because the code compiles and runs, that the work stops here. Languages like C++ are inherently more complex to develop in, meaning that you almost have to have a better grasp of what your code actually does to get them to work. Java is far easier to learn and use, so some of the developers it gets are not necessarily the best programmers in the business (VB works like this too). The easiest way to tell an experienced Java programmer who knows their code from somebody who can simply reproduce the syntax is to look at how they handle their Strings. Try it sometime, there is a big difference.
>Don't be so reactionary, I'm not saying Be is "evil", just not as good as Linux
Assuming the above was not a troll, I'd be interested to hear what your support is for that statement. What do you see about Linux as being better than BeOS specifically?
There are more than two states here. Why does this issue always have to end up as an argument between the 'forget the design crap lets just pound out the code' bunch and the 'abstract design concepts are always more important than trivial issues of implementation' crowd? This is not an either/or thing. Both perspectives are wrong if they fail to admit that both are important. Teaching abstract design to someone who has only the vaguest experience coding is usually a waste of time because they can't relate what you're teaching to anything and consequently won't TRULY undestand (or likely even remember) what they are being taught. On the other side, simply diving into development without a good design leads to frequent rewrites to account for things you didn't think of the first time around and programs that are not easily maintainable or extendable. Sure, you can write software this way but 9 times out of 10 it won't be even close to the quality you could have gotten simply by thinking things through first.
Frankly I'm getting tired of the smug certainty that focusing on the abstract is the way schools should go because it's the 'Right' way to do things. Schools like that produce people who can memorize stuff they don't understand, long enough to pass the tests and get their degrees. Then they get sent to people like me in the real world to re-teach them what the schools should have taught in the order they should have taught it. And I'm also tired of having to rewrite the crap that the 'code first and ask questions later' people produce. If it has to be done over then the time it took to plan first might not have been wasted after all hmmmm?
I think everyone would benefit from a little dose of reality on this whole issue. How about it?
I have seen the same mentality you've just expressed running rampant in universities lately. It sounds great in theory, but in practice it results in students who are nearly useless as programmers. I won't argue the theoretical merits of teaching higher level concepts to students, because in theory, I agree with you. Unfortunately the idea of teaching higher level concepts and abstract issues of design to a student first, does not work all that well.
Why? Because they have nothing to relate this new knowledge to. Again and again I've seen students walking out of classes taught the way you mention completely confused. Never having seen how design is applied in the real world, they fail to truly 'get' it and fall back on wrote memorization to learn enough to simply pass the next test. Once the class is over, they promptly forget what they 'learned', because this knowledge never really made much sense.
The people who really seem to make use of the more abstract programming concepts, are the ones for whom basic issues of programming syntax and logic are no longer a mystery. These people can comfortably transition from programmer to designer, because they now have a context in which to apply the abstract concepts they will be taught.
Here is a perfect example of the potential of free and open source software. It can allow our schools to freedom to encourage the next generation of technical people. As the article stated, it is a great fit for schools because it is free, which means that a school can afford to stay with the current technology (meaning that the students aren't learning stuff that is 5 years old and useless in the real world). There is only one problem with this scenario, which is part of the reason the school mentioned in the article is the exception rather than the rule, most schools don't have anyone to guide them down this path.
This is where all of you come in. You have a chance to influence the next generation of techies to use, support, and possibly even contribute code to the software you love. Volunteer your time at your local school. Help them setup and maintain their networks. Offer to speak on technology issues in classes and clubs. Make yourself available to mentor students with an interest in technology.
Schools won't use open source and free software unless someone shows them that it is an option and explains why it is a good option, be that person.
I could be missing something here, but it seems to me that a tradeoff is being made. You are getting better compression of your data at the potential cost of reliability. Multiple instances of data = redundancy. If you have only one copy of something and for some reason something goes wrong with it, you have a problem everywhere it is needed. To some extent this parallels problems that the windows registry gives us. With the registry, you have a single point of failure, if it gets corrupted, very bad things start happening. Another example (as others here have pointed out) is DLLs. Multiple programs use the same DLLs. If they disagree over what version the DLL should be or if the DLL goes missing, you can end up with multiple simultaneous problems. I'm not saying this whole idea is a bad one or that it definitely cannot be made to work well, just that there is some risk here which makes the quality of the implementation very critical.
I feel obligated to point out that this was not an article about Linux. It was an article about open source. They themselves made a good effort to express the two as not being equivalent. Linux is a small subset of open source. Taking this article as them being "pro-Linux" is a little like taking the comment "GUIs can be very useful" as an endorsement of Windows.
I agree with all your points but I think you left out or failed to emphasize the most important one.
Microsoft uses Office (and to a lesser extent some of the other apps) to continually reinforce its grip on the OS market. Right now I sincerely doubt they would ever consider porting it to an operating system that might threaten theirs (Linux or BeOS)(MacOS is not a threat because it does not run on X86). If the applications division of the company is split from the OS portion, ports become a lot more likely. Any platform getting a port of something like Office will find itself getting some real respect as a windows alternative. Splitting the company may not be a perfect solution, but combined with the momentum that alternative OSs are beginning to build, it might be enough.
When I read comments like the ones Dvorak made on Java I really have to wonder about his accuracy in other areas. He seems to equate Javas obvious failures with the failure of Java itself. Applets, while they have some niche applications, were a flop. JavaOS and the Java based NC were also flops. Why? Because the idea was to force a technology into an area where it really did not fit very well. What does this have to do with Java itself failing? Very little really.
Java is very good for some things and its use is exploding in those areas. The reason most people seem to think it has failed is that these areas fall into an end user blind spot. People could see applets (and later the lack thereof), but the areas where Java is now primarily being used are not nearly as visible to your average person. Java is now being used for middleware. It lives on the server side these days in the form of servlets and beans. Just because you don't see it does not mean it is not there. The number of Java programmers and the use of Java in development is swiftly growing.
As for Jini, it is still relatively new and its fate has yet to be determined. For my part I suspect it will catch on, but not in its incarnation as an interface to hardware devices (despite the hype, this is not its only use).
Ultimately I think it would be far more accurate to say that Sun was a flop this year. The technology they created is prospering, the companies standing in the industry is not.
As an aside, if you are a Java programmer and are looking for work (in NYC), the company I work for is now hiring. If you're interested send me a resume at: jkeck@finansys.com
I've never been a believer in conspiracy theories and I don't put too much stock in the one I'm about to mention, but it does make one think. AMD has a big new chip plant in Germany that the German government has been a strong supporter of, if I remember right. If Germany is the one leading the charge on this PIII ban, it does make you wonder if there is some sort of connection.
It is not really a question of 'Is Java being used?' it is a question of 'Where is Java being used?'. You are right in that there are not a huge number of end user Gui Java apps out there supplanting native versions of software. Cross platform issues and past problems with the languages architecture & speed (which have since mostly been corrected) kept this from happening. However there are areas where Java is very well suited to the tasking and is rapidly taking over. The article mentioned one area but only caught the tip of the iceberg. Middleware. Java is rapidly taking over the middleware arena. This is an area which is not very visible to the average person but is a huge market with a lot of developers. Another place Java is really catching on is the server side of the web. How many of the dynamic pages you see were generated by a servlet, and if they were could you tell?
As an aside, I'd like to comment on the issue of speed problems. If you think that Java must be slow, you are mistaken. Java 1.0 was slow. Java 1.1 and 1.2 do not have to be. My experience has been that Java makes it easy to write slow programs, but that a properly architected program written by someone who understands how Java functions executes very fast. There are also options like native compilation, Hotspot, JIT compilers, and modern VMs which combined with a properly written program will result in very good performance. No, Java is not fundamentally as fast as something like C or Fortran, but the idea that it is too slow to be useful is obsolete.
...its just that lately it seems to be a bit scarce at ZD publications. I regularly read a variety of ZD publications and have lately come to feel that they are of questionable value for anything but the broadest news and announcements. Anchordesk in particular seems to be taking its cue from the tabloids and doing its best to bring this...creative... angle of news reporting to the tech world. More and more I am leaning towards sources like Slashdot because they are self correcting. By their very nature, they tend to adjust towards reality in the information they present. Since the content is representative of a large group of opinions, you get not only a feeling for what the average opinion is, but also what the counter points are. It's not perfect, it is prone to groupthink and filtration in some cases, but I feel that it provides a better source of news and views then what we are seeing from many of the more traditional mediums. At least with the Slashdot model, the people providing the bulk of the information (the Slashdot community) are separate from the people with a financial stake in having more hits on the site (Andover.net). IMHO this makes a big difference.
Just at a quick glance I can't see any fundamental shifts in who has access to information here. Those who can invade your privacy now, will still be able to do so, and those who can't now probably still won't be able to. The only real change I see here is in ease of use for end users and possibly fewer headaches for sysadmins (depending on what they want to accomplish). No more messing around with address configuration in software.
Slashdot Mirror
I agree with you, don't add more stuff, just add more flexibility and let people go where they will with it. The more you add, the more people there are who will have to tear that out or change it to do what the want to do with it (i.e. that thing you never considered it could be used for). I do want to disagree with the poster about ethernet. Don't add it to the product, make it so others can add it IF they need it and add exactly what they need. Slap in a couple of PCMCIA slots and a better hard drive interface and leave it alone otherwise.
>There are performance issues, however. But using Java for a front-end, and c++ for the back-end
>can result in a quick application, with cross-platform, consistant UI.
The thing about a statement like this is that it is so often repeated that it has become something that people generally accept as fact without asking if it is actually true. Speaking as an experienced Java programmer I can tell you that it is not true. Java is not a slow language. It used to be slow due to design problems back in the 1.0 days and due to slow VMs, but it is not slow now. Some people will say that it is inherently slower than C++ due to the garbage collection and everything I've read says they're right. The real issue is how much slower. The answer: not much. If a program is written using a modern version of Java and runs on a reasonably up to date VM, it will not be slow due to the performance of the language itself. If it is slow, it is the fault of the programmer. One of the tradeoffs of Javas ease of use, is that it makes it disgustingly easy to write really lousy code. I see it all the time, sloppy programmers who don't really understand what the language and VM are actually doing with the code they write, will needlessly write their program to be an order of magnitude slower and more resource hungry. Beginner programmers assume that just because the code compiles and runs, that the work stops here. Languages like C++ are inherently more complex to develop in, meaning that you almost have to have a better grasp of what your code actually does to get them to work. Java is far easier to learn and use, so some of the developers it gets are not necessarily the best programmers in the business (VB works like this too). The easiest way to tell an experienced Java programmer who knows their code from somebody who can simply reproduce the syntax is to look at how they handle their Strings. Try it sometime, there is a big difference.
>Don't be so reactionary, I'm not saying Be is "evil", just not as good as Linux
Assuming the above was not a troll, I'd be interested to hear what your support is for that statement. What do you see about Linux as being better than BeOS specifically?
http://209.15.80.234/bekarma/mirror.html
There are more than two states here. Why does this issue always have to end up as an argument between the 'forget the design crap lets just pound out the code' bunch and the 'abstract design concepts are always more important than trivial issues of implementation' crowd? This is not an either/or thing. Both perspectives are wrong if they fail to admit that both are important. Teaching abstract design to someone who has only the vaguest experience coding is usually a waste of time because they can't relate what you're teaching to anything and consequently won't TRULY undestand (or likely even remember) what they are being taught. On the other side, simply diving into development without a good design leads to frequent rewrites to account for things you didn't think of the first time around and programs that are not easily maintainable or extendable. Sure, you can write software this way but 9 times out of 10 it won't be even close to the quality you could have gotten simply by thinking things through first.
Frankly I'm getting tired of the smug certainty that focusing on the abstract is the way schools should go because it's the 'Right' way to do things. Schools like that produce people who can memorize stuff they don't understand, long enough to pass the tests and get their degrees. Then they get sent to people like me in the real world to re-teach them what the schools should have taught in the order they should have taught it. And I'm also tired of having to rewrite the crap that the 'code first and ask questions later' people produce. If it has to be done over then the time it took to plan first might not have been wasted after all hmmmm?
I think everyone would benefit from a little dose of reality on this whole issue. How about it?
I have seen the same mentality you've just expressed running rampant in universities lately. It sounds great in theory, but in practice it results in students who are nearly useless as programmers. I won't argue the theoretical merits of teaching higher level concepts to students, because in theory, I agree with you. Unfortunately the idea of teaching higher level concepts and abstract issues of design to a student first, does not work all that well.
Why? Because they have nothing to relate this new knowledge to. Again and again I've seen students walking out of classes taught the way you mention completely confused. Never having seen how design is applied in the real world, they fail to truly 'get' it and fall back on wrote memorization to learn enough to simply pass the next test. Once the class is over, they promptly forget what they 'learned', because this knowledge never really made much sense.
The people who really seem to make use of the more abstract programming concepts, are the ones for whom basic issues of programming syntax and logic are no longer a mystery. These people can comfortably transition from programmer to designer, because they now have a context in which to apply the abstract concepts they will be taught.
Here is a perfect example of the potential of free and open source software. It can allow our schools to freedom to encourage the next generation of technical people. As the article stated, it is a great fit for schools because it is free, which means that a school can afford to stay with the current technology (meaning that the students aren't learning stuff that is 5 years old and useless in the real world). There is only one problem with this scenario, which is part of the reason the school mentioned in the article is the exception rather than the rule, most schools don't have anyone to guide them down this path.
This is where all of you come in. You have a chance to influence the next generation of techies to use, support, and possibly even contribute code to the software you love. Volunteer your time at your local school. Help them setup and maintain their networks. Offer to speak on technology issues in classes and clubs. Make yourself available to mentor students with an interest in technology.
Schools won't use open source and free software unless someone shows them that it is an option and explains why it is a good option, be that person.
I could be missing something here, but it seems to me that a tradeoff is being made. You are getting better compression of your data at the potential cost of reliability. Multiple instances of data = redundancy. If you have only one copy of something and for some reason something goes wrong with it, you have a problem everywhere it is needed. To some extent this parallels problems that the windows registry gives us. With the registry, you have a single point of failure, if it gets corrupted, very bad things start happening. Another example (as others here have pointed out) is DLLs. Multiple programs use the same DLLs. If they disagree over what version the DLL should be or if the DLL goes missing, you can end up with multiple simultaneous problems. I'm not saying this whole idea is a bad one or that it definitely cannot be made to work well, just that there is some risk here which makes the quality of the implementation very critical.
Look at the display made by Microvision.
http://www.mvis.com/
This is what will remove the biggest limitations
of handheld devices. Low power, compact, high
quality displays.
I feel obligated to point out that this was not an article about Linux. It was an article about open source. They themselves made a good effort to express the two as not being equivalent. Linux is a small subset of open source. Taking this article as them being "pro-Linux" is a little like taking the comment "GUIs can be very useful" as an endorsement of Windows.
I agree with all your points but I think you left out or failed to emphasize the most important one.
Microsoft uses Office (and to a lesser extent some of the other apps) to continually reinforce its grip on the OS market. Right now I sincerely doubt they would ever consider porting it to an operating system that might threaten theirs (Linux or BeOS)(MacOS is not a threat because it does not run on X86). If the applications division of the company is split from the OS portion, ports become a lot more likely. Any platform getting a port of something like Office will find itself getting some real respect as a windows alternative. Splitting the company may not be a perfect solution, but combined with the momentum that alternative OSs are beginning to build, it might be enough.
When I read comments like the ones Dvorak made on Java I really have to wonder about his accuracy in other areas. He seems to equate Javas obvious failures with the failure of Java itself. Applets, while they have some niche applications, were a flop. JavaOS and the Java based NC were also flops. Why? Because the idea was to force a technology into an area where it really did not fit very well. What does this have to do with Java itself failing? Very little really.
Java is very good for some things and its use is exploding in those areas. The reason most people seem to think it has failed is that these areas fall into an end user blind spot. People could see applets (and later the lack thereof), but the areas where Java is now primarily being used are not nearly as visible to your average person. Java is now being used for middleware. It lives on the server side these days in the form of servlets and beans. Just because you don't see it does not mean it is not there. The number of Java programmers and the use of Java in development is swiftly growing.
As for Jini, it is still relatively new and its fate has yet to be determined. For my part I suspect it will catch on, but not in its incarnation as an interface to hardware devices (despite the hype, this is not its only use).
Ultimately I think it would be far more accurate to say that Sun was a flop this year. The technology they created is prospering, the companies standing in the industry is not.
As an aside, if you are a Java programmer and are looking for work (in NYC), the company I work for is now hiring. If you're interested send me a resume at: jkeck@finansys.com
I've never been a believer in conspiracy theories and I don't put too much stock in the one I'm about to mention, but it does make one think. AMD has a big new chip plant in Germany that the German government has been a strong supporter of, if I remember right. If Germany is the one leading the charge on this PIII ban, it does make you wonder if there is some sort of connection.
>I love Java, but I'm not a completly fanatical idiot about it. Right now, you simply couldn't implement the low-level parts of a decent OS in Java.
Does this mean that you do not feel that JBED can work well? (See link for details)
JBED info here
It is not really a question of 'Is Java being used?' it is a question of 'Where is Java being used?'. You are right in that there are not a huge number of end user Gui Java apps out there supplanting native versions of software. Cross platform issues and past problems with the languages architecture & speed (which have since mostly been corrected) kept this from happening. However there are areas where Java is very well suited to the tasking and is rapidly taking over. The article mentioned one area but only caught the tip of the iceberg. Middleware. Java is rapidly taking over the middleware arena. This is an area which is not very visible to the average person but is a huge market with a lot of developers. Another place Java is really catching on is the server side of the web. How many of the dynamic pages you see were generated by a servlet, and if they were could you tell?
As an aside, I'd like to comment on the issue of speed problems. If you think that Java must be slow, you are mistaken. Java 1.0 was slow. Java 1.1 and 1.2 do not have to be. My experience has been that Java makes it easy to write slow programs, but that a properly architected program written by someone who understands how Java functions executes very fast. There are also options like native compilation, Hotspot, JIT compilers, and modern VMs which combined with a properly written program will result in very good performance. No, Java is not fundamentally as fast as something like C or Fortran, but the idea that it is too slow to be useful is obsolete.
...its just that lately it seems to be a bit
scarce at ZD publications. I regularly read
a variety of ZD publications and have lately
come to feel that they are of questionable
value for anything but the broadest news and
announcements. Anchordesk in particular seems
to be taking its cue from the tabloids and
doing its best to bring this...creative...
angle of news reporting to the tech world. More
and more I am leaning towards sources like
Slashdot because they are self correcting. By
their very nature, they tend to adjust towards
reality in the information they present. Since
the content is representative of a large group
of opinions, you get not only a feeling for what
the average opinion is, but also what the counter
points are. It's not perfect, it is prone to
groupthink and filtration in some cases, but I
feel that it provides a better source of news and
views then what we are seeing from many of the
more traditional mediums. At least with the
Slashdot model, the people providing the bulk
of the information (the Slashdot community) are
separate from the people with a financial stake
in having more hits on the site (Andover.net).
IMHO this makes a big difference.
Just at a quick glance I can't see any fundamental shifts in who has access to information here. Those who can invade your privacy now, will still be able to do so, and those who can't now probably still won't be able to. The only real change I see here is in ease of use for end users and possibly fewer headaches for sysadmins (depending on what they want to accomplish). No more messing around with address configuration in software.