Pretty soon, leasing a "dedicated server" will be simply leasing a dedicated "instance" of a server, for lack of a better word. Though not nearly the isolation of a true VM, BSD jails have been usable for this for quite some time, I remember seeing this for hosting companies back when FreeBSD 4.2 came out.
It isn't. It's new, called UIMA. We eval'ed this under an NDA. I's pretty cool, though what we saw was an SDK more than a package you can install. it is more of an infrastructure, so can be used to create new engines.
Richard Feynman was one of the people who investigated the first shuttle disaster, and as a pain in the ass cantankerous old coot, really didn't care about standard Washington procedures and really got to the core of the matter. He cronicles a lot of it in What Do You Care What Other People Think?, ISBN: 0393320928 (get it from wherever, no Amazon kickbacks here). A very interesting read, I ended up reading it right after the second shuttle disaster, and thought that a lot of the human problems that caused the first blow up could be fingered in the second.
If you haven't read Feynman before, you'll probably like him. Funny guy, pretty damn smart, and managed with luck, brains, skill and stubbornness to get in the middle of some of the biggest science in the last century.
My mistake, article was (though unclear) talking more about Virtual Machines than Virtual Memory, I guess I'm too frustrated at having the occasional kernel panic...
Not sure if that was a real post or meant to be humor, but mucking with the VM is not the way to be more stable. There were major stability issues in the early 2.4 kernels due to the new VM that was swapped in. My company is still dealing with bugs in what can be seen as a VM issue - the 4Gb user/4Gb kernel split in RedHat advanced server, and this is in the new U5 update.
My turn for a rant. I have DBC DSL in chicago. Originally, i wanted it on my house line, but SBC coudn't get my house line installed in time. I'm on call tech support, so work provided a second line with DSL on it. It was ordered a week after my main line, but still got installed before my main line did. Whatever, it's free. Fast forward 12 months, my work is no longer paying for the second DSL line. Fine, was nice to get it for free, but I can pay, no problem.
1) I get to pay the switching charge since they have to send a tech to come to my apartment building to make the switch. I don't have multiple circuits to my phone "port", they have to send a guy to the punchdown block in the boiler room.
2) I say I don't need a modem, since i have one thta's just a few months old. They don't give me one, but they manage to bill me for it. I still have to call and have it removed from my bill.
3) They never gave me a modem, nor any info on getting my account set up. I had to get to second level tech support (which was surprisingly good, though he treated me like an idiot at times). Took a day or so from the install to get me logged on. As a funny aside, they refused my first login name initially. My last name is Homolka, and they must be filtering out "homo".
All of this doesn't sound really bad but my bigger issue is their support lines. Every time I call they ask "we'd like you to be satisfied, are you satisfied?" and I say no, and ist out my complaints. Once someone said "well, maybe they'll check this call in training and do something for you". you mean, you can't trigger a look? You're that powerless?
I'd go toother DSL, but in some ways, I'm still paying to SBC since they own the lines. Sometimes I want to do VOIP and ditch SBC, but i need broadband, so it's either SBC (underneath) or Comcast, which also sucks (my cable is out today BTW).
Me - Guy whose gf has no idea why I read "slashthing" 5 times a day.
Acctually, this isn't that hard. Most compilers have some limit to the number of errors that they spit out, the general though being "if you got this many errors, you missed something early on and put your code into an uparseable state - fix it and we should be good". Sometimes missing a semicolon is enough to get some weird parsing state where everything else after is gobbledegook. Anything involving templates, such as the STL, is probably an easy source as well.
Troll "ripped from the headlines" (Law and Order style)... An Air France flight rolled off the runway in Toronto, on its way to Paris. No deaths if I remember.
So it's MS Windows - is that 95, 98, ME, NT3.51, NT4, win2k, XP, 2003? All have their quirks, advantages, disadvantes, and users always hide their start menu entries in different spots if they've had the thing for a while.
Was always fun to have to install a newer version of IE to get some kind of common control. Back in the day I needed to install a new version of IE to get a (16bit) properties dialog to work the way I needed. The "single windows UI" is a myth.
JBoss is not the only open source Appserver. Geronimo is moving quickly (recently passing the has passed the J2EE 1.4.1, on it's way to full J2EE compliance). IBM is providing support, a lot of it by donating Gluecode and other code to the apache project, and even BEA (even though they still sell WebLogic) contributing code to the Apache project, allowing their tools to target geronimo.
it's not super old in other places. I work in a downtown chicago building that was finished 2 years ago and it shows. I don't think the pic that I see on MickeySoft is more than 3 years old, I don't see much construction, and I remember when this building was going up (I was downtime a lot before I ever worked here). It's inconsistent, and maybe that's more dangerous, it might be new, might be old, you jsut don't know.
This would require you to enforce it for all modules as well. Custom third party modules have a variety of ways to parse the config file ranging from simple "apache API will give me these values" up to "here is a file stream and a location in it, we trust you to stop parsing when you see the end of your config, so have at it". Going to a strict XML parsing would remove current functionality, and make it harder for third party developers, one of apache's strengths.
I always liked seeing Rice Boy cars in my neighborhood (nothing against asians, I'm dating one, just the cars are all tricked out Civics rather than clapped out Z28s). Weird thing is, most of the mods you see on the outside hurt performance.
Huge rear wing
In a front wheel driver, magnifies understeer, hurting cornering balance and abilities. Needs to be tuned to car aerodynamics, or just adds drag for no reason.
Painted brake drums
Aside from the fact that no real performance cars should have the disc/drum combo (too difficult to modulate 2 separate braking curves) the paint retards head transfer, increasing brake fade.
Huge exhaust pipe
Exhaust performance is about tuning, not size. If your pipe is too wide, it doesn't scavenge right, increasing backpressure, lowering engine performance. You need the right size, airflow speed is critical here
Flashy colored interior
kicks up too many interior reflections. there's a reason racing cars are all flat black inside.
probably some others, but that's my rant for the day.
We have issues with corporate deployments as well, mostly for the same reasons. No way to lock things down. we're a hedge fund, and we also have a group policy to empty IE cache every week (to keep prying eyes from seeing what we're researching). FireFox has none of that. Not saying that it should, i doubt having all this stuff is their priority right now. Just saying that IE has some features that FireFox doesn't have even on a roadmap, and these are features that kicked it out the door.
Hmm, and they stopped issuing ZIPfile downloads, meaning I have to install every time. Pain in the ass.
This is a bad argument. It's a brand new chipset that's supposed to be a clean break from ia32, whereas AMD did the more evolutionary approach in creating x86-64 (now generally renamed AMD64). Do you congratulate Motorola for it's input in PowerPC creating a new desktop class RISC chip, or do you fault them for creating something incompatible with 68K. THere are sometimes you need to break compatibility
Itanium was supposed to be the new base chip, the thing to finally replace x86. Picture high performance and high volume, the other RISC guys weren't supposed to be able to compete. SGI and DEC cowered, shutting down ALPHA and MIPS.
HP decied they wanted in on Itanium early. They partnered with Intel on chip design. Intel designed the first chip, while HP's was the much more highly regarded Itanium II. They bet the farm on it, in some ways more than Intel itself, phasing out PA-RISC, and now trying to force people to move off of Vax and Tandem onto Itanium. They pretty much have no choice, having fired the other processor guys they had.
If you want to hear about respect, how about HP, the #1 Itanium vendor, killing off workstation class Itanium computers. If this isn't a white flag, resigning the chip into a nice market, i'm not sure what other interpretation there can be.
True, Intel needed to make the change at sometime to 64 bit, their Pentium lines only doing 32bit will eventually be dropped by everybody as 64bit becomes a standard. Intel is currently shipping Pentium 4's with AMD64 extensions, renamed EM64T. We have a box or two in house, waiting for someone with time to recompile our base libs in 64 bit.
Its amazing so many architectures are now powered by the same chips Not so weird, at one time Motorla 68000 powered everything, Sun, SGI, NeXT, Apple Mac and LaserWriter alike. Then RISC came, and everyone went their own separate ways.
Some of it is to get some lessons from ANSI C++: C++ style comments (most compilers accept already, make it part of the standard) declarations anywhere, not just at code block start. The new C++ style scoped for. New data types like _Bool, _Complex, _Imaginary Some others probably....
The biggest new thing is probably the restrict keyword. First, any new keyword has the possibility of breaking old code, so once could say that C0x is not fully backwards compatible with all former code.
Then actually getting yrou head around it. This is a new keyword to help give information about pointers and what they may alias (multiple pointers to the same variable). Compilers have difficulty with certain optimizations becasue they can never guarantee aliasing. With restrict, you're trying to give it hints. I'm sure this will be a cause of bugs in code for years.
1) prisoner's dilemma, nice. I first heard this talking about morality, weird to know the moral choice was to not admit to your crime. 2) expected value for coin flip > 1, always take it ((50% * +50) + (50% * -25)) 3) flame wars...
A long time ago, in a Galaxy far far away, I actually had Word 6.0 running on Solaris 2.6/SPARC, under WABI, which provided a Windows 3.1 environment. It worked slowly, but other than that, pretty well. In fact, I got a macro virus, Concept1. Was weird for the guy running Solaris to be able to identify to the company (all 8 of us) that we had a Word macro virus. They started to scrub the fileservers and eventually traced it back to a Word doc from a vendor.
Who says microsoft doesn't do cross platform, I got a cross platform virus people, years before java.
Slashdot Mirror
I used it - to my horror. It was difficult to use man pages on the system, how do you screw up man pages?
Pretty soon, leasing a "dedicated server" will be simply leasing a dedicated "instance" of a server, for lack of a better word.
Though not nearly the isolation of a true VM, BSD jails have been usable for this for quite some time, I remember seeing this for hosting companies back when FreeBSD 4.2 came out.
It isn't. It's new, called UIMA. We eval'ed this under an NDA. I's pretty cool, though what we saw was an SDK more than a package you can install. it is more of an infrastructure, so can be used to create new engines.
Richard Feynman was one of the people who investigated the first shuttle disaster, and as a pain in the ass cantankerous old coot, really didn't care about standard Washington procedures and really got to the core of the matter. He cronicles a lot of it in What Do You Care What Other People Think?, ISBN: 0393320928 (get it from wherever, no Amazon kickbacks here). A very interesting read, I ended up reading it right after the second shuttle disaster, and thought that a lot of the human problems that caused the first blow up could be fingered in the second.
If you haven't read Feynman before, you'll probably like him. Funny guy, pretty damn smart, and managed with luck, brains, skill and stubbornness to get in the middle of some of the biggest science in the last century.
My mistake, article was (though unclear) talking more about Virtual Machines than Virtual Memory, I guess I'm too frustrated at having the occasional kernel panic...
Not sure if that was a real post or meant to be humor, but mucking with the VM is not the way to be more stable. There were major stability issues in the early 2.4 kernels due to the new VM that was swapped in. My company is still dealing with bugs in what can be seen as a VM issue - the 4Gb user/4Gb kernel split in RedHat advanced server, and this is in the new U5 update.
My turn for a rant. I have DBC DSL in chicago. Originally, i wanted it on my house line, but SBC coudn't get my house line installed in time. I'm on call tech support, so work provided a second line with DSL on it. It was ordered a week after my main line, but still got installed before my main line did. Whatever, it's free. Fast forward 12 months, my work is no longer paying for the second DSL line. Fine, was nice to get it for free, but I can pay, no problem.
1) I get to pay the switching charge since they have to send a tech to come to my apartment building to make the switch. I don't have multiple circuits to my phone "port", they have to send a guy to the punchdown block in the boiler room.
2) I say I don't need a modem, since i have one thta's just a few months old. They don't give me one, but they manage to bill me for it. I still have to call and have it removed from my bill.
3) They never gave me a modem, nor any info on getting my account set up. I had to get to second level tech support (which was surprisingly good, though he treated me like an idiot at times). Took a day or so from the install to get me logged on.
As a funny aside, they refused my first login name initially. My last name is Homolka, and they must be filtering out "homo".
All of this doesn't sound really bad but my bigger issue is their support lines. Every time I call they ask "we'd like you to be satisfied, are you satisfied?" and I say no, and ist out my complaints. Once someone said "well, maybe they'll check this call in training and do something for you". you mean, you can't trigger a look? You're that powerless?
I'd go toother DSL, but in some ways, I'm still paying to SBC since they own the lines. Sometimes I want to do VOIP and ditch SBC, but i need broadband, so it's either SBC (underneath) or Comcast, which also sucks (my cable is out today BTW).
Me - Guy whose gf has no idea why I read "slashthing" 5 times a day.
Acctually, this isn't that hard. Most compilers have some limit to the number of errors that they spit out, the general though being "if you got this many errors, you missed something early on and put your code into an uparseable state - fix it and we should be good". Sometimes missing a semicolon is enough to get some weird parsing state where everything else after is gobbledegook. Anything involving templates, such as the STL, is probably an easy source as well.
There can be a general assumption that someone regularly reding slashdot does not have a girlfriend...
-- Me, guy who's gf has no idea why i need to check Slashdot 5 times a day...
Troll "ripped from the headlines" (Law and Order style)... An Air France flight rolled off the runway in Toronto, on its way to Paris. No deaths if I remember.
From the BBC: http://news.bbc.co.uk/2/hi/americas/4740539.stm
dot maybe?
So it's MS Windows - is that 95, 98, ME, NT3.51, NT4, win2k, XP, 2003? All have their quirks, advantages, disadvantes, and users always hide their start menu entries in different spots if they've had the thing for a while.
Was always fun to have to install a newer version of IE to get some kind of common control. Back in the day I needed to install a new version of IE to get a (16bit) properties dialog to work the way I needed. The "single windows UI" is a myth.
JBoss is not the only open source Appserver. Geronimo is moving quickly (recently passing the has passed the J2EE 1.4.1, on it's way to full J2EE compliance). IBM is providing support, a lot of it by donating Gluecode and other code to the apache project, and even BEA (even though they still sell WebLogic) contributing code to the Apache project, allowing their tools to target geronimo.
it's not super old in other places. I work in a downtown chicago building that was finished 2 years ago and it shows. I don't think the pic that I see on MickeySoft is more than 3 years old, I don't see much construction, and I remember when this building was going up (I was downtime a lot before I ever worked here). It's inconsistent, and maybe that's more dangerous, it might be new, might be old, you jsut don't know.
This would require you to enforce it for all modules as well. Custom third party modules have a variety of ways to parse the config file ranging from simple "apache API will give me these values" up to "here is a file stream and a location in it, we trust you to stop parsing when you see the end of your config, so have at it". Going to a strict XML parsing would remove current functionality, and make it harder for third party developers, one of apache's strengths.
Never saw dodgeball, I thought of Herbie (one film they called him Ocho, 5 + 3)
In a front wheel driver, magnifies understeer, hurting cornering balance and abilities. Needs to be tuned to car aerodynamics, or just adds drag for no reason.
Aside from the fact that no real performance cars should have the disc/drum combo (too difficult to modulate 2 separate braking curves) the paint retards head transfer, increasing brake fade.
Exhaust performance is about tuning, not size. If your pipe is too wide, it doesn't scavenge right, increasing backpressure, lowering engine performance. You need the right size, airflow speed is critical here
kicks up too many interior reflections. there's a reason racing cars are all flat black inside.
probably some others, but that's my rant for the day.
We have issues with corporate deployments as well, mostly for the same reasons. No way to lock things down. we're a hedge fund, and we also have a group policy to empty IE cache every week (to keep prying eyes from seeing what we're researching). FireFox has none of that. Not saying that it should, i doubt having all this stuff is their priority right now. Just saying that IE has some features that FireFox doesn't have even on a roadmap, and these are features that kicked it out the door.
Hmm, and they stopped issuing ZIPfile downloads, meaning I have to install every time. Pain in the ass.
This is a bad argument. It's a brand new chipset that's supposed to be a clean break from ia32, whereas AMD did the more evolutionary approach in creating x86-64 (now generally renamed AMD64). Do you congratulate Motorola for it's input in PowerPC creating a new desktop class RISC chip, or do you fault them for creating something incompatible with 68K. THere are sometimes you need to break compatibility
Not so much respect as a "we have to" thing.
Itanium was supposed to be the new base chip, the thing to finally replace x86. Picture high performance and high volume, the other RISC guys weren't supposed to be able to compete. SGI and DEC cowered, shutting down ALPHA and MIPS.
HP decied they wanted in on Itanium early. They partnered with Intel on chip design. Intel designed the first chip, while HP's was the much more highly regarded Itanium II. They bet the farm on it, in some ways more than Intel itself, phasing out PA-RISC, and now trying to force people to move off of Vax and Tandem onto Itanium. They pretty much have no choice, having fired the other processor guys they had.
If you want to hear about respect, how about HP, the #1 Itanium vendor, killing off workstation class Itanium computers. If this isn't a white flag, resigning the chip into a nice market, i'm not sure what other interpretation there can be.
True, Intel needed to make the change at sometime to 64 bit, their Pentium lines only doing 32bit will eventually be dropped by everybody as 64bit becomes a standard.
Intel is currently shipping Pentium 4's with AMD64 extensions, renamed EM64T. We have a box or two in house, waiting for someone with time to recompile our base libs in 64 bit.
Its amazing so many architectures are now powered by the same chips
Not so weird, at one time Motorla 68000 powered everything, Sun, SGI, NeXT, Apple Mac and LaserWriter alike. Then RISC came, and everyone went their own separate ways.
Some of it is to get some lessons from ANSI C++:
C++ style comments (most compilers accept already, make it part of the standard)
declarations anywhere, not just at code block start.
The new C++ style scoped for.
New data types like _Bool, _Complex, _Imaginary
Some others probably....
The biggest new thing is probably the restrict keyword. First, any new keyword has the possibility of breaking old code, so once could say that C0x is not fully backwards compatible with all former code.
Then actually getting yrou head around it. This is a new keyword to help give information about pointers and what they may alias (multiple pointers to the same variable). Compilers have difficulty with certain optimizations becasue they can never guarantee aliasing. With restrict, you're trying to give it hints. I'm sure this will be a cause of bugs in code for years.
1) prisoner's dilemma, nice. I first heard this talking about morality, weird to know the moral choice was to not admit to your crime.
2) expected value for coin flip > 1, always take it ((50% * +50) + (50% * -25))
3) flame wars...
A long time ago, in a Galaxy far far away, I actually had Word 6.0 running on Solaris 2.6/SPARC, under WABI, which provided a Windows 3.1 environment. It worked slowly, but other than that, pretty well. In fact, I got a macro virus, Concept1. Was weird for the guy running Solaris to be able to identify to the company (all 8 of us) that we had a Word macro virus. They started to scrub the fileservers and eventually traced it back to a Word doc from a vendor.
Who says microsoft doesn't do cross platform, I got a cross platform virus people, years before java.