AIX vs. Solaris? DB2 vs. MySQL? This certainly bodes well for IBM's Java offerings and it means they can stop developing their own JRE, if they haven't already. They can also cannibalize Sun's server customers. On the other hand, it seems like this has to mean certain parts of Sun's business die. AIX and Solaris don't both need to exist within the same company. SPARC and POWER don't need to exist within the same company. DB2 and MySQL might, since they target different markets.
I don't. That doesn't change the fact that I'd fire Josh because he's a jackass and I don't like working with jackasses. If someone thought I were a jackass, I wouldn't fault them for firing me because they didn't want to put up with me.
That said, perhaps the fact that I haven't been fired yet suggests I'm not a jackass. Or, alternately, that I'm so badass at what I do that my employer has decided to overlook my jackassery. Or the third possibility that my employer is so clueless, or so desperate, that he is either ignorant of my jackassery or feels like he has no choice but to put up with it. Personally, I think it's option #1.
Funny you should say that. The person in the cube next to me mutters to themselves while working. In French, no less. Possibly a good thing I don't speak French.
He proceeded to convert all the java class files to single lines
That's the most ridiculous thing I've ever heard. I could see maybe running them all through some sort of source-code formatter and checking the re-formatted versions back in. But single line files? Wtf?
While there's certainly a steep learning curve in becoming a developer, at the end of the day its just a matter of hands-on experience. Fold flap A into slot B enough times and you eventually get the hang of it.
I can't disagree more with this statement. IMO there's a certain aesthetic element to coding that most developers just don't "get". You're designing something with moving parts that fit together in a certain way, etc. You want it to be fast, but you also want it to be as "simple" as possible. You want it to be easily extended. Etc. From what I've seen, this isn't something people just "pick up through experience" or learn from a book.
The gag about EE folks moving to CS also existed at my University, mainly because the Engineering college made you "apply" to finish the degree after your first two years, whereas CS was open to anyone. As someone who earned undergraduate degrees in CS and Math, I'll say that the Math students similarly looked down on the Engineering students as being devoid of intellectual curiosity. There was some truth to that stereotype.
Sometimes solving a problem in the most straightforward and human-readable fashion does not result in the most high-performing solution. Sometimes the best code will necessarily be less easy to read than the "most readable" version. That's not a bad thing.
Also, I feel like the idea that "the code is the documentation" gets crapped on a little more than it should. Programming languages have well-defined syntax. If one understands the syntax, one should typically be able to read a block of code and figure out what it does without the benefit of comments. For the most part, I find other developers' comments to be redundant and unhelpful. That's not always the case, but often it is.
I will add, though, that poor developers often write code that is unnecessarily complex. Just because a block of code is complex, that doesn't mean it necessarily represents a brilliant or effective solution. It may just be that the developer sucks, and that's why the code is way more complicated than it need be.
Personally I try to make my code as "concise" as possible, just as a personal preference. Sometimes that makes it slightly less readable than a more verbose (and less efficient) version might be. But then, rarely do I come back to something I've written and end up scratching my head wondering what it does.
My XP installation boots faster than the last few Ubuntu versions I've tried. A substantial portion of most uers' Windows boot time can be attributed to bloatware and/or anti-virus apps, neither of which I tolerate on my Windows installation.
I'm 33. Most of the video I watch, either broadcast or DVD, is on a TV. The same is true for most of my friends. So...if TV's dead, I'm certainly not seeing it.
Yeah, last time I did a comparison it was nightly FF vs. nightly Webkit vs. Chrome's development branch. For benchmarks I used SunSpider and Dromaeo.
Here are the results if you're curious, though they're somewhat stale by now. And here's an addendum where I include numbers for the IE8 RC. The original test used IE8 Beta2.
The speed boost is attributed to TraceMonkey. I've been testing nightly builds for a while now with TraceMonkey enabled and they're generally outperformed (barely) by Webkit nightly builds, and pretty much trounced by Chrome. So if the author is betting on TraceMonkey to give Firefox as massive lead in Javascript performance then he may be in for an unpleasant surprise.
He then raves about how eight critical flaws will be fixed in the upcoming version. Say what? That means there are eight critical unpatched flaws in the current released code that have yet to be repaired. That's a bad thing, not a good thing.
Obviously I don't support child porn...but it seems like all this guy needed to do is obfuscate the encrypted files so that they don't look like encrypted files.
If I were Microsoft, I'd just rip out IE. Ship it on a separate install disk. "Windows Extras" or something. Then they can't be faulted for unfairly promoting IE anymore. But forcing them to bundle Firefox is just silly.
I'm curious, will they force Apple to ship Firefox instead of just Safari?
My point was that DRM has, at times, actually dissuaded some people from pirating software. It's possible that technical reasons make it unfeasible now, but that's different from just saying "DRM never works".
Speaking personally, many years ago, copy protection most certainly *did* prevent me from pirating games. Instead I bought used copies, which came with documentation (i.e. which the game prompted you for), adding to the "resale value" of used games and potentially causing people to purchase more new (since they can turn around and sell them to someone else, reducing their overall cost). Had there been no copy protection I'm pretty sure I wouldn't have bothered to part with my money.
Hence my correction to the initial post. PGO is a method of optimizing compiled code. If the other posters are accurate, then it was employed for the Windows build but not for those packages available on various distros' repositories. While that's not a knock on "linux the operating system", it is definitely an issue for "linux the software ecosystem". If the primary means of obtaining applications in linux-land is likely to give me code that is significantly less optimized than what I'd get in Windows-land, then that's a problem for linux-land.
Okay, I just realized you were talking about a compiler-level optimization and not some "feature" of FireFox that's just not enabled when its built on linux. Ooops. So it's not an app support issue, it's a "distro builders are dumb" problem, assuming they're the one who failed to compile it correctly.
Still bad news for linux, just not "as" bad. Instead of it having crappy performance, it's just not the platform where FF developers decide to implement their newest features first. So it's more of an app support issue.
I know of at least one study in a peer-reviewed journal citing a link between Hib and Hep-B vaccination and development of childhood asthma. The Hep-B link was stronger. Issues like this tend to get lost in all the thimerosal/autism noise.
Prices vary widely outside the United States. It's not entirely fair to cherry pick Japan and France, which are two of the cheapest around. Japan in particular is geographically different in that it has a much higher population density.
Current U.S. prices also vary by which "tier" you purchase. Take AT&T DSL, for example. In my area you can get the 6 Mbit/s service for 4.50 EUR/Month per Mbit/s, or 4.0 GBP/Month per Mbit/s. On the cable side you can get 7 Mbit/s for $40/month, which is slightly cheaper.
While it doesn't approach how cheap broadband is in France or Japan, it's cheaper than what you'd pay in the UK and not too far off what you'd pay in Germany.
One caveat: those cost statistics are from a year and a half ago. No idea how they've changed in the mean time.
While cable does have a higher performance ceiling, at the most common price point for cable there is a comparable DSL alternative that is the same speed. At least, where I live there is. The "standard" cable package is 7 Mb/s for $40/mo. The "premium" DSL package is 6 MB/s for $35/mo. So unless you're paying extra for "premium" cable, you could do about as well with DSL.
Install time? Clicks per install? Installed footprint? Does anybody care about those?
The file copy tests were marginally useful, but not exactly controlled. But it certainly looks like the Linux USB drivers and related I/O code is better than what exists in Windows.
Then again, on what is possibly the most useful and meaningful benchmark, Windows wins. The Richards thing is not disk I/O bound, so we're talking about memory allocation/deallocation and probably some underlying C library calls. Since we're on identical hardware, the difference is either due to the Windows memory manager, faster library routines, or a more optimized version of the python interpreter. (Which wouldn't really be a win for Windows per se.)
I'd like to see something like...oh...a standard database benchmark (e.g. TPC) run on a couple different databases (Postgres and Oracle would be fine) installed under both Ubuntu and under Windows 7 on identical hardware. This would, of course, be influenced by how well optimized these database implementations are on each operating system, but there's little we can do about that. The test would essentially be Windows+Oracle vs. Ubuntu+Oracle, or Windows+Postgres vs. Ubuntu+Postgres.
I think you missed my point. I wasn't trying to suggest that this money would be wasted because it would be spent on something other than the project it's earmarked for. My point was that even if it is spent on the broad band project, and they accomplish their goals, it may still have been a waste. The Korean government's contribution amounts to about $20 USD from every man, woman and child in South Korea. That doesn't seem like much, until you remember that the private telecoms are also spending about 25 times that much. Where do you think they'll get that money? Answer: their customers. In other words, current Korean broadband customers will pay more in the next five years in order to finance this new infrastructure. That's money they're not spending on clothes, entertainment, etc. In other words, every other part of the Korean economy.
Then there's all those jobs. Each one represents someone who isn't working somewhere else. Meaning companies not involved in this project will need to pay more for their labor than they might have otherwise.
All this might be worth it, if the end result of the project was something really useful. But considering they already have 100Mb/s, is the extra speed really that useful? How about when compared with what else you could do with $25 billion USD?
If the new infrastructure is really going to be that useful and beneficial, couldn't the telecoms foot the bill themselves, since they're the ones who're going to eventually profit from it?
Also: It looks doubtful that South Korea is less corrupt than the United States. Don't believe me? US: 7.3, South Korea: 5.6.
Slashdot Mirror
AIX vs. Solaris? DB2 vs. MySQL? This certainly bodes well for IBM's Java offerings and it means they can stop developing their own JRE, if they haven't already. They can also cannibalize Sun's server customers. On the other hand, it seems like this has to mean certain parts of Sun's business die. AIX and Solaris don't both need to exist within the same company. SPARC and POWER don't need to exist within the same company. DB2 and MySQL might, since they target different markets.
I don't. That doesn't change the fact that I'd fire Josh because he's a jackass and I don't like working with jackasses. If someone thought I were a jackass, I wouldn't fault them for firing me because they didn't want to put up with me.
That said, perhaps the fact that I haven't been fired yet suggests I'm not a jackass. Or, alternately, that I'm so badass at what I do that my employer has decided to overlook my jackassery. Or the third possibility that my employer is so clueless, or so desperate, that he is either ignorant of my jackassery or feels like he has no choice but to put up with it. Personally, I think it's option #1.
Funny you should say that. The person in the cube next to me mutters to themselves while working. In French, no less. Possibly a good thing I don't speak French.
That's the most ridiculous thing I've ever heard. I could see maybe running them all through some sort of source-code formatter and checking the re-formatted versions back in. But single line files? Wtf?
I can't disagree more with this statement. IMO there's a certain aesthetic element to coding that most developers just don't "get". You're designing something with moving parts that fit together in a certain way, etc. You want it to be fast, but you also want it to be as "simple" as possible. You want it to be easily extended. Etc. From what I've seen, this isn't something people just "pick up through experience" or learn from a book.
The gag about EE folks moving to CS also existed at my University, mainly because the Engineering college made you "apply" to finish the degree after your first two years, whereas CS was open to anyone. As someone who earned undergraduate degrees in CS and Math, I'll say that the Math students similarly looked down on the Engineering students as being devoid of intellectual curiosity. There was some truth to that stereotype.
If he'd chosen the pseudonym "Joel" instead.
I'd fire Josh just because he's jackass and I don't like working with jackasses.
Sometimes solving a problem in the most straightforward and human-readable fashion does not result in the most high-performing solution. Sometimes the best code will necessarily be less easy to read than the "most readable" version. That's not a bad thing.
Also, I feel like the idea that "the code is the documentation" gets crapped on a little more than it should. Programming languages have well-defined syntax. If one understands the syntax, one should typically be able to read a block of code and figure out what it does without the benefit of comments. For the most part, I find other developers' comments to be redundant and unhelpful. That's not always the case, but often it is.
I will add, though, that poor developers often write code that is unnecessarily complex. Just because a block of code is complex, that doesn't mean it necessarily represents a brilliant or effective solution. It may just be that the developer sucks, and that's why the code is way more complicated than it need be.
Personally I try to make my code as "concise" as possible, just as a personal preference. Sometimes that makes it slightly less readable than a more verbose (and less efficient) version might be. But then, rarely do I come back to something I've written and end up scratching my head wondering what it does.
My XP installation boots faster than the last few Ubuntu versions I've tried. A substantial portion of most uers' Windows boot time can be attributed to bloatware and/or anti-virus apps, neither of which I tolerate on my Windows installation.
I'm 33. Most of the video I watch, either broadcast or DVD, is on a TV. The same is true for most of my friends. So...if TV's dead, I'm certainly not seeing it.
Yeah, last time I did a comparison it was nightly FF vs. nightly Webkit vs. Chrome's development branch. For benchmarks I used SunSpider and Dromaeo.
Here are the results if you're curious, though they're somewhat stale by now. And here's an addendum where I include numbers for the IE8 RC. The original test used IE8 Beta2.
The speed boost is attributed to TraceMonkey. I've been testing nightly builds for a while now with TraceMonkey enabled and they're generally outperformed (barely) by Webkit nightly builds, and pretty much trounced by Chrome. So if the author is betting on TraceMonkey to give Firefox as massive lead in Javascript performance then he may be in for an unpleasant surprise.
He then raves about how eight critical flaws will be fixed in the upcoming version. Say what? That means there are eight critical unpatched flaws in the current released code that have yet to be repaired. That's a bad thing, not a good thing.
Obviously I don't support child porn...but it seems like all this guy needed to do is obfuscate the encrypted files so that they don't look like encrypted files.
Nobody complain about that silly beta label anymore.
If I were Microsoft, I'd just rip out IE. Ship it on a separate install disk. "Windows Extras" or something. Then they can't be faulted for unfairly promoting IE anymore. But forcing them to bundle Firefox is just silly.
I'm curious, will they force Apple to ship Firefox instead of just Safari?
My point was that DRM has, at times, actually dissuaded some people from pirating software. It's possible that technical reasons make it unfeasible now, but that's different from just saying "DRM never works".
Speaking personally, many years ago, copy protection most certainly *did* prevent me from pirating games. Instead I bought used copies, which came with documentation (i.e. which the game prompted you for), adding to the "resale value" of used games and potentially causing people to purchase more new (since they can turn around and sell them to someone else, reducing their overall cost). Had there been no copy protection I'm pretty sure I wouldn't have bothered to part with my money.
Hence my correction to the initial post. PGO is a method of optimizing compiled code. If the other posters are accurate, then it was employed for the Windows build but not for those packages available on various distros' repositories. While that's not a knock on "linux the operating system", it is definitely an issue for "linux the software ecosystem". If the primary means of obtaining applications in linux-land is likely to give me code that is significantly less optimized than what I'd get in Windows-land, then that's a problem for linux-land.
Okay, I just realized you were talking about a compiler-level optimization and not some "feature" of FireFox that's just not enabled when its built on linux. Ooops. So it's not an app support issue, it's a "distro builders are dumb" problem, assuming they're the one who failed to compile it correctly.
Still bad news for linux, just not "as" bad. Instead of it having crappy performance, it's just not the platform where FF developers decide to implement their newest features first. So it's more of an app support issue.
I know of at least one study in a peer-reviewed journal citing a link between Hib and Hep-B vaccination and development of childhood asthma. The Hep-B link was stronger. Issues like this tend to get lost in all the thimerosal/autism noise.
Here's a link to the abstract.
Prices vary widely outside the United States. It's not entirely fair to cherry pick Japan and France, which are two of the cheapest around. Japan in particular is geographically different in that it has a much higher population density.
Current U.S. prices also vary by which "tier" you purchase. Take AT&T DSL, for example. In my area you can get the 6 Mbit/s service for 4.50 EUR/Month per Mbit/s, or 4.0 GBP/Month per Mbit/s. On the cable side you can get 7 Mbit/s for $40/month, which is slightly cheaper.
While it doesn't approach how cheap broadband is in France or Japan, it's cheaper than what you'd pay in the UK and not too far off what you'd pay in Germany.
One caveat: those cost statistics are from a year and a half ago. No idea how they've changed in the mean time.
While cable does have a higher performance ceiling, at the most common price point for cable there is a comparable DSL alternative that is the same speed. At least, where I live there is. The "standard" cable package is 7 Mb/s for $40/mo. The "premium" DSL package is 6 MB/s for $35/mo. So unless you're paying extra for "premium" cable, you could do about as well with DSL.
Install time? Clicks per install? Installed footprint? Does anybody care about those?
The file copy tests were marginally useful, but not exactly controlled. But it certainly looks like the Linux USB drivers and related I/O code is better than what exists in Windows.
Then again, on what is possibly the most useful and meaningful benchmark, Windows wins. The Richards thing is not disk I/O bound, so we're talking about memory allocation/deallocation and probably some underlying C library calls. Since we're on identical hardware, the difference is either due to the Windows memory manager, faster library routines, or a more optimized version of the python interpreter. (Which wouldn't really be a win for Windows per se.)
I'd like to see something like...oh...a standard database benchmark (e.g. TPC) run on a couple different databases (Postgres and Oracle would be fine) installed under both Ubuntu and under Windows 7 on identical hardware. This would, of course, be influenced by how well optimized these database implementations are on each operating system, but there's little we can do about that. The test would essentially be Windows+Oracle vs. Ubuntu+Oracle, or Windows+Postgres vs. Ubuntu+Postgres.
I think you missed my point. I wasn't trying to suggest that this money would be wasted because it would be spent on something other than the project it's earmarked for. My point was that even if it is spent on the broad band project, and they accomplish their goals, it may still have been a waste. The Korean government's contribution amounts to about $20 USD from every man, woman and child in South Korea. That doesn't seem like much, until you remember that the private telecoms are also spending about 25 times that much. Where do you think they'll get that money? Answer: their customers. In other words, current Korean broadband customers will pay more in the next five years in order to finance this new infrastructure. That's money they're not spending on clothes, entertainment, etc. In other words, every other part of the Korean economy.
Then there's all those jobs. Each one represents someone who isn't working somewhere else. Meaning companies not involved in this project will need to pay more for their labor than they might have otherwise.
All this might be worth it, if the end result of the project was something really useful. But considering they already have 100Mb/s, is the extra speed really that useful? How about when compared with what else you could do with $25 billion USD?
If the new infrastructure is really going to be that useful and beneficial, couldn't the telecoms foot the bill themselves, since they're the ones who're going to eventually profit from it?
Also: It looks doubtful that South Korea is less corrupt than the United States. Don't believe me? US: 7.3, South Korea: 5.6.