How can people honestly believe that a modern government could harbour ANY kind of conspiracy...
Tell that to the people behind the SR-71, the F-117, and the B-2. If the government wants to keep something a secret until they want to reveal it, they will do it.
And these were aircraft which obviously had to fly out in the open, but the general public was very successfully kept in the dark. Now, by extension, how much easier is it to keep unseen things a secret?
Re:Legislation goes a bit too far
on
Mega-Geek March?
·
· Score: 2
It smacks of 'affirmative action' programs and admission standards...
Openness of government is essential, except, perhaps, where the government is "classified". Acheiving this doesn't require Affirmative Action, which has been and always will be a debatable practice.
A good compromise would be for government customers to begin demanding that their software use open file formats and protocols at a minimum. If a closed-source software package is so good, then opening its file format won't destroy its user base, right? Opening up formats and protocols will go very far in leveling the playing field between all the different options, and everyone that uses software will benefit, regardless of Open Source vs. Closed Source.
Yes, most people don't use the features. But if you actually do use one of the features, good luck trying to replicate it in a free software package.
This is a very good reason for people to not use MS Office. Why should a corporation spend tens of thousands of dollars on features only ten people in the company use?
When feature utilization reaches this level in normal software, that software appeals to a niche, while many other people find cheaper alternatives.
For example: Pro/ENGINEER vs. Autocad, Photoshop vs. xv or GIMP, Sun cc vs. GCC, MS Office vs. OpenOffice. There are many more.
It is a basic fact that MS Office has gone out of control in a number of ways. For those people who can live with this, they can be happy with their proprietary formats and interfaces doing what they need to do best. Everyone else can find immense simplicity and flexibility in OpenOffice's good functionality and open file formats.
As long as the other free office suites also follow the model of open file formats, interoperability will be the rule rather than an exception, and MS Office will gradually find itself falling out of fashion. For office productivity, this would be a very very good thing.
My mom is a little out of the loop on the whole computer thing and she bought a PowerBook G4. So far she has broken the computer twice.
Perhaps she should stick to a cheaper computer; the iBook is much more reasonable in cost.
Just right there in less than a year is $700 for my silly mom if she wouldn't have been on warranty. When her warranty expires she will be buying Applecare.
This is another sign that the Powerbook wasn't a good match for your mother. If a product needs to be protected from its user, then something really is amiss.
The right product will: meet the user's needs and not be so expensive as to require financing nor extended warranties. This results in a low-risk purchase that simply works better in the long run. The question should be: "Do I really need a full-blown PowerBook at greater than $2500, or can I get by just fine without spending as much money? The answer can be suprising.
A laptop *is* likely to break at some point over a three-year span either from mechanical failure or (more likely) by human accident, especially if you are in a college environment.
Any object is likely to break if the owner is careless, and default warranties cover nearly all cases where the object breaks due to carelessness on the part of the manufacturer. If the object needs to be protected from its owner, perhaps a more robust version of that object is needed (like an armored laptop).
At a college, it is arguable that the laptop will be stolen before it breaks, which is covered under normal household insurance policies. If it breaks due to someone else's carlessness, they are responsible for the cost of repairing it.
I just generally consider the cost of something, even a computer, as a risk once the default warranty expires. If it breaks, I replace it, if not, I use it as long as it is practical. This works as long as I make purchases within my ability to spend (financing small purchases also makes no sense).
The applecare warranty is a good value.
Only if it provides something above and beyond just warranty service. If it includes hand-holding tech support for learning how to use the computer, it may be worth the cost. But, if it is just a warranty and nothing more, it isn't worth spending a sizable fraction of the purchase price for it.
You're foolish not the get the applecare warranty.
Extendend warranties are a form of insurance. Insurance's purpose in life is to save us from financial catastrophes, such as a tree crushing a home or a major illness. In these cases, insurance is a wise purchase.
However, purchasing insurance on small things, such as a PC or a dishwasher is not a wise investment. For small purchases, the insurance is more expensive than a potential repair, because things just don't break frequently. It is more likely that you will never use the insurance, which is why companies enjoy selling it so much. Add up all the money saved by not buying small-time insurance, and that occasional repair is simply not a big deal, even if it is $350.
It is better to skip the insurance and spend the money on repairs when they are needed. In the long run, you will come out better off. An added benefit is that you maintain only a small number of insurances policies to track, which is helpful given their nitpicking complexity.
...these postings feel more like a lynch gang than an informed discussion.
Lynch gangs usually act irrationally. The better Slashdot posts on the subject of Passport, however, are far from irrational; in fact, they are very objective.
The reason: We understand that Passport is a single point of failure, and its design is fundamentally flawed. It inherently does nothing to protect the information housed within it, and only the most naive individual would continue to believe otherwise in the face of the truth.
I will never sign up for Passport voluntarily, and I will actively seek to do business with Microsoft's competitors in this market. If the WWW decomposes to the point of working only in the context of Passport, then this will surely be a sign of the failure of the WWW. I wholeheartedly hope this will never happen, as the WWW, so far, has been one of the greatest resources in the history of Mankind.
We must all work hard to prevent Microsoft from spoiling the WWW further than they already have. This goes beyond resisting Passport, as the potential of Palladium and other overblown DRM schemes is just as dangerous. Already, Microsoft has done damage using an illegally-gained market share for Internet Explorer.
Why are the general public and the U.S. government allowing them to continue unchecked?
This question baffles and sickens me, as the only answers, it seems, are based in the psychology of cults and organized crime.
This is, of course, opposed to the self-serving testimony of the RIAA.
This is why our only hope is that some of the regulators are shamelessly cynical. If even one person in a powerful position sees just how transparent the media industry's motives are, then, at least, there is a chance for a positive outcome.
On the other hand, what is it they say about hope?
Given that you read something like 1/3 more slowly off a video screen opposed to a paper book, this would suck pretty big-time.
I disagree with this, since video screens have adjustable font sizes and contrast. I can adjust my monitor, for example, to compensate for the inaccuracies in my glasses prescription. On a smaller PDA screen, this can still be done, but with some more limitations. Speed readers, on the other hand, would probably hate video screens, just because they read so damn fast.
Besides, nothing (yet) can replace a book you own, a highlighter and a pen for making notes in the margin and taking the whole thing to a study lounge to get away from your roommate's beer party (the sacrifices we make for good grades).
I absolutely agree with this. One thing I like to do is take a book, find the errata for it, and manually correct the whole thing. Errata in text books is a real PITA, and some books are so bad that they can't be trusted until they are corrected. Would e-books allow correction on-the-fly under the DMCA and/or Palladium-type schemes? I know that it would be possible to upgrade to book version 1.01, but making notes and corrections in real time is pretty much a neccessity.
And then there won't be any supported stable platform for running Windows apps, except WINE and related projects.
This is very true. Even though Open Source software is often accused of volatility, it has an inherent stability as well, because the motives driving it are different. As projects like WINE mature, I would bet that whole consulting businesses develop around supporting Windows applications that, ironically, no longer work in Windows.
A small company (or just a well-run, tight margins, efficient company) that only has the time to develop for one platform will choose windows...
I really think that the "Windows only, because it is popular" arguments are pretty weak now-a-days, because cross-platform development is getting easier all the time. Sure, someone can grab at that first 80 or 90% initially, but why not also tap into that remaining customer base? This also has the nice side-effect of protecting your company from Microsoft's future (whatever that will be) by betting on Microsoft, Apple, and Linux/UNIX simultaneously.
Another added benefit of keeping several platforms in mind at one time, is that Microsoft's proprietary development tools take a proper place among everything else. For example, Visual Studio is simply a modular part of the work flow (compiler & linker only; no binary "makefiles"). This can really help when some tools need to be changed out for others without crippling the project.
Cross-platform need not be gut-wrenchingly hard, either. I work on a UNIX application that's been under development for over a decade. However, it is modular enough and uses APIs wisely enough, that a Windows version isn't out of reach. This could be true, also, of well-planned Windows applications.
I think most of the current arguments for Windows-only applications stem from either short-sightedness, arrogance, or stubbornness. Historically, this wasn't the case, but today things are just different.
I'm tempted to ask how long you've been in the industry to say this.
The specific system I had in mind when writing this was an old HP3000 minicomputer from Way Back In The Day. The 500MB disk drives weighed in at over 100lbs. each.
Okay, this makes sense. My experience is the SCSI-2 and later era, where pretty much everything plays nicely. It's even possible now to put 36GB hard drives into SPARCStations, although I'm not sure how many people would do that. Also, the secondary market for old Sun hardware is pretty healthy. I suppose, perhaps another five or ten years from now, that even SPARCStations will become unmaintainable, which is about where that HP 3000 system got to.
So, the argument changes quite a bit depending on whether the hardware is just old or is really old.
We have (had) about 4500 Sun boxen (Servers, workstations, SunRays, etc) ranging from old Ultra-1s to E-10k.
With that much hardware, what do you expect? That's many thousands of processors, peripheral cards, etc. One call once-a-week doesn't sound bad for such a huge installation of hardware. If all of those machines were Dell boxes, you would probably be making calls several times a day!
Suffice to say, it seems like EVERY new Sun machine has at least 1 CPU problem (read: ecache error) within 1 month of operation.
I thought the E-Cache issue is several years old, now, and has been dealt with. Are you talking about brand-new machines delivered within the last few months?
Other issues could just be the initial shake-down for a newly-installed big server. Electronics fail in a distribution that is heavily skewed towards the first moments of operation. Once the initial "wear-in" period occurs, the server should be good for years with few break-downs, if any.
However, as soon as one of their techs shows up on site, any number of idiotic things can happen.
If you don't like their techs, you are fully capable of learning how to fix things yourself. Even big servers don't take a rocket scientist to maintain; it takes a willingness to learn, critical thinking, a static-strap, and some patience. That's it.
So from my point of view, all that crap about Sun hardware and service being great just doesn't add up. I'd much rather have a fleet of Dells running Linux than one E-4500.
So, how are you going to manage a fleet of Dells? Also, clusters or grids scale well for some types of computation, but can be a real PITA for general-purpose computing. There is a big difference in applicability between a cluster and a single SMP server.
Also a thing to note that people don't usually take into consideration: The more processors a machine has, the more likely it is to fail. The more processors you have, the more chances you have for a CPU panic.
You should be able to work around down processors without much trouble. Thankfully, I haven't had to deal with this.
In Enterprise systems, you have to have hardware support - you can't just assume that the box will keep working or that you will be able to fix it. It is not at all uncommon for the hardware support costs on a very old box to be substantially more than it would cost to replace the box.
Actually, even this is arguable. With competent administrators, continuing hardware support is probably just a few thousand a year for several servers (replacing hard drives, the occasional memory DIMM, etc.).
The hardware support you seem to be speaking of is the "sucker-grade" support that manufacturers sell with multi-thousand-dollar/year pricetags just for the contingency that something will break.
By the way, it is usually easier to diagnose a fully-firmware-equipped UNIX box than it is a new PC-based box, so fixes can actually be made more quickly on the older hardware.
In all of this, however, it does require competent people to work on the machines day-to-day. Otherwise, any option will be expensive, as the admins just flounder about trying every little thing that pops into their head.
Go price an Sun 450 with 4 processors. Then price an equivalent Dell/Compaq....
Sun should know better.
Interesting thing is, I would still take the Sun E450, since Dell or Compaq don't sell an equivalent machine. The 450 will hold five independent SCSI controllers for 20 drives. All in one enclosure. It's also robust as hell. They don't have to cost more than $10,000, either, if you find a good used one. Plus, once you find out just how much work an E450 can do, it just might be the only server you need for much of a small company's infrastructure.
Sun competes on things beyond price/performance. Consistency and reliability are one such thing. A more balanced architecture is another (576-bit memory busses, SCSI/FC-AL standard, large CPU caches). Well-engineered enclosures is another. I would also bet that each Sun server design goes through much more testing and quality control than most Intel-based servers (I've read that the UltraSPARC CPUs have a very low errata rate relative to Intel CPUs).
Sun still makes a strong case for itself in its markets. In some ways the prices can be hard to stomach, but, if a company is to the point of affording a real IT infrastructure, they should be as concerned about risk as much as they are up-front cost. Sun equipment tends to be low-risk and very long-lived. It is somewhat harder to claim this for Intel-server Brand X, although I'm sure there are a handful of winners out there.
A lot of these arguments apply to IBM (Power), SGI (MIPS), and other hard-core UNIX server companies as well. The prices vary pretty widely, but they all share a core quality that makes them worthwhile as a long-term investment.
...someone who knows c++ and enjoys working in c++ will be just as productive as someone who knows java and enjoys working in java.
This is very largely true, given the learning curve for any complex system.
Another issue when choosing a language is the requirements for the software. Each language may lend itself well to a different problem, with dramatic results. Both Java and C++ are extremely versatile, but it would be a stretch to say they are "one size fits all."
making improvements to Java, which make it incompatable with older JVM's, and simplifying and improving the GUI bit, exactly what MS did and got slated for?
It doesn't matter what Microsoft tried to do or actually did, because once they make things to work only in their own sandbox, they lose all credibility.
The one thing Microsoft should get credit for is perverting everything they touch. They deliberately try to break things in subtle ways to make themselves look favorable. Microsoft ranks among the dirtiest politicians in the world.
This is why I cringe when I read someone defending Microsoft's actions. Odds are, they are defending the problem and not the solution.
Not Sun, many people would take the time to jump to C# rather than Java3
This isn't true. There is no advantage to migrating to C# over a new version of Java. The amount of work to move to a Java 3 would be substantially less than moving to C#. Note that the article mostly talks about just stripping out deprecated material from the Java API and cleaning things up. A few of the points are major changes, but they shouldn't cripple most Java-based projects.
Not the users. IN GENERAL the cost the rework would be greater than the cost of the benefits of additional clarity.
An introduction of a Java 3 doesn't mean Java 2 just dies off. Sun will have to support both for several years. Existing projects can go on as if nothing changed, and new projects can start with the latest version of Java.
This is nothing new to software development. Just be thankful that Java has been pretty non-volitile over the years. Compare this to Microsoft's tendency to put its tools through a two-year turnover cycle.
Microsoft would like nothing better than for Sun to release a version of Java that isn't backwards compatible.
Don't forget that Sun offers directly from their home page the current Java Runtime for Microsoft Windows. Just because Microsoft doesn't ship a current version of Java doesn't mean that Java is any worse off for it.
If Sun decides to make something like this proposed Java 3, they will simply make it available for Windows just as they do for Solaris and Linux. Also, Sun will obviously provide support for several years for Java 2, just as they do for their other products, such as Solaris, that go through transitions.
Java isn't what is important, it is the write (bytecode) once, run (bytecode) anywhere that is important. Whether that bytecode is generated from Java, Python, Scheme, Ruby, or Joes New Language For His CS210 course, doesn't really matter.
I'm suprised that this fact just doesn't come up often when debating Java vs. the.NET runtime. Any language that will work well on a stack-based architecture will compile very nicely to the JVM.
Even though I really know nothing about C#, I suppose even it could be ported to the JVM, because C# is touted as a Java clone. Microsoft markets a "migration path" from Java to C#; does Sun do something similar for C# to Java? I have seen that there are Visual Basic to Java options.
Another interesting question is whether the GCJ implementation can handle non-Java bytecode. Imagine being able to write Java, Scheme, or Python source, distribute it as bytecode, and, if the end-user wants native binaries, they just run it through GCJ. JVM + GCJ has a lot of potential to be very formidable against.NET.
If you're comfortable with 90's technology, please continue to use it, but know that your keeping your clients behind the times and limiting your job possibilities. I've already spent too much time defending what is obviously here to stay, so unless someone can come up with a new argument against JS, I'm finished here.
Baselessness and belittlement do not make for a convincing argument.
Okay, it's on the order of a GB per week. How much does a GB cost?
For a commercial site, is that cost significant relative to the revenue?
Is the possible bandwidth savings of JavaScript (remember, the JS code gets sent, too) more important to the business than the consistency of server-side processing?
I used JavaScript once a long time ago and it didn't work becuase the coder was an idiot
A long time ago, a year ago, last month, last week, and today. Overall, JavaScript implementation just hasn't gotten better.
And please stop thinking that beefing up your server with extra power is going to speed up the user's machine or their connection any. It doesn't work that way.
Actually, it does. 3KB/sec modem throughput is pretty fast for text, and latencies on the Internet are fast relative to the user "click" and server processing. The server processing aspect of the latency is controllable by the owners of the website, and additional CPUs coupled with a good OS, such as Solaris, will improve the responsiveness.
Server-side processing also guarantees success indpendently of the user's browser configuration.
Slashdot Mirror
How can people honestly believe that a modern government could harbour ANY kind of conspiracy...
Tell that to the people behind the SR-71, the F-117, and the B-2. If the government wants to keep something a secret until they want to reveal it, they will do it.
And these were aircraft which obviously had to fly out in the open, but the general public was very successfully kept in the dark. Now, by extension, how much easier is it to keep unseen things a secret?
It smacks of 'affirmative action' programs and admission standards...
Openness of government is essential, except, perhaps, where the government is "classified". Acheiving this doesn't require Affirmative Action, which has been and always will be a debatable practice.
A good compromise would be for government customers to begin demanding that their software use open file formats and protocols at a minimum. If a closed-source software package is so good, then opening its file format won't destroy its user base, right? Opening up formats and protocols will go very far in leveling the playing field between all the different options, and everyone that uses software will benefit, regardless of Open Source vs. Closed Source.
Yes, most people don't use the features. But if you actually do use one of the features, good luck trying to replicate it in a free software package.
This is a very good reason for people to not use MS Office. Why should a corporation spend tens of thousands of dollars on features only ten people in the company use?
When feature utilization reaches this level in normal software, that software appeals to a niche, while many other people find cheaper alternatives.
For example: Pro/ENGINEER vs. Autocad, Photoshop vs. xv or GIMP, Sun cc vs. GCC, MS Office vs. OpenOffice. There are many more.
It is a basic fact that MS Office has gone out of control in a number of ways. For those people who can live with this, they can be happy with their proprietary formats and interfaces doing what they need to do best. Everyone else can find immense simplicity and flexibility in OpenOffice's good functionality and open file formats.
As long as the other free office suites also follow the model of open file formats, interoperability will be the rule rather than an exception, and MS Office will gradually find itself falling out of fashion. For office productivity, this would be a very very good thing.
My mom is a little out of the loop on the whole computer thing and she bought a PowerBook G4. So far she has broken the computer twice.
Perhaps she should stick to a cheaper computer; the iBook is much more reasonable in cost.
Just right there in less than a year is $700 for my silly mom if she wouldn't have been on warranty. When her warranty expires she will be buying Applecare.
This is another sign that the Powerbook wasn't a good match for your mother. If a product needs to be protected from its user, then something really is amiss.
The right product will: meet the user's needs and not be so expensive as to require financing nor extended warranties. This results in a low-risk purchase that simply works better in the long run. The question should be: "Do I really need a full-blown PowerBook at greater than $2500, or can I get by just fine without spending as much money? The answer can be suprising.
A laptop *is* likely to break at some point over a three-year span either from mechanical failure or (more likely) by human accident, especially if you are in a college environment.
Any object is likely to break if the owner is careless, and default warranties cover nearly all cases where the object breaks due to carelessness on the part of the manufacturer. If the object needs to be protected from its owner, perhaps a more robust version of that object is needed (like an armored laptop).
At a college, it is arguable that the laptop will be stolen before it breaks, which is covered under normal household insurance policies. If it breaks due to someone else's carlessness, they are responsible for the cost of repairing it.
I just generally consider the cost of something, even a computer, as a risk once the default warranty expires. If it breaks, I replace it, if not, I use it as long as it is practical. This works as long as I make purchases within my ability to spend (financing small purchases also makes no sense).
The applecare warranty is a good value.
Only if it provides something above and beyond just warranty service. If it includes hand-holding tech support for learning how to use the computer, it may be worth the cost. But, if it is just a warranty and nothing more, it isn't worth spending a sizable fraction of the purchase price for it.
You're foolish not the get the applecare warranty.
Extendend warranties are a form of insurance. Insurance's purpose in life is to save us from financial catastrophes, such as a tree crushing a home or a major illness. In these cases, insurance is a wise purchase.
However, purchasing insurance on small things, such as a PC or a dishwasher is not a wise investment. For small purchases, the insurance is more expensive than a potential repair, because things just don't break frequently. It is more likely that you will never use the insurance, which is why companies enjoy selling it so much. Add up all the money saved by not buying small-time insurance, and that occasional repair is simply not a big deal, even if it is $350.
It is better to skip the insurance and spend the money on repairs when they are needed. In the long run, you will come out better off. An added benefit is that you maintain only a small number of insurances policies to track, which is helpful given their nitpicking complexity.
where the hell does all that money go?
To feed a very hungry bureaucracy.
...these postings feel more like a lynch gang than an informed discussion.
Lynch gangs usually act irrationally. The better Slashdot posts on the subject of Passport, however, are far from irrational; in fact, they are very objective.
The reason: We understand that Passport is a single point of failure, and its design is fundamentally flawed. It inherently does nothing to protect the information housed within it, and only the most naive individual would continue to believe otherwise in the face of the truth.
I will never sign up for Passport voluntarily, and I will actively seek to do business with Microsoft's competitors in this market. If the WWW decomposes to the point of working only in the context of Passport, then this will surely be a sign of the failure of the WWW. I wholeheartedly hope this will never happen, as the WWW, so far, has been one of the greatest resources in the history of Mankind.
We must all work hard to prevent Microsoft from spoiling the WWW further than they already have. This goes beyond resisting Passport, as the potential of Palladium and other overblown DRM schemes is just as dangerous. Already, Microsoft has done damage using an illegally-gained market share for Internet Explorer.
Why are the general public and the U.S. government allowing them to continue unchecked?
This question baffles and sickens me, as the only answers, it seems, are based in the psychology of cults and organized crime.
This is, of course, opposed to the self-serving testimony of the RIAA.
This is why our only hope is that some of the regulators are shamelessly cynical. If even one person in a powerful position sees just how transparent the media industry's motives are, then, at least, there is a chance for a positive outcome.
On the other hand, what is it they say about hope?
Given that you read something like 1/3 more slowly off a video screen opposed to a paper book, this would suck pretty big-time.
I disagree with this, since video screens have adjustable font sizes and contrast. I can adjust my monitor, for example, to compensate for the inaccuracies in my glasses prescription. On a smaller PDA screen, this can still be done, but with some more limitations. Speed readers, on the other hand, would probably hate video screens, just because they read so damn fast.
Besides, nothing (yet) can replace a book you own, a highlighter and a pen for making notes in the margin and taking the whole thing to a study lounge to get away from your roommate's beer party (the sacrifices we make for good grades).
I absolutely agree with this. One thing I like to do is take a book, find the errata for it, and manually correct the whole thing. Errata in text books is a real PITA, and some books are so bad that they can't be trusted until they are corrected. Would e-books allow correction on-the-fly under the DMCA and/or Palladium-type schemes? I know that it would be possible to upgrade to book version 1.01, but making notes and corrections in real time is pretty much a neccessity.
And then there won't be any supported stable platform for running Windows apps, except WINE and related projects.
This is very true. Even though Open Source software is often accused of volatility, it has an inherent stability as well, because the motives driving it are different. As projects like WINE mature, I would bet that whole consulting businesses develop around supporting Windows applications that, ironically, no longer work in Windows.
A small company (or just a well-run, tight margins, efficient company) that only has the time to develop for one platform will choose windows...
I really think that the "Windows only, because it is popular" arguments are pretty weak now-a-days, because cross-platform development is getting easier all the time. Sure, someone can grab at that first 80 or 90% initially, but why not also tap into that remaining customer base? This also has the nice side-effect of protecting your company from Microsoft's future (whatever that will be) by betting on Microsoft, Apple, and Linux/UNIX simultaneously.
Another added benefit of keeping several platforms in mind at one time, is that Microsoft's proprietary development tools take a proper place among everything else. For example, Visual Studio is simply a modular part of the work flow (compiler & linker only; no binary "makefiles"). This can really help when some tools need to be changed out for others without crippling the project.
Cross-platform need not be gut-wrenchingly hard, either. I work on a UNIX application that's been under development for over a decade. However, it is modular enough and uses APIs wisely enough, that a Windows version isn't out of reach. This could be true, also, of well-planned Windows applications.
I think most of the current arguments for Windows-only applications stem from either short-sightedness, arrogance, or stubbornness. Historically, this wasn't the case, but today things are just different.
I'm tempted to ask how long you've been in the industry to say this.
The specific system I had in mind when writing this was an old HP3000 minicomputer from Way Back In The Day. The 500MB disk drives weighed in at over 100lbs. each.
Okay, this makes sense. My experience is the SCSI-2 and later era, where pretty much everything plays nicely. It's even possible now to put 36GB hard drives into SPARCStations, although I'm not sure how many people would do that. Also, the secondary market for old Sun hardware is pretty healthy. I suppose, perhaps another five or ten years from now, that even SPARCStations will become unmaintainable, which is about where that HP 3000 system got to.
So, the argument changes quite a bit depending on whether the hardware is just old or is really old.
We have (had) about 4500 Sun boxen (Servers, workstations, SunRays, etc) ranging from old Ultra-1s to E-10k.
With that much hardware, what do you expect? That's many thousands of processors, peripheral cards, etc. One call once-a-week doesn't sound bad for such a huge installation of hardware. If all of those machines were Dell boxes, you would probably be making calls several times a day!
Suffice to say, it seems like EVERY new Sun machine has at least 1 CPU problem (read: ecache error) within 1 month of operation.
I thought the E-Cache issue is several years old, now, and has been dealt with. Are you talking about brand-new machines delivered within the last few months?
Other issues could just be the initial shake-down for a newly-installed big server. Electronics fail in a distribution that is heavily skewed towards the first moments of operation. Once the initial "wear-in" period occurs, the server should be good for years with few break-downs, if any.
However, as soon as one of their techs shows up on site, any number of idiotic things can happen.
If you don't like their techs, you are fully capable of learning how to fix things yourself. Even big servers don't take a rocket scientist to maintain; it takes a willingness to learn, critical thinking, a static-strap, and some patience. That's it.
So from my point of view, all that crap about Sun hardware and service being great just doesn't add up. I'd much rather have a fleet of Dells running Linux than one E-4500.
So, how are you going to manage a fleet of Dells? Also, clusters or grids scale well for some types of computation, but can be a real PITA for general-purpose computing. There is a big difference in applicability between a cluster and a single SMP server.
Also a thing to note that people don't usually take into consideration: The more processors a machine has, the more likely it is to fail. The more processors you have, the more chances you have for a CPU panic.
You should be able to work around down processors without much trouble. Thankfully, I haven't had to deal with this.
Actually, you're very wrong here.
In Enterprise systems, you have to have hardware support - you can't just assume that the box will keep working or that you will be able to fix it. It is not at all uncommon for the hardware support costs on a very old box to be substantially more than it would cost to replace the box.
Actually, even this is arguable. With competent administrators, continuing hardware support is probably just a few thousand a year for several servers (replacing hard drives, the occasional memory DIMM, etc.).
The hardware support you seem to be speaking of is the "sucker-grade" support that manufacturers sell with multi-thousand-dollar/year pricetags just for the contingency that something will break.
By the way, it is usually easier to diagnose a fully-firmware-equipped UNIX box than it is a new PC-based box, so fixes can actually be made more quickly on the older hardware.
In all of this, however, it does require competent people to work on the machines day-to-day. Otherwise, any option will be expensive, as the admins just flounder about trying every little thing that pops into their head.
Go price an Sun 450 with 4 processors. Then price an equivalent Dell/Compaq. ...
Sun should know better.
Interesting thing is, I would still take the Sun E450, since Dell or Compaq don't sell an equivalent machine. The 450 will hold five independent SCSI controllers for 20 drives. All in one enclosure. It's also robust as hell. They don't have to cost more than $10,000, either, if you find a good used one. Plus, once you find out just how much work an E450 can do, it just might be the only server you need for much of a small company's infrastructure.
Sun competes on things beyond price/performance. Consistency and reliability are one such thing. A more balanced architecture is another (576-bit memory busses, SCSI/FC-AL standard, large CPU caches). Well-engineered enclosures is another. I would also bet that each Sun server design goes through much more testing and quality control than most Intel-based servers (I've read that the UltraSPARC CPUs have a very low errata rate relative to Intel CPUs).
Sun still makes a strong case for itself in its markets. In some ways the prices can be hard to stomach, but, if a company is to the point of affording a real IT infrastructure, they should be as concerned about risk as much as they are up-front cost. Sun equipment tends to be low-risk and very long-lived. It is somewhat harder to claim this for Intel-server Brand X, although I'm sure there are a handful of winners out there.
A lot of these arguments apply to IBM (Power), SGI (MIPS), and other hard-core UNIX server companies as well. The prices vary pretty widely, but they all share a core quality that makes them worthwhile as a long-term investment.
...someone who knows c++ and enjoys working in c++ will be just as productive as someone who knows java and enjoys working in java.
This is very largely true, given the learning curve for any complex system.
Another issue when choosing a language is the requirements for the software. Each language may lend itself well to a different problem, with dramatic results. Both Java and C++ are extremely versatile, but it would be a stretch to say they are "one size fits all."
making improvements to Java, which make it incompatable with older JVM's, and simplifying and improving the GUI bit, exactly what MS did and got slated for?
It doesn't matter what Microsoft tried to do or actually did, because once they make things to work only in their own sandbox, they lose all credibility.
The one thing Microsoft should get credit for is perverting everything they touch. They deliberately try to break things in subtle ways to make themselves look favorable. Microsoft ranks among the dirtiest politicians in the world.
This is why I cringe when I read someone defending Microsoft's actions. Odds are, they are defending the problem and not the solution.
Not Sun, many people would take the time to jump to C# rather than Java3
This isn't true. There is no advantage to migrating to C# over a new version of Java. The amount of work to move to a Java 3 would be substantially less than moving to C#. Note that the article mostly talks about just stripping out deprecated material from the Java API and cleaning things up. A few of the points are major changes, but they shouldn't cripple most Java-based projects.
Not the users. IN GENERAL the cost the rework would be greater than the cost of the benefits of additional clarity.
An introduction of a Java 3 doesn't mean Java 2 just dies off. Sun will have to support both for several years. Existing projects can go on as if nothing changed, and new projects can start with the latest version of Java.
This is nothing new to software development. Just be thankful that Java has been pretty non-volitile over the years. Compare this to Microsoft's tendency to put its tools through a two-year turnover cycle.
Microsoft would like nothing better than for Sun to release a version of Java that isn't backwards compatible.
Don't forget that Sun offers directly from their home page the current Java Runtime for Microsoft Windows. Just because Microsoft doesn't ship a current version of Java doesn't mean that Java is any worse off for it.
If Sun decides to make something like this proposed Java 3, they will simply make it available for Windows just as they do for Solaris and Linux. Also, Sun will obviously provide support for several years for Java 2, just as they do for their other products, such as Solaris, that go through transitions.
Java isn't what is important, it is the write (bytecode) once, run (bytecode) anywhere that is important. Whether that bytecode is generated from Java, Python, Scheme, Ruby, or Joes New Language For His CS210 course, doesn't really matter.
.NET runtime. Any language that will work well on a stack-based architecture will compile very nicely to the JVM.
.NET.
I'm suprised that this fact just doesn't come up often when debating Java vs. the
Even though I really know nothing about C#, I suppose even it could be ported to the JVM, because C# is touted as a Java clone. Microsoft markets a "migration path" from Java to C#; does Sun do something similar for C# to Java? I have seen that there are Visual Basic to Java options.
Another interesting question is whether the GCJ implementation can handle non-Java bytecode. Imagine being able to write Java, Scheme, or Python source, distribute it as bytecode, and, if the end-user wants native binaries, they just run it through GCJ. JVM + GCJ has a lot of potential to be very formidable against
If you're comfortable with 90's technology, please continue to use it, but know that your keeping your clients behind the times and limiting your job possibilities. I've already spent too much time defending what is obviously here to stay, so unless someone can come up with a new argument against JS, I'm finished here.
Baselessness and belittlement do not make for a convincing argument.
So its smarter to ignore JS because it offends your geek-pride...
It appears this conversation has been exhausted.
you do the math....
Okay, it's on the order of a GB per week. How much does a GB cost?
For a commercial site, is that cost significant relative to the revenue?
Is the possible bandwidth savings of JavaScript (remember, the JS code gets sent, too) more important to the business than the consistency of server-side processing?
I used JavaScript once a long time ago and it didn't work becuase the coder was an idiot
A long time ago, a year ago, last month, last week, and today. Overall, JavaScript implementation just hasn't gotten better.
And please stop thinking that beefing up your server with extra power is going to speed up the user's machine or their connection any. It doesn't work that way.
Actually, it does. 3KB/sec modem throughput is pretty fast for text, and latencies on the Internet are fast relative to the user "click" and server processing. The server processing aspect of the latency is controllable by the owners of the website, and additional CPUs coupled with a good OS, such as Solaris, will improve the responsiveness.
Server-side processing also guarantees success indpendently of the user's browser configuration.