I've bet the farm on Linux with great success. Our CentOS server cluster does well north of 99.95% uptime and excellent performance. Our normal client(s) replace anywhere from 1-5 servers with our out-sourced solution, where we serve over a hundred clients from a single cluster of 6 servers. Just ONE of our clients replaced some dozen heavily-loaded servers with our software solution, solving the same set of problems, and we didn't even notice the load increase.
I'm not commenting on reality, I'm commenting on perception. Reality is that OSS forms a good foundation for high-class service. But suits see the world in the form of well-dressed salesmen and bottles of fine wine delivered to their houses, which isn't how the OSS typically works.
>What's your credibility to suggest anything at all when you have to come >to (of all places) Slashdot for advice?
Presumably better than if he was the type to pretend he knows everything.
I'm sorry, I didn't mean this as an attack - just a statement of reality. Is the OP trusted by the organization he's representing? I suspect not. Look at it from the perspective of a mid-level suit to see what I was trying to communicate. I respect OP for learning, and eventually he/she/it will understand this lots better. But don't confuse those who pretend to know with those who actually do!
Is this a rational fear?
Does it matter? It's there, and it's both real, and reinforced by widespread anecdote. Who hasn't heard of a migration disaster or three? Having a demonstrably strong organization willing to commit to supporting your OSS solution goes a long, long way: how else do you think IBM manages to sell OSS-based solutions?
isn't this essentially the classic definition of FUD?
Yes. And FUD often works because it's a real effect that manifests on real people.
To evaluate the success of your recommendations, take a look in the mirror. What's your credibility to suggest anything at all when you have to come to (of all places) Slashdot for advice?
Large corps have lots at stake, and they really, really, REALLY are terrified of any solutions that aren't basically guaranteed to work by large, trusted vendors. Stuff that they consider to be a competitive advantage will be enshrouded in mystery while everything else will be outsourced to the most commodity vendor.
Now, compare 'Drupal' to 'Microsoft'. Maybe everybody HERE knows how painful it can be to get MS stuff to work, but nobody is going to be fired for saying MS because it's the biggest commodity vendor in the software space.
Look in the mirror: are you trusted there? When you are fired, who is MEGACORP going to go to when there's a problem?
These questions are being answered by PEOPLE who are afraid that if they make a risky decision, they will suffer the consequences. (get fired/sued/whatever) To sell your OSS solution you have to that there's no/little risk in going with it.
Take a look at Black Box Voting and check it out. A while back, they had a YT video where a hacker was (easily!) able to preload a flash disk with values to rig the vote without there ever being any sign of a problem by the voting machines.
Yes, this is / was Diebold, but unless we use some nice sequential hash algorithms and/or cryptography, along with a verified "clean" starting point, it's not possible to trust electronic voting machines.
Further, the problem is that verifying e-votes and e-voting machines has to be done by a professional programmer and security expert. By definition, this makes verification (and trust) basically impossible for the average person. This means that by operating from authority, programmers and security "experts" could (and have!) certify voting machines and equipment and the general population would have no easy, trustable method to know if they're being hoodwinked.
Sorry, voting machines are a bad, bad, bad idea. As somebody who programs/maintains large databases of sensitive data, I can't say with confidence that I'd even be able to trust an open or OSS solution because of the difficulty in ensuring that the software that's been reviewed is the same as the software that's actually running.
For example, what if your compiler was compromised with a virus, so that the compiler itself produced software that was virus laden?
Sorry, e-voting is too complex. The people responsible for their security are parties of interest, and so by definition can never be trusted. E-voting is a bad, bad, bad idea.
Beverly Harris (at Black Box Voting) is a quintessential example of a modern American Hero. History should remember her with the warmth and love given to Benjamin Franklin and Thomas Paine! I can't say enough how much I respect this average US mom who simply demanded that votes be counted accurately. In so doing, she's changed the world for the better. She's received several hundred dollars from me, and I donate more every year. You would do well to throw $5 her way, and maybe download and use her press pack... it's YOUR freedom at stake!
Since you are asking Slashdot, you don't have a lawyer, or you wouldn't need the reams of idiocy you'll find as a response**. They didn't mention the DMCA, which would at least allow you to defer the problem to the original poster. (who could ask to have it put back up after you notify them)
You got nothing. So, take it down, and resist the urge to post it to wikileaks while enjoying a $0.75 cup of coffee at a local coffee shop with Wifi because that might be considered (ahem) copyright infringement... how long you resist that urge is up to you.
Some fights are worth fighting. It's rarely worth fighting a fight you have no resources to win. Pearson is a big, big, big uber-ultra teh evil mega-corporation, but this is unlikely to be controversial enough to benefit from the Streissand Effect.
PS: Strikes me that this is a good time for us to truly develop cross-core, cross-system, and cross-cluster application development by unifying cross-thread, cross-core, and cross-system communications under a common API.
Really, as a clustered application developer, why should I have to worry about where a process runs, on what server? The POSIX process scheduler should be extended to run applications on the cluster with the SAME API as launching a thread.
It's stupid that I have to think about all the details when all I want is:
A) X done. B) In another process/thread.
Somebody who "gets" this could make a pretty serious dent ini the marketplace!
Languages like PHP/Perl, as a rule, are not designed for threading - at ALL. This makes multi-core performance a non-starter. Sure, you can run more INSTANCES of the language with multiple cores, but you can't get any single instance of a script to run any faster than what a single core can do.
I have, so, so, SOOOO many times wished I could split a PHP script into threads, but it's just not there. The closest you can get is with (heavy, slow, painful) forking and multiprocess communication through sockets or (worse) shared memory.
Truth be told, there's a whole rash of security issues through race conditions that we'll soon have crawling out of nearly every pore as the development community slowly digests multi-threaded applications (for real!) in the newly commoditized multi-CPU environment.
so some bad happens. Who do you report it to? Local police? State police? Federal police? And if so, do you mean: FBI? CIA? Department of Homeland Security? National Security Agency? Peace Corps? Coast Guard? National Guard?
We don't need another agency. We need about half of these dissolved / merged so it's understandable who's in charge of what!
Wonder why 'we' are never happy here on slashdot? Why no matter what MicroSoft does, they are vilified by 'we'?
Here's a hint: take your user Id, and subtract 1. That's about how many DIFFERENT people registered here before yoi did. Each with their own ideas about priority and values, and what to lambaste MS for.
I lambaste them for lame things like email not working right with IMAP4 servers in WinMobile 5, 6, 6.1, and 6.5. That's 3 YEARS that some as simple as deleting an email hasn't worked right in a device primarily bought to (ahem) read email.
Modded insightful, the advice to do whatever you like is surprisingly close, IMHO,
See, college really exists to get you familiar with the basics. It then takes EXPERIENCE to advance to coding kernels and the like. How much of that you do is determined by what you LIKE to do. So, you go to school, then you bang around for a while on small jobs and hobby programming 'till you find your groove. Then, you go into whatever field you like.
make a very comfortable living as a programmer, and I'm damned good at it. But kernels and drivers? Not my gig. But I'm mean when you need a perl/php script with javascript and ANSI SQL. I literally dream in code.
So bang around, find out what your favorite cup of tea is, and then pick your industry based on that experience. Programmers are in a unique position: you basically never have to choose between love and money, because you are so much better at what you love that you'll generally make more money there, anyway!
Clustered filesystems are, as a breed, ridiculously over complicated. perhaps king of the hill is OCFS. To get it working right, your entire cluster has to to perform a series of steps IN SYNCH. EG: your entire cluster must all be done with step 1 before they all do step 2, etc. Just too complex, and no way to be redundant without blowing loads of cash on highly complex hardware....
Sorry... NO!
If you want simple, redundant storage, you really have to do it in the application layer. Doing it at the OS level requires too much abstraction to do well while maintaining decent performance. The closest I've seen to a decent clustered filesystem from an administrative standpoint is Gluster F/S...
Going one further, there isn't even a compelling advantage, in many cases, of any medium at all. When I watch shows/movies at Netflix instant play or Hulu or any of the bazillion other portals springing up everywhere, I get to watch something with good enough quality that I generally don't notice artifacts on a 24" wide screen viewed in a normal 12x14 bedroom (occasional dropped frames) with a 3 Mb DSL service. It's basically never bad enough to ruin the viewing experience.
So I get:
1) Something to watch NOW. No wait.
2) No need to go to a store,
3) Nothing to lose,
4) Reasonable selection, (better than I ever had in purchased DVDs - once you've watched a movie 1 or 2 times, do you EVER watch it again?).... and
4) Reasonably good quality.
I might care more if I had a bigger screen, or if I sat 18" away from the screen, but I don't. And I don't expect to spend any more in a few years for 10-20 Mb Internet than I pay for my 3 Mb now, so if/when I care at that point, the problem will already be solved.
I might care if I had a favorite movie that I wanted to watch over and over, but I don't. The last movie set I cared about was the Matrix set, and I have all the DVDs. But my son and I wanted to do a Matrix marathon, and even though I own the DVDs, we ended up doing it all on Netflix Instant Play!
So... what does a Blue-Ray disk give me, again? That's right. Nothing.
I am surprised that people even bothered to do research on this. I could have told you this without looking at any metrics.
Sure, you coulda SAID that this was so. You could have even pulled a convincing number from your arse. But just because what you say is validated by real research, it's the real research that demonstrates the validity of what you said.
If I said that "it's cold in Houston, TX", sometimes, I'd be right, notwithstanding any research. But actually getting somebody to report the temperature with a thermometer in Houston, TX both costs more and has much more value than my original statement.
It's the proof of validity that's valuable, now the correctness of the statement.
Think about it: Every January, the mediums come up with some predictions for the coming year that you'll read in the trash rags at the grocery lines. Just because they get it right sometimes doesn't mean that their predictions are worth any more than the cheap, crappy paper they're printed on.
Wrenches can be used to kill people. So can metal broomhandles and tennis rackets. This fact does not make anyone who carries a wrench, a broomhandle, or a tennis racket a bad murderer.
(My father used to keep a tennis racket in the car to arm himself when he was in a bad neighborhood - works well and you won't be accused of brandishing a weapon)
There are few (any?) other companies that 'get' OSS more/better than Red Hat. They license their patents to the community at large, effectively subverting the usual use of patents, much as using the copyleft license uses copyrights to flip normal copyright conventions on their ear.
Red Hat is one of the really, really good guys. With Red Hat patenting CGI/SOAP, they've effectively prevented others from doing so, assuring all of our freedom to use CGI/SOAP. (No, I don't work for RH, though I was once a shareholder)
Blaming RH for its patents are like calling somebody who uses a shotgun to get mistletoe out of a tree a murderer.
I can't imagine OSX has a good Silverlight implementation. I couldn't even find evidence of Moonlight being ported over.
I can't either, but that's because I don't have to. I have an Intel Mac Mini, and Silverlight/Netflix works perfectly - so well that I've replaced my cable service with Netflix and Internet TV.
Sad that Linux isn't supported (yet) but I understand that Moonlight is making good progress...
The answer to our traffic woes is probably not flying cars, but rather something like self-driving cars on defined tracks. Most of our traffic problems are caused by people following too closely and overreacting to developments ahead of them (braking harder than necessary, etc), not to mention the general scourge of distracted driving.
OK, I'm following....
If the whole process of freeway merging, maintaining safe distance, responding to stimuli outside the vehicle, etc, was handled by an unemotional computer (perhaps interfacing with a central traffic planning computer in more congested areas), things should smooth out.
Assume you have a number of lanes. Assume that you are spacing cars 2 seconds apart. (The minimum distance that most people can drive and still have enough response time to not hit the car in front in an emergency)
Now run cars on your highway, spaced exactly 2 seconds apart, all going exactly 30 miles per hour. Count the number of cars that pass on the highway in a given period of time. Now, run those same cars on the same highway at 60 MPH, and notice how many cars pass on the highway during the same amount of time.
It's not going to be much different! Yes, the cars go faster. Yes, individual cars get home faster. But the actual throughput of the highway doesn't rise much as you speed up a well-spaced highway, because by definition, basing the capacity of the highway on a fixed response time (2 seconds) you limit the actual maximum throughput of the highway.
And most highways are already much denser than this theoretical throughput, now. People don't tend to drive a whole 2 seconds apart. And putting l337 computers in charge won't change things much - only about 3/4 of one of those 2 seconds is actually spent in response time. The rest is spent in actually braking the car. If a computer was in charge, you'd only slightly more than double the safe limit capacity on a highway.
It's an improvement, but probably not much (if any) over the amount that the highways are carrying today. How do you plan to deal with the actual limits of driving?
Want less congestion? BUILD MORE ROADS. How hard is that to understand?
I first tried Gentoo on an AMD K62 with 96 MB RAM. Back then, a usable system took at least a week if there were no compiler errors. Which there were, I never got it working because I gave up after the first week or so.
But now, Moore's law has largely mitigated this nightmare - you can have a working system in a few hours! Of course, the faster systems also obviate a primary need for Gentoo (performance) but...
The typical argument goes something like: 'MySQL suxorz - nobody uses it for serios work' followed by: 'Yeah? well explain that to =HIGH VOLUME SITE=!'
Such responses show a misunderstanding of what serious work is being discussed.
MySQL does a fabulous job with simple, high-volume transactions, exactly the type seen by Yahoogle/Flicker/Blogsites. They need to sore simple data (EG text) and be able to retrieve it quickly, and for these uses, MySQL is probably a better bet than Postgres or DB2.
But 'serios work' means thing like strong, ACID compliant transactions, row-level locking, strong integrity of field types, and a query scheduler that holds its own when you combine inner, outer, nested, subqueries mashing together a dozen or more tables with millions/billions of records/combinations.
Postgres will do this, MySQL won't. MySQL isn't bad because of this, it's just a tool not well suited to this specific job. I use MySQL for website CMS, I use Postgres for financial applications.
Does your dishwasher suck because it does a piss-poor job cleaning your socks? Use the right tool for the job.
Funny - I've never had any trouble watching Dollhouse whenever I want to. It's always on at the same time - when I want to see it - just by going to http://www.fox.com/fod/
Seriously - Internet TV changes everything. I moved a couple months back, ditched the dual-dish + DVR that we had the old house while we moved. In the meantime, we've been using hulu, netflix, wtso.net, casttv.com, and the like to watch our shows.
Internet TV rocks! We watch what we want, when we want to, with little/no commercial interruptions, LEGALLY! Even when we've never heard of the show before... Downloading torrents is crap. I'm talking about click & play, armed with nothing more than a 3 Mb Internet connection and a Mac mini with a big screen.
Seriously, give it a shot. I don't think I'll ever end up buying an HDTV.
Meanwhile, Linux user seem to be content pointing and laughing at Microsoft's efforts and pointing out that Linux is so much more secure.
Because it is. There. I said it.
The relatively simple, understandable Unix security model has a very long history, and has grown gracefully as the strength, power, speed, and ability of the individual computers have. Everything is a file, and all files have the three permissions: Users, Groups, and Other. Each of these can have read, write, and execute permissions. Simple, understandable, easy to enforce. It's so taken for granted as such that it's routinely used in embedded devices (such as routers) where updates are few and far between, yet they are rarely, if ever, compromised.
Compare/contrast that with the Windows security model, where there are actually alternate file spaces within the existing file system. With the Windows API, it's trivial to save a file that's in an alternate namespace and thus cannot be found with *any* normal Windows system call. There are many examples of strangeness like this!
There was a recent article I read about the confessions of a grey-hat programmer... he describes Windows as incredibly complex, labyrinthine, and basically impossible to secure well. He laughed at so-called "security vendors" like anti-virus.
First, Microsoft tried to make the browser part of their operating system, without paying much attention to security. Now, they're trying to make a browser into an operating system with security first in mind?
Looks like an about-face if you ask me...
Funny how the vendor of one of the world's most insecure operating systems now considers that they're going to one-up the competition with the most secure browser / operating system? I guess they'd have an excellent track record of finding out what not to do... ?
Projects like this are always limited by a single factor: energy density.
Loads of heavy batteries that only seem to last an hour or so, or loud, smelly, fault-prone ICEs are par for the course. See, millions of years of evolution have resulted in bodies that are surprisingly efficient in a wide variety of circumstances and pack loads of energy into a very little weight. When your body truly runs out of energy in sustained exhaustion, it can even burn its own motor (muscle tissue) for a last bit of energy!
The problems are many and severe. It will be a while before exoskeletons are worth much.
What you are forgetting is that most companies, especially large companies ARE boring places staffed by a high percentage of mediocre people. Large organizations have a large amount of administrative overhead, and the vetting process is long, convoluted, and inefficient. It's just the nature of the beast.
1) IT staffed by control freaks? Well duh! It's the only way they can appear to be doing something and not getting their asses handed back to them if anything goes wrong...
2) Trust? How much do YOU trust people you know just barely well enough to remember their name? And anytime you get more than 5 people together, they start grouping up and taking sides. Disputes soon follow. Care to guess what it's like when there are 500?
3) Hiring standards? Have you seen who applies to Monster.com ads? As an employer, I can say the domain name is appropriate...
4) unrealistc expectations... It's often hard enough to simply establish expectations at all. 5) Morale? You want to talk about morale!?!? Large companies spend months rolling out big updates like using actual coffe in the coffe makers at their 2,000 store fronts, or on 6 month programs toget locations to clean their bathrooms. Wait until you spend a man-week working yer ass off because somebody didn't know what 'historic' meant, only to find you didn't need to do anything at all. Then see what your morale is like.
6) Unmotivated employees? Your average wage slave is motivated by a desire to do as little as possible and not get yelled at.
Go work at/for/with some large organizations sometime. You'll see why Dilbert is so popular - not because it's quirky and off-beat but because IT'S TRUE!
Slashdot Mirror
And you're telling me this because... why again?
I've bet the farm on Linux with great success. Our CentOS server cluster does well north of 99.95% uptime and excellent performance. Our normal client(s) replace anywhere from 1-5 servers with our out-sourced solution, where we serve over a hundred clients from a single cluster of 6 servers. Just ONE of our clients replaced some dozen heavily-loaded servers with our software solution, solving the same set of problems, and we didn't even notice the load increase.
I'm not commenting on reality, I'm commenting on perception. Reality is that OSS forms a good foundation for high-class service. But suits see the world in the form of well-dressed salesmen and bottles of fine wine delivered to their houses, which isn't how the OSS typically works.
>What's your credibility to suggest anything at all when you have to come
>to (of all places) Slashdot for advice?
Presumably better than if he was the type to pretend he knows everything.
I'm sorry, I didn't mean this as an attack - just a statement of reality. Is the OP trusted by the organization he's representing? I suspect not. Look at it from the perspective of a mid-level suit to see what I was trying to communicate. I respect OP for learning, and eventually he/she/it will understand this lots better. But don't confuse those who pretend to know with those who actually do!
Is this a rational fear?
Does it matter? It's there, and it's both real, and reinforced by widespread anecdote. Who hasn't heard of a migration disaster or three? Having a demonstrably strong organization willing to commit to supporting your OSS solution goes a long, long way: how else do you think IBM manages to sell OSS-based solutions?
isn't this essentially the classic definition of FUD?
Yes. And FUD often works because it's a real effect that manifests on real people.
To evaluate the success of your recommendations, take a look in the mirror. What's your credibility to suggest anything at all when you have to come to (of all places) Slashdot for advice?
Large corps have lots at stake, and they really, really, REALLY are terrified of any solutions that aren't basically guaranteed to work by large, trusted vendors. Stuff that they consider to be a competitive advantage will be enshrouded in mystery while everything else will be outsourced to the most commodity vendor.
Now, compare 'Drupal' to 'Microsoft'. Maybe everybody HERE knows how painful it can be to get MS stuff to work, but nobody is going to be fired for saying MS because it's the biggest commodity vendor in the software space.
Look in the mirror: are you trusted there? When you are fired, who is MEGACORP going to go to when there's a problem?
These questions are being answered by PEOPLE who are afraid that if they make a risky decision, they will suffer the consequences. (get fired/sued/whatever) To sell your OSS solution you have to that there's no/little risk in going with it.
Good luck.
The point is... how would you know?
Take a look at Black Box Voting and check it out. A while back, they had a YT video where a hacker was (easily!) able to preload a flash disk with values to rig the vote without there ever being any sign of a problem by the voting machines.
Yes, this is / was Diebold, but unless we use some nice sequential hash algorithms and/or cryptography, along with a verified "clean" starting point, it's not possible to trust electronic voting machines.
Further, the problem is that verifying e-votes and e-voting machines has to be done by a professional programmer and security expert. By definition, this makes verification (and trust) basically impossible for the average person. This means that by operating from authority, programmers and security "experts" could (and have!) certify voting machines and equipment and the general population would have no easy, trustable method to know if they're being hoodwinked.
Sorry, voting machines are a bad, bad, bad idea. As somebody who programs/maintains large databases of sensitive data, I can't say with confidence that I'd even be able to trust an open or OSS solution because of the difficulty in ensuring that the software that's been reviewed is the same as the software that's actually running.
For example, what if your compiler was compromised with a virus, so that the compiler itself produced software that was virus laden?
Sorry, e-voting is too complex. The people responsible for their security are parties of interest, and so by definition can never be trusted. E-voting is a bad, bad, bad idea.
Beverly Harris (at Black Box Voting) is a quintessential example of a modern American Hero. History should remember her with the warmth and love given to Benjamin Franklin and Thomas Paine! I can't say enough how much I respect this average US mom who simply demanded that votes be counted accurately. In so doing, she's changed the world for the better. She's received several hundred dollars from me, and I donate more every year. You would do well to throw $5 her way, and maybe download and use her press pack... it's YOUR freedom at stake!
Your choice is pretty simple:
1) Take it down, or
2) Deal with their lawyers.
Since you are asking Slashdot, you don't have a lawyer, or you wouldn't need the reams of idiocy you'll find as a response**. They didn't mention the DMCA, which would at least allow you to defer the problem to the original poster. (who could ask to have it put back up after you notify them)
You got nothing. So, take it down, and resist the urge to post it to wikileaks while enjoying a $0.75 cup of coffee at a local coffee shop with Wifi because that might be considered (ahem) copyright infringement... how long you resist that urge is up to you.
Some fights are worth fighting. It's rarely worth fighting a fight you have no resources to win. Pearson is a big, big, big uber-ultra teh evil mega-corporation, but this is unlikely to be controversial enough to benefit from the Streissand Effect.
You have much better things to worry about.
** Feel free to consider this post idiotic
PS: Strikes me that this is a good time for us to truly develop cross-core, cross-system, and cross-cluster application development by unifying cross-thread, cross-core, and cross-system communications under a common API.
Really, as a clustered application developer, why should I have to worry about where a process runs, on what server? The POSIX process scheduler should be extended to run applications on the cluster with the SAME API as launching a thread.
It's stupid that I have to think about all the details when all I want is:
A) X done.
B) In another process/thread.
Somebody who "gets" this could make a pretty serious dent ini the marketplace!
Languages like PHP/Perl, as a rule, are not designed for threading - at ALL. This makes multi-core performance a non-starter. Sure, you can run more INSTANCES of the language with multiple cores, but you can't get any single instance of a script to run any faster than what a single core can do.
I have, so, so, SOOOO many times wished I could split a PHP script into threads, but it's just not there. The closest you can get is with (heavy, slow, painful) forking and multiprocess communication through sockets or (worse) shared memory.
Truth be told, there's a whole rash of security issues through race conditions that we'll soon have crawling out of nearly every pore as the development community slowly digests multi-threaded applications (for real!) in the newly commoditized multi-CPU environment.
GLUSTER FS IS NOT GFS.
Please read up on it. GFS is better than OCFS, but Gluster FS is WAY better than both.
so some bad happens. Who do you report it to? Local police?
State police?
Federal police? And if so, do you mean:
FBI?
CIA?
Department of Homeland Security?
National Security Agency?
Peace Corps?
Coast Guard?
National Guard?
We don't need another agency. We need about half of these dissolved / merged so it's understandable who's in charge of what!
Wonder why 'we' are never happy here on slashdot? Why no matter what MicroSoft does, they are vilified by 'we'?
Here's a hint: take your user Id, and subtract 1. That's about how many DIFFERENT people registered here before yoi did. Each with their own ideas about priority and values, and what to lambaste MS for.
I lambaste them for lame things like email not working right with IMAP4 servers in WinMobile 5, 6, 6.1, and 6.5. That's 3 YEARS that some as simple as deleting an email hasn't worked right in a device primarily bought to (ahem) read email.
Modded insightful, the advice to do whatever you like is surprisingly close, IMHO,
See, college really exists to get you familiar with the basics. It then takes EXPERIENCE to advance to coding kernels and the like. How much of that you do is determined by what you LIKE to do. So, you go to school, then you bang around for a while on small jobs and hobby programming 'till you find your groove. Then, you go into whatever field you like.
make a very comfortable living as a programmer, and I'm damned good at it. But kernels and drivers? Not my gig. But I'm mean when you need a perl/php script with javascript and ANSI SQL. I literally dream in code.
So bang around, find out what your favorite cup of tea is, and then pick your industry based on that experience. Programmers are in a unique position: you basically never have to choose between love and money, because you are so much better at what you love that you'll generally make more money there, anyway!
What did you Linus was doing when
Clustered filesystems are, as a breed, ridiculously over complicated. perhaps king of the hill is OCFS. To get it working right, your entire cluster has to to perform a series of steps IN SYNCH. EG: your entire cluster must all be done with step 1 before they all do step 2, etc. Just too complex, and no way to be redundant without blowing loads of cash on highly complex hardware....
Sorry... NO!
If you want simple, redundant storage, you really have to do it in the application layer. Doing it at the OS level requires too much abstraction to do well while maintaining decent performance. The closest I've seen to a decent clustered filesystem from an administrative standpoint is Gluster F/S...
Going one further, there isn't even a compelling advantage, in many cases, of any medium at all. When I watch shows/movies at Netflix instant play or Hulu or any of the bazillion other portals springing up everywhere, I get to watch something with good enough quality that I generally don't notice artifacts on a 24" wide screen viewed in a normal 12x14 bedroom (occasional dropped frames) with a 3 Mb DSL service. It's basically never bad enough to ruin the viewing experience.
So I get:
1) Something to watch NOW. No wait.
2) No need to go to a store,
3) Nothing to lose,
4) Reasonable selection, (better than I ever had in purchased DVDs - once you've watched a movie 1 or 2 times, do you EVER watch it again?).... and
4) Reasonably good quality.
I might care more if I had a bigger screen, or if I sat 18" away from the screen, but I don't. And I don't expect to spend any more in a few years for 10-20 Mb Internet than I pay for my 3 Mb now, so if/when I care at that point, the problem will already be solved.
I might care if I had a favorite movie that I wanted to watch over and over, but I don't. The last movie set I cared about was the Matrix set, and I have all the DVDs. But my son and I wanted to do a Matrix marathon, and even though I own the DVDs, we ended up doing it all on Netflix Instant Play!
So... what does a Blue-Ray disk give me, again? That's right. Nothing.
Physical media is doomed.
I am surprised that people even bothered to do research on this. I could have told you this without looking at any metrics.
Sure, you coulda SAID that this was so. You could have even pulled a convincing number from your arse. But just because what you say is validated by real research, it's the real research that demonstrates the validity of what you said.
If I said that "it's cold in Houston, TX", sometimes, I'd be right, notwithstanding any research. But actually getting somebody to report the temperature with a thermometer in Houston, TX both costs more and has much more value than my original statement.
It's the proof of validity that's valuable, now the correctness of the statement.
Think about it: Every January, the mediums come up with some predictions for the coming year that you'll read in the trash rags at the grocery lines. Just because they get it right sometimes doesn't mean that their predictions are worth any more than the cheap, crappy paper they're printed on.
Wrenches can be used to kill people. So can metal broomhandles and tennis rackets. This fact does not make anyone who carries a wrench, a broomhandle, or a tennis racket a bad murderer.
(My father used to keep a tennis racket in the car to arm himself when he was in a bad neighborhood - works well and you won't be accused of brandishing a weapon)
There are few (any?) other companies that 'get' OSS more/better than Red Hat. They license their patents to the community at large, effectively subverting the usual use of patents, much as using the copyleft license uses copyrights to flip normal copyright conventions on their ear.
Red Hat is one of the really, really good guys. With Red Hat patenting CGI/SOAP, they've effectively prevented others from doing so, assuring all of our freedom to use CGI/SOAP. (No, I don't work for RH, though I was once a shareholder)
Blaming RH for its patents are like calling somebody who uses a shotgun to get mistletoe out of a tree a murderer.
I can't imagine OSX has a good Silverlight implementation. I couldn't even find evidence of Moonlight being ported over.
I can't either, but that's because I don't have to. I have an Intel Mac Mini, and Silverlight/Netflix works perfectly - so well that I've replaced my cable service with Netflix and Internet TV.
Sad that Linux isn't supported (yet) but I understand that Moonlight is making good progress...
The answer to our traffic woes is probably not flying cars, but rather something like self-driving cars on defined tracks. Most of our traffic problems are caused by people following too closely and overreacting to developments ahead of them (braking harder than necessary, etc), not to mention the general scourge of distracted driving.
OK, I'm following....
If the whole process of freeway merging, maintaining safe distance, responding to stimuli outside the vehicle, etc, was handled by an unemotional computer (perhaps interfacing with a central traffic planning computer in more congested areas), things should smooth out.
Assume you have a number of lanes. Assume that you are spacing cars 2 seconds apart. (The minimum distance that most people can drive and still have enough response time to not hit the car in front in an emergency)
Now run cars on your highway, spaced exactly 2 seconds apart, all going exactly 30 miles per hour. Count the number of cars that pass on the highway in a given period of time. Now, run those same cars on the same highway at 60 MPH, and notice how many cars pass on the highway during the same amount of time.
It's not going to be much different! Yes, the cars go faster. Yes, individual cars get home faster. But the actual throughput of the highway doesn't rise much as you speed up a well-spaced highway, because by definition, basing the capacity of the highway on a fixed response time (2 seconds) you limit the actual maximum throughput of the highway.
And most highways are already much denser than this theoretical throughput, now. People don't tend to drive a whole 2 seconds apart. And putting l337 computers in charge won't change things much - only about 3/4 of one of those 2 seconds is actually spent in response time. The rest is spent in actually braking the car. If a computer was in charge, you'd only slightly more than double the safe limit capacity on a highway.
It's an improvement, but probably not much (if any) over the amount that the highways are carrying today. How do you plan to deal with the actual limits of driving?
Want less congestion? BUILD MORE ROADS. How hard is that to understand?
I first tried Gentoo on an AMD K62 with 96 MB RAM. Back then, a usable system took at least a week if there were no compiler errors. Which there were, I never got it working because I gave up after the first week or so.
But now, Moore's law has largely mitigated this nightmare - you can have a working system in a few hours! Of course, the faster systems also obviate a primary need for Gentoo (performance) but...
I'm sticking with Fedora, thanks!
PS: Your company is pissing away tens of thousands of dollars on Oracle, when you could use PostgreSQL for free!
And no, I haven't read your requirements, but I'd be intrigued to find out what needs Oracle answers that PostgreSQL can't!
The typical argument goes something like: 'MySQL suxorz - nobody uses it for serios work' followed by: 'Yeah? well explain that to =HIGH VOLUME SITE=!'
Such responses show a misunderstanding of what serious work is being discussed.
MySQL does a fabulous job with simple, high-volume transactions, exactly the type seen by Yahoogle/Flicker/Blogsites. They need to sore simple data (EG text) and be able to retrieve it quickly, and for these uses, MySQL is probably a better bet than Postgres or DB2.
But 'serios work' means thing like strong, ACID compliant transactions, row-level locking, strong integrity of field types, and a query scheduler that holds its own when you combine inner, outer, nested, subqueries mashing together a dozen or more tables with millions/billions of records/combinations.
Postgres will do this, MySQL won't. MySQL isn't bad because of this, it's just a tool not well suited to this specific job. I use MySQL for website CMS, I use Postgres for financial applications.
Does your dishwasher suck because it does a piss-poor job cleaning your socks? Use the right tool for the job.
Funny - I've never had any trouble watching Dollhouse whenever I want to. It's always on at the same time - when I want to see it - just by going to http://www.fox.com/fod/
Seriously - Internet TV changes everything. I moved a couple months back, ditched the dual-dish + DVR that we had the old house while we moved. In the meantime, we've been using hulu, netflix, wtso.net, casttv.com, and the like to watch our shows.
Internet TV rocks! We watch what we want, when we want to, with little/no commercial interruptions, LEGALLY! Even when we've never heard of the show before... Downloading torrents is crap. I'm talking about click & play, armed with nothing more than a 3 Mb Internet connection and a Mac mini with a big screen.
Seriously, give it a shot. I don't think I'll ever end up buying an HDTV.
Meanwhile, Linux user seem to be content pointing and laughing at Microsoft's efforts and pointing out that Linux is so much more secure.
Because it is. There. I said it.
The relatively simple, understandable Unix security model has a very long history, and has grown gracefully as the strength, power, speed, and ability of the individual computers have. Everything is a file, and all files have the three permissions: Users, Groups, and Other. Each of these can have read, write, and execute permissions. Simple, understandable, easy to enforce. It's so taken for granted as such that it's routinely used in embedded devices (such as routers) where updates are few and far between, yet they are rarely, if ever, compromised.
Compare/contrast that with the Windows security model, where there are actually alternate file spaces within the existing file system. With the Windows API, it's trivial to save a file that's in an alternate namespace and thus cannot be found with *any* normal Windows system call. There are many examples of strangeness like this!
There was a recent article I read about the confessions of a grey-hat programmer... he describes Windows as incredibly complex, labyrinthine, and basically impossible to secure well. He laughed at so-called "security vendors" like anti-virus.
First, Microsoft tried to make the browser part of their operating system, without paying much attention to security. Now, they're trying to make a browser into an operating system with security first in mind?
Looks like an about-face if you ask me...
Funny how the vendor of one of the world's most insecure operating systems now considers that they're going to one-up the competition with the most secure browser / operating system? I guess they'd have an excellent track record of finding out what not to do... ?
PS: Good luck with retro compatibility!
Projects like this are always limited by a single factor: energy density.
Loads of heavy batteries that only seem to last an hour or so, or loud, smelly, fault-prone ICEs are par for the course. See, millions of years of evolution have resulted in bodies that are surprisingly efficient in a wide variety of circumstances and pack loads of energy into a very little weight. When your body truly runs out of energy in sustained exhaustion, it can even burn its own motor (muscle tissue) for a last bit of energy!
The problems are many and severe. It will be a while before exoskeletons are worth much.
What you are forgetting is that most companies, especially large companies ARE boring places staffed by a high percentage of mediocre people. Large organizations have a large amount of administrative overhead, and the vetting process is long, convoluted, and inefficient. It's just the nature of the beast.
1) IT staffed by control freaks? Well duh! It's the only way they can appear to be doing something and not getting their asses handed back to them if anything goes wrong...
2) Trust? How much do YOU trust people you know just barely well enough to remember their name? And anytime you get more than 5 people together, they start grouping up and taking sides. Disputes soon follow. Care to guess what it's like when there are 500?
3) Hiring standards? Have you seen who applies to Monster.com ads? As an employer, I can say the domain name is appropriate...
4) unrealistc expectations... It's often hard enough to simply establish expectations at all. 5) Morale? You want to talk about morale!?!? Large companies spend months rolling out big updates like using actual coffe in the coffe makers at their 2,000 store fronts, or on 6 month programs toget locations to clean their bathrooms. Wait until you spend a man-week working yer ass off because somebody didn't know what 'historic' meant, only to find you didn't need to do anything at all. Then see what your morale is like.
6) Unmotivated employees? Your average wage slave is motivated by a desire to do as little as possible and not get yelled at.
Go work at/for/with some large organizations sometime. You'll see why Dilbert is so popular - not because it's quirky and off-beat but because IT'S TRUE!