Think of the overclockers - they crank up the speed and voltages, add cooling to compensate, and when the cooling can't handle the load (the system gets too hot) it becomes unstable. It doesn't necessarily die, but stability becomes unreliable (it generally locks up, sometimes dies.)
Ball-bearing fans, balance fans blowing in and fans blowing out, and my favorite is a twin fan exhaust in the top 5.25" bay of the system like this - note that this is the first url that Google found that looks like the one I'm talking about, I don't necessarily recommend this specific vendor.
If uptime (reliability) is your concern then crank up the airflow. Your CPU will last forever or until the CPU fan dies, whichever comes first. It is not something I can draw here in/. but Google the bathtub curve of hardware failures, and understand that as the system internal temperature (just the temp in the case, not on the CPU) goes up the shorter the time between sides of the bathtub curve are spread apart.
There are two types of downtime - planned and unplanned. Server death is the worst kind of unplanned downtime.
I don't see where it would go any faster, honestly.
Look at the charts - when they are not using AA/Oversampling the single 6600GT goes just as fast as the 3D1 and the 6800GT - at all resolutions. The applications are CPU bound and better video isn't going to change that.
Granted there is always the 1337 crew pointing at the 8xOversampling / 4xAA numbers - but quite honestly the FPS hit going there on any card isn't worth the questionable increase in quality. Given the 60Hz an LCD puts out - and given that it takes a $600 LCD to run 1600x1200 - I'd say that quite honestly these cards are pretty much the same. Hardware has once again outrun software, and right now videocards are processing data faster than the CPU can serve it up. This 3D1 is cool, and it is going to take the early adopters to buy it in order to finance the second generation of the technology (ie, runs on other random motherboards) but I think I'm going to pass.
Maybe the guy working on it is color-blind. Don't laugh - we had a guy in our office that was color blind, none of us knew it until we let him set up Windows 3.0's color scheme to best suit his needs (ok it was a long time ago.) Made the Hotdog Stand color scheme look mild in comparison - it was frightful to us, but the contrast worked great for him.
I did totally unscientific set of benchmarks on these machines (see the link in my sig, goes to my Journal).
I was almost able to saturate the GigE pipe outbound (116MB/s outbound on a theoretical 125MB/s peak wire speed.)
Pure memory reads (ramdrive) pegged at about 520MB/s (.5GBps or over 4Gbps.)
I keep hearing that the higher end boxes are faster but... is it really so? I mean that with the same CPU, same kinds of memory, same hard drives, same PCI architecture, same GigE network card, same... everything. The only things I could imagine going faster would be many different connections all requesting different transactions, or ??? but really the only variables I can see are motherboard (how would that increase throughput in any significant level) and drive array controllers (which I can see having a performance boost but in theory I could put the same drive controller in a low end box?)
Good points - I inferred but didn't actually state the environment I manage with this philosophy - exactly the size environment the OP is describing. Small / home office environment in which the users are all perfectly capable of building machines from the ground up, managing the network, adding users or printers themselves.
The machines I use are basic, entry level boxes (the 400sc was their entry level box, recently supplanted by the 420sc) and are little more than nicely apportioned desktops with crap video cards and Intel chipset motherboards (including the integrated GigE.) They are expandable enough for my needs (DVD burner or tape drive, plus up to 5 hard drives internally if you get creative (using the space designed for a floppy drive to put more hard drives, ditto the second 5.25" bay.)) I'm not buying hardware capable of running corporate Fortune 500 business, but at an average of $350 apiece they are plenty for my needs.
I actually picked up a few 400sc boxes (40G drives, 256M memory) for $300 apiece - isn't that the price point of the XBox when it first came out (and everybody went freaky trying to convert them to run desktop software / Linux)?
I don't look at Dell as a solution provider. I look at Dell as a whitebox dealer that happens to send me the machines pre-built using the same hardware in each machine, keeps a product line long enough that I can get more of the same later, keeps a fairly nice stash of drivers for all the hardware in one place, ships it for free, and sends me a replacement drive via personal messenger (on-site service) if a drive fails.
I don't expect the service levels on a $400 box to be good regardless of who I buy it from so I guess that's why I haven't been disappointed by Dell tech support yet. They do give a pretty good recount of the hardware going into the machines. Also their web support is good enough to get all the drivers in one place, and if you put in the service tag number it actually bundles only the drivers associated with the hardware in your machine on one page.
I'm not sure what you mean about the non-glamorous busses - I'm not real happy with the lack of an AGP slot in the machine that replaced the 400sc (the 420sc has PCI-Express) but that's only because I also liked slapping in an AGP card to make a nice cheap desktop (not a concern if we are really talking about servers.)
And I laughingly agree about their 'discounts' - watch their web site for a few weeks to catch a much better price. Opportunity is your friend, and 'have to buy today' is not.
That said - if the OP is going to build his own, he is going to build his own. If that's the case, here is the most important bit of wisdom I have to offer:
Heat Kills.
Computers die for one reason, and one reason only. Heat. I have autopsied well over 100 dead computers (servers, desktops) in the past two decades and with extremely few exceptions the deaths were caused by CPU fan failure (a $10 part), power supply fan failure (a $10 part), or hard drive failure. If I really wanted to push it, I would say the drive failures were heat related, generally due to weak power supply exhaust fans.
Heat kills. It doesn't matter where you buy your parts if you are building your own systems (hey, we all did it early in our careers when money was more scarce than time) then spend the extra eight dollars to get the best CPU fan you can find, get top quality fans pulling air into the machine, and get top quality fans pushing air out of the machine.
Heat kills. Don't overclock your servers, and don't go for the fastest of anything. Insure there is airflow space between your drives and make sure the room your machines are in is cool. Maybe even consider underclocking your CPU - a machine running 80% as fast but giving you an extra two years of uninterrupted uptime is a lot better than a server running 105% and dying at a rate of two hardware failures a year.
If you are hand building, build conservative machines and buy identical spare parts while you still can. Nothing sucks more than having to retire a machine because you can't get spare parts, except replacing a machine because a $10 cheapo fan stopped blowing.
You might take a look at SuSE 9.2 in a week when it becomes available for free FTP - Novell has done some pretty nice things with SuSE in the past year. I was a RedHat fanboy for a while and after a few months with SuSE 9.1 I wouldn't go back to RH if it was free.
YaST as a system management interface, available from the command line as well as in the X GUI of choice, sort of brings it all together.
I haven't played with freebsd yet so I can't make any parallels, but if RPM is driving your hate for Linux I can fully understand (and offer up an alternative.)
you should still be getting systems for less than half of the price for identical systems from IBM, HP, Dell
I bought my last Dell servers, 400sc boxes with P4 2.8GHz hyperthreaded CPU's, quarter to half a gig of Micron or Crucial dual channel DDR pc3200, 160G IDE drives, Intel gigabit ethernet (no fault tolerance in hard drive or power supply, I admit that) for about $400 apiece. When they arrive I throw in enough memory to bring them up to 1G (bought from Crucial, generally) and for about $500 apiece I'm running servers that have all identical hardware, supported by the factory, including Linux driver support (Windows driver support also, natch.)
I might be able build white box machines to those specs for $400 apiece if I pricewatch.com'ed every component from a different vendor, hassled with buying every piece from someone else, put them together myself, obligating myself to support it myself in the process, and spending a full weekend assembling them when all the parts came in (I can build a machine from 'parts in their packaging' to logging into Win2000 the first time it boots in about an hour if everything works right the first pass - which it occasionally doesn't.) If you can do it for half of that ($200 per machine at that level ready to load an OS onto) I would be interested.
That said, if you want to keep your long term costs down - do it my way. I did the 'hand build every machine using parts from every vendor under the sun, using a different build configuration for each machine) the first time and after about 4 years it became a system management and support nightmare. We had to hire two more techs at $35k a year apiece just to keep the systems running - that's about $100,000 in fully burdened salary costs per year because we saved about $200 per machine (about 150 machines over 5 years, so lets be generous and assume we saved $30k total over five years, or $6k per year.) If we had had 150 identical machines we could have maintained it with a single tech instead of three.
When you pay extra getting hardware from a single vendor you aren't paying for 'better hardware' - you are paying for a single point of contact that can get your systems running again in a hurry when they stop running. The most freakish nightmare you can give your company a decade from now is to let your network evolve over time using the cheapest route available on a day to day basis. Maybe 25% of the cost over the life of the system is hardware. 25% is software. 50% is wetware. Skimp on hardware costs to drop that 25% to 20% and the wetware costs can double.
You laugh. A friend of mine had his son banging on the keyboard at roughly 6 months, and understanding the connection between mouse movements and cursor movement by about 18 months. Kid is a serious UT2004 opponent at the ripe old age of 5, starts the game and connects to a good server on his own.
I suspect he hadn't read that far ahead in the book he was teaching, either that or he was on a power trip and wanted to teach me to shut up and sit down.
Probably both.
You sure think you know a lot about my assignment and professor even though you were not there.
You implied that he had taught you character based interaction, and hadn't taught you GUI yet. You also implied that he had given you an assignment. It then follows that regardless of what he asked for in black and white (do a case statement) what he really wanted was a program that extended on all the lessons learned in class thus far, as applicable. Given that he accepted it after you did exactly that... it seems I know quite a bit about your assignment and professor even though I wasn't there.
My earlier post stands. An important part of being an employed software engineer is giving the manager / customer / professor what he wants, not what he asks for in black and white. Another importand part of being an employed software engineer is never outsmarting and never ever one-upping the manager / customer / professor. If you had gone to him before writing the code, discussed it with him and socially engineered the discussion towards GUI code, led him to recommend that you do it in a GUI as you explored the language and had him believing it was his idea all along - the code would have been the same and you would haven scored an A. A good manager / professor isn't there to control you, isn't there to deny you anything or hassle you about long distance phone calls or taking too long at lunch. A good manager / professor is there to enable you to be as good as you can be - it's ok for you to be smarter than they are but you have to learn to do so in a very smooth, almost manipulative manner. You get away with bitch-slapping a supervisor exactly one time, destroying your upward mobility in the process. Nobody wants someone under them that makes them feel stupid (or worse, look stupid to other subordinates.)
You learn how to make him (whoever you happen to be working for at the time) look good and the world is your oyster.
I came into this thread to scream quit doing technology outreach programs (quit teaching poor people how to use computers)! and this, quite frankly, is a perfect example of why. I don't care if a welfare mom can buy a new computer for $87 worth of food stamps... that doesn't make it a good idea.
This goes double strength for going to third world countries and teaching them tech. Shit - what's a tribal bungaman gonna do with HTML? Teach them something useful like growing food on a farm.
It could be that the intent of the lesson was to demonstrate familiarity with the cin / cout or System.in.read() / System.out.print() functionality of the programming language, as evidenced by all of the lesson content both in class and in the chapter associated with the assignment. I would have bounced you for taking a literal interpretation of the assignment ('take some input, do something with, display the result to user') and doing it via GUI widgets.
I have a hard time believing that he spent weeks going over random language syntax and then assigned a homework that wasn't even remotely related to anything he had gone over or suggested that you read in the book. When a prof 'teaches' you a bunch of stuff and then gives you an assignment, the unspoken agreement is that you will apply the lessons you just 'learned' when you do your assignment.
In the real word taking the business user's spec document and implementing it to the letter without taking into account a) his intent and b) his needs is a sure-fire way to find yourself out in the cold. You sound young, sharp, bright, and motivated and by golly you were going to show him... but you forgot the golden rule - he who has the gold (or gives the grades) makes the rules. It is one thing to do a really good job - quite another thing to try and outsmart the boss, out think the professor, and twist the spec document around so you can code what you feel like coding and not what the user wanted.
The gold plating isn't so much for corrosion resistance or for the amazing electrical conductive qualities - it is because gold makes the best interface between two electrical conduits. When the two wires or connectors come together - gold does this best. As far as the actual wire goes for moving the electricity around silver is better than gold, once the connection has been made - IIRC satellites use gold connectors, but silver wires (since money is no object, and failure isn't an option.)
Actually the speculation I have heard says that if the flattops (aircraft carriers) had been in-port that day (Dec 7, 1941) then the US would have been well and good fuxored.
As for the second point, think bigger. Much. There are a few groups that can be very thankful that a man as compassionate, understanding and peaceful as W is in office - and not me.
Looking at the destruction in the SriL and India coastline I just have to wonder - there are a million H1-B's in the US making more than $50k apiece per year... if each one of them sent $500 in aid back to help the countries they are from we are talking about $500M. Where is all the ex-patriot support from the H1-B crew? I mean wouldn't that make sense?
I don't know about you, but if I'm going to be taking on a country that's far more powerful than I am, I'm not going to give them the preparation time that comes with a war declaration.
That's how the Japanese thought, right up until we nuked them. Twice.
I'm not trying to be offensive, just pointing out the facts. If I was going to be offensive I would... well I dunno, 'cause I'm not real good at offending people. Heh.
The loophole is technical in nature, not legal. There is a subtle difference between 'can he do it (technically)' and 'may he do it (legally)'. There are plenty of things that you 'can' do that you 'may not' do.
Actually nobody where I work surfs for porn - well a few guys did last year and within minutes the stormtroopers walked up and grabbed them, escorted them out of the building while HR out-processed them on way out. Zero tolerance policy enforced by some fairly visible insta-firing a few people that didn't catch a clue early enough and... no porn, no spyware, no adware.
It's actually pretty simply, and brutally effective - particularly in today's economic environment.
'Just say No' actually works, if applied correctly.
I spent a few years doing remote developent work (program from home) and got used to being able to control every environmental aspect of my work day... and was welcomed back to the real world this year.
I get a handful of the neon squishy ones at the gun range and keep them in my drawer - over the course of the week I spend as much time with them in as out. Serves a double purpose : in addition to the sweet serenity within my head, their bright neon yellow / pink give those around me a subtle reminder that quiet is a virtue in a shared environment.
Speaking of which - anybody considering what to do about the 32-bit time thing? Granted we have some time, but we got some time before we have to worry about the asteroid and I would guess the 32-bit time thingy is a lot more likely to actually hit.
Slashdot Mirror
In my honest opinion - you are 100% correct.
/. but Google the bathtub curve of hardware failures, and understand that as the system internal temperature (just the temp in the case, not on the CPU) goes up the shorter the time between sides of the bathtub curve are spread apart.
Think of the overclockers - they crank up the speed and voltages, add cooling to compensate, and when the cooling can't handle the load (the system gets too hot) it becomes unstable. It doesn't necessarily die, but stability becomes unreliable (it generally locks up, sometimes dies.)
Ball-bearing fans, balance fans blowing in and fans blowing out, and my favorite is a twin fan exhaust in the top 5.25" bay of the system like this - note that this is the first url that Google found that looks like the one I'm talking about, I don't necessarily recommend this specific vendor.
If uptime (reliability) is your concern then crank up the airflow. Your CPU will last forever or until the CPU fan dies, whichever comes first. It is not something I can draw here in
There are two types of downtime - planned and unplanned. Server death is the worst kind of unplanned downtime.
when can we SLI this card
I don't see where it would go any faster, honestly.
Look at the charts - when they are not using AA/Oversampling the single 6600GT goes just as fast as the 3D1 and the 6800GT - at all resolutions. The applications are CPU bound and better video isn't going to change that.
Granted there is always the 1337 crew pointing at the 8xOversampling / 4xAA numbers - but quite honestly the FPS hit going there on any card isn't worth the questionable increase in quality. Given the 60Hz an LCD puts out - and given that it takes a $600 LCD to run 1600x1200 - I'd say that quite honestly these cards are pretty much the same. Hardware has once again outrun software, and right now videocards are processing data faster than the CPU can serve it up. This 3D1 is cool, and it is going to take the early adopters to buy it in order to finance the second generation of the technology (ie, runs on other random motherboards) but I think I'm going to pass.
Maybe the guy working on it is color-blind.
Don't laugh - we had a guy in our office that was color blind, none of us knew it until we let him set up Windows 3.0's color scheme to best suit his needs (ok it was a long time ago.) Made the Hotdog Stand color scheme look mild in comparison - it was frightful to us, but the contrast worked great for him.
I did totally unscientific set of benchmarks on these machines (see the link in my sig, goes to my Journal).
... is it really so? I mean that with the same CPU, same kinds of memory, same hard drives, same PCI architecture, same GigE network card, same ... everything. The only things I could imagine going faster would be many different connections all requesting different transactions, or ??? but really the only variables I can see are motherboard (how would that increase throughput in any significant level) and drive array controllers (which I can see having a performance boost but in theory I could put the same drive controller in a low end box?)
I was almost able to saturate the GigE pipe outbound (116MB/s outbound on a theoretical 125MB/s peak wire speed.)
Pure memory reads (ramdrive) pegged at about 520MB/s (.5GBps or over 4Gbps.)
I keep hearing that the higher end boxes are faster but
Maybe we just need to kill all the bad people.
Pretty simple, actually.
Good points - I inferred but didn't actually state the environment I manage with this philosophy - exactly the size environment the OP is describing. Small / home office environment in which the users are all perfectly capable of building machines from the ground up, managing the network, adding users or printers themselves.
The machines I use are basic, entry level boxes (the 400sc was their entry level box, recently supplanted by the 420sc) and are little more than nicely apportioned desktops with crap video cards and Intel chipset motherboards (including the integrated GigE.) They are expandable enough for my needs (DVD burner or tape drive, plus up to 5 hard drives internally if you get creative (using the space designed for a floppy drive to put more hard drives, ditto the second 5.25" bay.)) I'm not buying hardware capable of running corporate Fortune 500 business, but at an average of $350 apiece they are plenty for my needs.
I actually picked up a few 400sc boxes (40G drives, 256M memory) for $300 apiece - isn't that the price point of the XBox when it first came out (and everybody went freaky trying to convert them to run desktop software / Linux)?
I don't look at Dell as a solution provider. I look at Dell as a whitebox dealer that happens to send me the machines pre-built using the same hardware in each machine, keeps a product line long enough that I can get more of the same later, keeps a fairly nice stash of drivers for all the hardware in one place, ships it for free, and sends me a replacement drive via personal messenger (on-site service) if a drive fails.
I don't expect the service levels on a $400 box to be good regardless of who I buy it from so I guess that's why I haven't been disappointed by Dell tech support yet. They do give a pretty good recount of the hardware going into the machines. Also their web support is good enough to get all the drivers in one place, and if you put in the service tag number it actually bundles only the drivers associated with the hardware in your machine on one page.
I'm not sure what you mean about the non-glamorous busses - I'm not real happy with the lack of an AGP slot in the machine that replaced the 400sc (the 420sc has PCI-Express) but that's only because I also liked slapping in an AGP card to make a nice cheap desktop (not a concern if we are really talking about servers.)
And I laughingly agree about their 'discounts' - watch their web site for a few weeks to catch a much better price. Opportunity is your friend, and 'have to buy today' is not.
That said - if the OP is going to build his own, he is going to build his own. If that's the case, here is the most important bit of wisdom I have to offer :
Heat Kills.
Computers die for one reason, and one reason only. Heat. I have autopsied well over 100 dead computers (servers, desktops) in the past two decades and with extremely few exceptions the deaths were caused by CPU fan failure (a $10 part), power supply fan failure (a $10 part), or hard drive failure. If I really wanted to push it, I would say the drive failures were heat related, generally due to weak power supply exhaust fans.
Heat kills. It doesn't matter where you buy your parts if you are building your own systems (hey, we all did it early in our careers when money was more scarce than time) then spend the extra eight dollars to get the best CPU fan you can find, get top quality fans pulling air into the machine, and get top quality fans pushing air out of the machine.
Heat kills. Don't overclock your servers, and don't go for the fastest of anything. Insure there is airflow space between your drives and make sure the room your machines are in is cool. Maybe even consider underclocking your CPU - a machine running 80% as fast but giving you an extra two years of uninterrupted uptime is a lot better than a server running 105% and dying at a rate of two hardware failures a year.
If you are hand building, build conservative machines and buy identical spare parts while you still can. Nothing sucks more than having to retire a machine because you can't get spare parts, except replacing a machine because a $10 cheapo fan stopped blowing.
If you are going to shop retail, shop at Fry's. If you can't shop at Fry's (ie, not in your area) you are probably better off not shopping retail.
What vendor?
I have never had that experience with machines from HP, Dell, or IBM.
You might take a look at SuSE 9.2 in a week when it becomes available for free FTP - Novell has done some pretty nice things with SuSE in the past year. I was a RedHat fanboy for a while and after a few months with SuSE 9.1 I wouldn't go back to RH if it was free.
YaST as a system management interface, available from the command line as well as in the X GUI of choice, sort of brings it all together.
I haven't played with freebsd yet so I can't make any parallels, but if RPM is driving your hate for Linux I can fully understand (and offer up an alternative.)
you should still be getting systems for less than half of the price for identical systems from IBM, HP, Dell
I bought my last Dell servers, 400sc boxes with P4 2.8GHz hyperthreaded CPU's, quarter to half a gig of Micron or Crucial dual channel DDR pc3200, 160G IDE drives, Intel gigabit ethernet (no fault tolerance in hard drive or power supply, I admit that) for about $400 apiece. When they arrive I throw in enough memory to bring them up to 1G (bought from Crucial, generally) and for about $500 apiece I'm running servers that have all identical hardware, supported by the factory, including Linux driver support (Windows driver support also, natch.)
I might be able build white box machines to those specs for $400 apiece if I pricewatch.com'ed every component from a different vendor, hassled with buying every piece from someone else, put them together myself, obligating myself to support it myself in the process, and spending a full weekend assembling them when all the parts came in (I can build a machine from 'parts in their packaging' to logging into Win2000 the first time it boots in about an hour if everything works right the first pass - which it occasionally doesn't.) If you can do it for half of that ($200 per machine at that level ready to load an OS onto) I would be interested.
That said, if you want to keep your long term costs down - do it my way. I did the 'hand build every machine using parts from every vendor under the sun, using a different build configuration for each machine) the first time and after about 4 years it became a system management and support nightmare. We had to hire two more techs at $35k a year apiece just to keep the systems running - that's about $100,000 in fully burdened salary costs per year because we saved about $200 per machine (about 150 machines over 5 years, so lets be generous and assume we saved $30k total over five years, or $6k per year.) If we had had 150 identical machines we could have maintained it with a single tech instead of three.
When you pay extra getting hardware from a single vendor you aren't paying for 'better hardware' - you are paying for a single point of contact that can get your systems running again in a hurry when they stop running. The most freakish nightmare you can give your company a decade from now is to let your network evolve over time using the cheapest route available on a day to day basis. Maybe 25% of the cost over the life of the system is hardware. 25% is software. 50% is wetware. Skimp on hardware costs to drop that 25% to 20% and the wetware costs can double.
No. We own private jets because commercial air travel, like paying taxes, is for the little people.
-Fiona
You laugh. A friend of mine had his son banging on the keyboard at roughly 6 months, and understanding the connection between mouse movements and cursor movement by about 18 months. Kid is a serious UT2004 opponent at the ripe old age of 5, starts the game and connects to a good server on his own.
I suspect he hadn't read that far ahead in the book he was teaching, either that or he was on a power trip and wanted to teach me to shut up and sit down.
... it seems I know quite a bit about your assignment and professor even though I wasn't there.
Probably both.
You sure think you know a lot about my assignment and professor even though you were not there.
You implied that he had taught you character based interaction, and hadn't taught you GUI yet. You also implied that he had given you an assignment. It then follows that regardless of what he asked for in black and white (do a case statement) what he really wanted was a program that extended on all the lessons learned in class thus far, as applicable. Given that he accepted it after you did exactly that
My earlier post stands. An important part of being an employed software engineer is giving the manager / customer / professor what he wants, not what he asks for in black and white. Another importand part of being an employed software engineer is never outsmarting and never ever one-upping the manager / customer / professor. If you had gone to him before writing the code, discussed it with him and socially engineered the discussion towards GUI code, led him to recommend that you do it in a GUI as you explored the language and had him believing it was his idea all along - the code would have been the same and you would haven scored an A. A good manager / professor isn't there to control you, isn't there to deny you anything or hassle you about long distance phone calls or taking too long at lunch. A good manager / professor is there to enable you to be as good as you can be - it's ok for you to be smarter than they are but you have to learn to do so in a very smooth, almost manipulative manner. You get away with bitch-slapping a supervisor exactly one time, destroying your upward mobility in the process. Nobody wants someone under them that makes them feel stupid (or worse, look stupid to other subordinates.)
You learn how to make him (whoever you happen to be working for at the time) look good and the world is your oyster.
I came into this thread to scream quit doing technology outreach programs (quit teaching poor people how to use computers)! and this, quite frankly, is a perfect example of why. I don't care if a welfare mom can buy a new computer for $87 worth of food stamps ... that doesn't make it a good idea.
This goes double strength for going to third world countries and teaching them tech.
Shit - what's a tribal bungaman gonna do with HTML?
Teach them something useful like growing food on a farm.
I could rant on, but I won't.
It could be that the intent of the lesson was to demonstrate familiarity with the cin / cout or System.in.read() / System.out.print() functionality of the programming language, as evidenced by all of the lesson content both in class and in the chapter associated with the assignment. I would have bounced you for taking a literal interpretation of the assignment ('take some input, do something with, display the result to user') and doing it via GUI widgets.
... but you forgot the golden rule - he who has the gold (or gives the grades) makes the rules. It is one thing to do a really good job - quite another thing to try and outsmart the boss, out think the professor, and twist the spec document around so you can code what you feel like coding and not what the user wanted.
I have a hard time believing that he spent weeks going over random language syntax and then assigned a homework that wasn't even remotely related to anything he had gone over or suggested that you read in the book. When a prof 'teaches' you a bunch of stuff and then gives you an assignment, the unspoken agreement is that you will apply the lessons you just 'learned' when you do your assignment.
In the real word taking the business user's spec document and implementing it to the letter without taking into account a) his intent and b) his needs is a sure-fire way to find yourself out in the cold. You sound young, sharp, bright, and motivated and by golly you were going to show him
The gold plating isn't so much for corrosion resistance or for the amazing electrical conductive qualities - it is because gold makes the best interface between two electrical conduits. When the two wires or connectors come together - gold does this best. As far as the actual wire goes for moving the electricity around silver is better than gold, once the connection has been made - IIRC satellites use gold connectors, but silver wires (since money is no object, and failure isn't an option.)
Actually the speculation I have heard says that if the flattops (aircraft carriers) had been in-port that day (Dec 7, 1941) then the US would have been well and good fuxored.
... if each one of them sent $500 in aid back to help the countries they are from we are talking about $500M. Where is all the ex-patriot support from the H1-B crew?
As for the second point, think bigger. Much.
There are a few groups that can be very thankful that a man as compassionate, understanding and peaceful as W is in office - and not me.
Looking at the destruction in the SriL and India coastline I just have to wonder - there are a million H1-B's in the US making more than $50k apiece per year
I mean wouldn't that make sense?
I don't know about you, but if I'm going to be taking on a country that's far more powerful than I am, I'm not going to give them the preparation time that comes with a war declaration.
... well I dunno, 'cause I'm not real good at offending people. Heh.
That's how the Japanese thought, right up until we nuked them.
Twice.
I'm not trying to be offensive, just pointing out the facts. If I was going to be offensive I would
The loophole is technical in nature, not legal.
There is a subtle difference between 'can he do it (technically)' and 'may he do it (legally)'.
There are plenty of things that you 'can' do that you 'may not' do.
Actually nobody where I work surfs for porn - well a few guys did last year and within minutes the stormtroopers walked up and grabbed them, escorted them out of the building while HR out-processed them on way out. Zero tolerance policy enforced by some fairly visible insta-firing a few people that didn't catch a clue early enough and ... no porn, no spyware, no adware.
It's actually pretty simply, and brutally effective - particularly in today's economic environment.
'Just say No' actually works, if applied correctly.
No joke about permissions - I would give ANYTHING to have file / directory permissions set up the way Netware's NDS (Netware 4.x and newer) does it.
Come to think of it, now that Novell is hacking on the Suse code base - how about adding NDS as the file permission scheme?
Database - Sybase ASE 12.5
If the box is
Running Linux
On one CPU
With 2G of RAM or less
And keeps the total database space to less than 5G
= free Sybase ASE 12.5
Sybase Linux Promo
Actually this is exactly how I do it.
... and was welcomed back to the real world this year.
I spent a few years doing remote developent work (program from home) and got used to being able to control every environmental aspect of my work day
I get a handful of the neon squishy ones at the gun range and keep them in my drawer - over the course of the week I spend as much time with them in as out. Serves a double purpose : in addition to the sweet serenity within my head, their bright neon yellow / pink give those around me a subtle reminder that quiet is a virtue in a shared environment.
Speaking of which - anybody considering what to do about the 32-bit time thing? Granted we have some time, but we got some time before we have to worry about the asteroid and I would guess the 32-bit time thingy is a lot more likely to actually hit.