Dev machine #1: P4-3.8GHz, 64bit one, with 3GB of RAM running about three virtual machines (MS Virtual Server), one SQL, one WinXP to test the UI, one runs apps. OS is W2K3 Enterprise (a requirement for the stuff I develop).
Dev machine #2: Dual Xeon 3.02GHz with 2GB of RAM. No VMs, no nothing, just a subset of apps that can't run on machine #1, and SQL server.
Monitor: Dell 2405FPW - bought with my own money. The one provided to me by the company sucked big hairy balls, and I value my vision.
Keyboard: MS Natural, what else.
Mouse: Logitech TrackMan Marble FX. It's too bad Logitech doesn't make this one anymore. I had severe wrist pain until I switched to it.
Desk - adjustable height. That's about the only important thing.
Chair - a shitty one right now (company won't buy a good one). I'm gonna buy a decent chair for myself, the one I use right now fucks up my back.
Which is what I think this technology is. It simply will not work where I live for example. There are folks on almost every channel (while they should be only on channels 1,6 and 11), and even single-channel wireless gets tricky.
I heard a nuclear missile launched from Russia will only take 9 minutes to get to the US. Did you guys know that during the cold war both countries kept their ICBMs targeted at each other all the time? That there was a plan that in the case of nuclear war nuclear warheads would get blown up over the Missisipi river sequentially to cause a giant wave and amplify devastation? I'm sure US had similar plans, maybe not related to rivers.
This is a dangerous game these fellas are playing. I'd rather see them concentrate on nonproliferation efforts than on brandishing weapons and flexing muscle, because all it takes is one madman that gets elected as president in a country that has nuclear weapons. Press a few buttons, break an ampule with cyanide, you won't be around 18 minutes later when response comes in.
Here's a simple problem for you. On a Fedora Core 4 system with SELinux enabled, configure Samba access on a directory that can be also accessed by Apache without disabling SELinux for either. Should be easy, right?
Turns out, you can't do this! For someone coming from a system with proper implementation of ACLs (such as Windows or Mac OS X) this is just unbelievable. And folks who wrote SELinux say this is by design, because if you give access to the same data to two apps, you can't isolate these apps. WTF? Do I want to isolate these apps? No, I simply want to be able to copy pictures to the webserver in my closet. If I can't do something as trivial as this, then folks who actually run serious stuff on Linux are bound to have problems too. My guess is, the first thing folks do is they disable SELinux and move on. That's what I did, eventually.
I've lived in Russia for quite a while, and it's funny how Russians perceive themselves, especially when it comes to the military. Everything is "world's first", everything is said to "have no equal in the world", and everyone believes it, whether or not it's true. So I'd take this "world's first" thing with a three-pound grain of salt if it comes from Russians.
That was just a little probing. After that he'd have to do a couple phone interviews, and after phone interviews (if he passes them) he'd have to go through a six-hour MS interview gauntlet.
Job offer comes if you successfully complete a six-hour interview, and he'd probably be rejected during phone interview without further consideration. I seriously envy the guy. He thinks so much about himself. Humility be damned.
>> The few random vulnerabilitys that have even made headlines have been snuffed out >> in a week or two by Apple themselves in Security Updates
The thing is, by the time things appear in headlines it's WAY TOO LATE to fix them. Most of Microsoft vulns that have appeared in the headlines were fixed by Microsoft MONTHS before, it's just that folks are not diligent enough about installing patches.
No, it's not because of me. I make every effort to help these poor folks in India isofar as this help does not affect my own schedule and performance. They _demand_ help. They're not like your typical FTE who would just come to the team, sit in his office for two weeks, swearing like a truck driver and start producing solid code. These fellas require handholding at everything. Devs in the US are between a rock and a hard place. If you help - they grow dependent on you and want you to do work for them (which ultimately leads to either overtime or schedule impact), if you don't - you get a lower review score because of your "teamwork skills".
There's another problem as well - program management. You see, program managers are a stupid bunch. They like to formulate their requirements fully only towards the end of the product cycle. Their communication skills often leave much to be desired, too. Guess what, this DOES NOT WORK if you outsource dev work. No issue can be resolved in less than 24 hours because of time difference. Even if you do resolve it over email, there's no guarantee Indian folks "got your point". If your dev is next door, you just walk into his office and draw shit on the whiteboard. If your dev is in India, you're massively out of luck.
I could bring up a dozen more issues explaining why outsourcing _development_ is a good idea, and outsourcing _pm_ is even worse. But I think that would be wasted effort. You're probably in Hyderabad or Bangalore right now, sipping chai and thinking that Americans are lazy and stupid, and your IIT diploma is better than the one from Stanford.:-)
Now try the same for Unicode (i.e. non English) strings. Your lower(x) doesn't work for them.
The first approach suggested by you has horrible select performance and leads to full table scans (that's on top of not working with international chars).
The second approach screws up INSERT performance, which would be OK if it worked well with Unicode strings. Which it doesn't.
So next time before you unleash your righteous criticism on someone, do some hands-on testing.
One thing to consider - collations and Unicode support. Believe it or not, folks, Postgres does NOT support case-insensitive string comparisons. Or, more exactly it does, but you end up doing full table scan and converting everything into upper/lowercase, which is not an option on all but the smallest of the datasets. And even converting to upper/lowercase is a BIG problem for PostgreSQL, because it's UNICODE support is quite poor. So if your project has even remote possibility of using non-English textual data in lookups, steer clear of PostgreSQL.
There's a discussion about including support for IBM ICU, but as of right now there's no proper collations/unicode support in PgSQL, aside from storing character data in UTF-8.
There aren't that many "best and brightest", you see. MSR China has been around for a looong while and I have yet to see anything significant to come out of it. As far as I'm concerned, as long as groups can be separated and as long as what MSFT gets back is state of the art software - go ahead, ship it over there. However, I have yet to see a top notch piece of code developed in PRC or India.
Over the entire dot-com "bust" there wasn't a single big layoff at Microsoft. Even when people are laid off, they're laid off in small groups and offered an opportunity to find a position elsewhere within he Co. As of right now I'm not aware of anyone technical who could not find such a position. There are also thousands of open positions, both here and in India/PRC.
And if you're GOOD, you have nohing to be afraid of at Microsoft, not for at least the next five years. Because there's a shortage of people who are GOOD.
Not all of it, either. Mostly manual test, requiring little to no skill. For this kind of test, it sure would be a waste to pay someone in Redmond $60-70K/yr. Automated test, infrastructure, security, perf/stres, and all other critical test remains in Redmond so far.
What was said above was only about China, though. Indian insiders seem to push real hard to get not only testing, but also development and program management to India. However, since they aren't (yet) an overwhelmin majority here, only low-impact work items go to India, so that if folks in India fuck it up (and they often do), we could fix the situation without slipping the schedule much. Overall, I'd say "split" development leads to worse code quality, but it's still a lot better than if all of it was developed in India.
In Russia, there are a few popular search engines, the most popular being Yandex ("Ya" is the last letter in the Russian alphabet and also a word meaning "I", so it's Index with "Ya" instead of "I"). It has tons of free services, it has paid "ad words" style advertising, and most importantly its spiders are optimized for Russian sites. For Russian language searches it's simply BETTER than Google, believe it or not. If Yandex doesn't find it, Google is used as a last resort.
How about just watching your freaking kids using the freaking pool? What if pool owner installs this $100K+ system and it fails to react to a drowning kid? What if no one is available to rescue the kid? What if the potential rescuer is also a poor swimmer?
There are thousands of "what ifs" here. The point is, watch after your kids until they're smart enough to watch after themselves (about 20-21 years or so). This is coming from a person who had a severe trauma at 1.5 years of age due to parents not watching.
Spending hundred thousand dollars is not a reason to be careless enough to let your kid (or friend) drown in the pool.
Slashdot Mirror
Dev machine #1: P4-3.8GHz, 64bit one, with 3GB of RAM running about three virtual machines (MS Virtual Server), one SQL, one WinXP to test the UI, one runs apps. OS is W2K3 Enterprise (a requirement for the stuff I develop).
Dev machine #2: Dual Xeon 3.02GHz with 2GB of RAM. No VMs, no nothing, just a subset of apps that can't run on machine #1, and SQL server.
Monitor: Dell 2405FPW - bought with my own money. The one provided to me by the company sucked big hairy balls, and I value my vision.
Keyboard: MS Natural, what else.
Mouse: Logitech TrackMan Marble FX. It's too bad Logitech doesn't make this one anymore. I had severe wrist pain until I switched to it.
Desk - adjustable height. That's about the only important thing.
Chair - a shitty one right now (company won't buy a good one). I'm gonna buy a decent chair for myself, the one I use right now fucks up my back.
Which is what I think this technology is. It simply will not work where I live for example. There are folks on almost every channel (while they should be only on channels 1,6 and 11), and even single-channel wireless gets tricky.
>> To recap in manager-speak, you're saying the utilization of utilize nets suboptimal perceptual leadership utility?
No, it simply doesn't leverage any paradigm leading to go-to market, customer focused synergies.
"Utilize" would be "use" if you were speaking plain English. And no, using the word "utilize" doesn't make you seem any smarter.
I heard a nuclear missile launched from Russia will only take 9 minutes to get to the US. Did you guys know that during the cold war both countries kept their ICBMs targeted at each other all the time? That there was a plan that in the case of nuclear war nuclear warheads would get blown up over the Missisipi river sequentially to cause a giant wave and amplify devastation? I'm sure US had similar plans, maybe not related to rivers.
This is a dangerous game these fellas are playing. I'd rather see them concentrate on nonproliferation efforts than on brandishing weapons and flexing muscle, because all it takes is one madman that gets elected as president in a country that has nuclear weapons. Press a few buttons, break an ampule with cyanide, you won't be around 18 minutes later when response comes in.
Here's a simple problem for you. On a Fedora Core 4 system with SELinux enabled, configure Samba access on a directory that can be also accessed by Apache without disabling SELinux for either. Should be easy, right?
Turns out, you can't do this! For someone coming from a system with proper implementation of ACLs (such as Windows or Mac OS X) this is just unbelievable. And folks who wrote SELinux say this is by design, because if you give access to the same data to two apps, you can't isolate these apps. WTF? Do I want to isolate these apps? No, I simply want to be able to copy pictures to the webserver in my closet. If I can't do something as trivial as this, then folks who actually run serious stuff on Linux are bound to have problems too. My guess is, the first thing folks do is they disable SELinux and move on. That's what I did, eventually.
Except different models of iPods are DIFFERENT. They aren't just the same model with different parts of it disabled in software.
I've lived in Russia for quite a while, and it's funny how Russians perceive themselves, especially when it comes to the military. Everything is "world's first", everything is said to "have no equal in the world", and everyone believes it, whether or not it's true. So I'd take this "world's first" thing with a three-pound grain of salt if it comes from Russians.
That was just a little probing. After that he'd have to do a couple phone interviews, and after phone interviews (if he passes them) he'd have to go through a six-hour MS interview gauntlet.
Job offer comes if you successfully complete a six-hour interview, and he'd probably be rejected during phone interview without further consideration. I seriously envy the guy. He thinks so much about himself. Humility be damned.
>> The few random vulnerabilitys that have even made headlines have been snuffed out
>> in a week or two by Apple themselves in Security Updates
The thing is, by the time things appear in headlines it's WAY TOO LATE to fix them. Most of Microsoft vulns that have appeared in the headlines were fixed by Microsoft MONTHS before, it's just that folks are not diligent enough about installing patches.
HTML rendering is WAY faster. No other visible differences, though.
digg.com rules.
No, it's not because of me. I make every effort to help these poor folks in India isofar as this help does not affect my own schedule and performance. They _demand_ help. They're not like your typical FTE who would just come to the team, sit in his office for two weeks, swearing like a truck driver and start producing solid code. These fellas require handholding at everything. Devs in the US are between a rock and a hard place. If you help - they grow dependent on you and want you to do work for them (which ultimately leads to either overtime or schedule impact), if you don't - you get a lower review score because of your "teamwork skills".
:-)
There's another problem as well - program management. You see, program managers are a stupid bunch. They like to formulate their requirements fully only towards the end of the product cycle. Their communication skills often leave much to be desired, too. Guess what, this DOES NOT WORK if you outsource dev work. No issue can be resolved in less than 24 hours because of time difference. Even if you do resolve it over email, there's no guarantee Indian folks "got your point". If your dev is next door, you just walk into his office and draw shit on the whiteboard. If your dev is in India, you're massively out of luck.
I could bring up a dozen more issues explaining why outsourcing _development_ is a good idea, and outsourcing _pm_ is even worse. But I think that would be wasted effort. You're probably in Hyderabad or Bangalore right now, sipping chai and thinking that Americans are lazy and stupid, and your IIT diploma is better than the one from Stanford.
Try functional indexes with Unicode strings, then come back and we'll talk. Until then, shut the fuck up.
It's NOT fine, neither on Windows, nor on Linux. I've tried both just two weeks ago. Try it yourself.
I know this, smartass.
Now try the same for Unicode (i.e. non English) strings. Your lower(x) doesn't work for them.
The first approach suggested by you has horrible select performance and leads to full table scans (that's on top of not working with international chars).
The second approach screws up INSERT performance, which would be OK if it worked well with Unicode strings. Which it doesn't.
So next time before you unleash your righteous criticism on someone, do some hands-on testing.
One thing to consider - collations and Unicode support. Believe it or not, folks, Postgres does NOT support case-insensitive string comparisons. Or, more exactly it does, but you end up doing full table scan and converting everything into upper/lowercase, which is not an option on all but the smallest of the datasets. And even converting to upper/lowercase is a BIG problem for PostgreSQL, because it's UNICODE support is quite poor. So if your project has even remote possibility of using non-English textual data in lookups, steer clear of PostgreSQL.
There's a discussion about including support for IBM ICU, but as of right now there's no proper collations/unicode support in PgSQL, aside from storing character data in UTF-8.
MySQL is much better in this regard.
>> fallucious
:-)
Apparently you're not as smart as you think.
There aren't that many "best and brightest", you see. MSR China has been around for a looong while and I have yet to see anything significant to come out of it. As far as I'm concerned, as long as groups can be separated and as long as what MSFT gets back is state of the art software - go ahead, ship it over there. However, I have yet to see a top notch piece of code developed in PRC or India.
Over the entire dot-com "bust" there wasn't a single big layoff at Microsoft. Even when people are laid off, they're laid off in small groups and offered an opportunity to find a position elsewhere within he Co. As of right now I'm not aware of anyone technical who could not find such a position. There are also thousands of open positions, both here and in India/PRC.
And if you're GOOD, you have nohing to be afraid of at Microsoft, not for at least the next five years. Because there's a shortage of people who are GOOD.
Not all of it, either. Mostly manual test, requiring little to no skill. For this kind of test, it sure would be a waste to pay someone in Redmond $60-70K/yr. Automated test, infrastructure, security, perf/stres, and all other critical test remains in Redmond so far.
What was said above was only about China, though. Indian insiders seem to push real hard to get not only testing, but also development and program management to India. However, since they aren't (yet) an overwhelmin majority here, only low-impact work items go to India, so that if folks in India fuck it up (and they often do), we could fix the situation without slipping the schedule much. Overall, I'd say "split" development leads to worse code quality, but it's still a lot better than if all of it was developed in India.
In Russia, there are a few popular search engines, the most popular being Yandex ("Ya" is the last letter in the Russian alphabet and also a word meaning "I", so it's Index with "Ya" instead of "I"). It has tons of free services, it has paid "ad words" style advertising, and most importantly its spiders are optimized for Russian sites. For Russian language searches it's simply BETTER than Google, believe it or not. If Yandex doesn't find it, Google is used as a last resort.
Just use 7zip. http://www.7-zip.org/ Or if you're inclined to spend a little bit of money, get WinRar.
If this serves as an excuse for parents to not watch their kids, there will be more harm than good.
How about just watching your freaking kids using the freaking pool? What if pool owner installs this $100K+ system and it fails to react to a drowning kid? What if no one is available to rescue the kid? What if the potential rescuer is also a poor swimmer?
There are thousands of "what ifs" here. The point is, watch after your kids until they're smart enough to watch after themselves (about 20-21 years or so). This is coming from a person who had a severe trauma at 1.5 years of age due to parents not watching.
Spending hundred thousand dollars is not a reason to be careless enough to let your kid (or friend) drown in the pool.
But will it enable customer driven, lean, agile XML based go-to market paradigms and rapid value driven innovation? That's the question.