Currently, there is a massive lack of funding for basic research (except stuff which enables showoffs, like putting humans on the moon). Marketing is a higher priority.
A significant proportion of research is just reshashing what hjas been done elsewhere, not truly new stuff.
After all, all that India and China are asking for is the same quality of life for their citizens as enjoyed by the more "developed" nations. Perhaps you need to rethink your assumptions here.
The problem is that the mainframe is still a huge, single point of failure. What we need is the ability to toss a few dozen _CHEAP_ systems at the problem, and figure out how to make it work.
Failures happen. Code around them.
There is only so high you can scale a mainframe up to. Then you need to start scaling out. Scalability isn't a few thousand users pounding your systems. It's about a few million users pounding your systems, with increases of one order of magnitude being common.
Learning how systems work WILL make you a better programmer. It will cost you in terms of time spent at work. The big difference is that your code will be better performing, and easier to maintain.
The bandwidth for a call is fixed. If you call someone but don't talk, that circuit is still running at full bandwidth.
Packet switching saves on the costs of keeping the virtual circuit alive by multiplexing the stream between multiple users. By always using the bandwidth, even when the humans aren't around, you are removing the benefits of multiplexing, but expect to pay the same price. Something has to give, and it's not likely to be the packet switching. You can get dedicated, always on, unshared connectivity. It's called a T1.
Well, it varies from basic tech support, to systems administration, to writing code, to architecting systems, to network maintainance, dealing with performance, security,...
What you are thinking of is probably business IT, which SUCKS.
The idea of profit is not evil. However, the idea of unrestricted, increasing growth in profits forever is. Make a reasonable profit. But don't expect the percentage growth to remain the same on and on.
The ISP probably buys 10 Mbit/s of transit, but sells it to 100 users at 1Mit/s each (1 1:10 oversubscription). If users wanted less oversubscription, those plans would be available (typically business class ADSL, or T1/E1 and above).
This would be reasonable for usage patterns which were previously considered normal (web surfing, email, some chat) over asymmetric links, with a large bandwidth bias in favour of the subscriber.
This is one of the reasons why ISPs forbid end-users from running servers (which have a bandwidth usage bias in the other direction).
It isn't necessary that you use only 1/4th of the bandwidth, it is that you use full bandwidth only 1/4th of the total time, and be nice to your neighbours. What we have on the other hand is a culture which encourages grabbing shared resources for yourself all the time instead of being nice to your neighbours. ISPs don't encourage the "be nice to your neighbours" viewpoint either.
In a civilised world, file-sharing wouldn't even be a problem because it would be dealt with by social measures (but we would also have world peace and no spam). In this scenario, the file-sharers are the abusers of the last-mile bandwidth commons, and cause trouble for the remaining users (just like spammers cause trouble for the rest of the Internet).
Distributed file sharing technologies re-define normal, throwing their existing calculations off kilter.
Slashdot Mirror
Large portions of chat, P2P, spam and websurfing is for porn.
Currently, there is a massive lack of funding for basic research (except stuff which enables showoffs, like putting humans on the moon). Marketing is a higher priority.
A significant proportion of research is just reshashing what hjas been done elsewhere, not truly new stuff.
And I thought Java was the replacement to Perl because there was only one way to do it?
Google for PostgreSQL dollar quoting.
Redhat support is Pune, not Bangalore
See California
As a large scale mail admin, let me give you a hint: Databases are not good email stores. You more likely want maildir format for your message store.
http://en.wikipedia.org/wiki/List_of_countries_by_greenhouse_gas_emissions_per_capita indicates that the US emitted 22.9 tonnes of CO2 per capita, China was at 3.9 tonnes, and India was at a measly 1.8 tonnes per capita. The US should have been emitting ~ 0.8 tonnes per capita of CO2 to be equitable to India, twice that for China.
After all, all that India and China are asking for is the same quality of life for their citizens as enjoyed by the more "developed" nations. Perhaps you need to rethink your assumptions here.
The problem is that the mainframe is still a huge, single point of failure. What we need is the ability to toss a few dozen _CHEAP_ systems at the problem, and figure out how to make it work.
Failures happen. Code around them.
There is only so high you can scale a mainframe up to. Then you need to start scaling out. Scalability isn't a few thousand users pounding your systems. It's about a few million users pounding your systems, with increases of one order of magnitude being common.
Don't Repeat Yourself is a good principle.
You can do that with RedHat (also see CentOS). Very definitely a redistribution of commercial application.
Learning how systems work WILL make you a better programmer. It will cost you in terms of time spent at work. The big difference is that your code will be better performing, and easier to maintain.
ISPs are not common carriers.
Flickr and Yahoogroups were companies purchased by Yahoo! !innovation.
That wasn't Microsoft software, that was Compaq who truly ushered in the PC revolution.
http://www.studentsfororwell.org/
The US has always been the land of the free*.
* Subject to terms and conditions, offer not valid where inapplicable.
The bandwidth for a call is fixed. If you call someone but don't talk, that circuit is still running at full bandwidth.
Packet switching saves on the costs of keeping the virtual circuit alive by multiplexing the stream between multiple users. By always using the bandwidth, even when the humans aren't around, you are removing the benefits of multiplexing, but expect to pay the same price. Something has to give, and it's not likely to be the packet switching. You can get dedicated, always on, unshared connectivity. It's called a T1.
That's not Microsoft, that's Compaq. Microsoft merely rode the gravy train.
Just toss the coin 'n' times, where 'n' is a positive integer such that 2^n >= k > 2^(n-1) where k is the number of possible choices.
Then they need to switch to local.example.com (where example.com is their domain, or they could just use example.(com|net|org).
It's even a RFC.
Multi-process FTW. You don't always need threads. Event driven programming + multiple processes can make your life easier.
Threads generally imply shared memory, and all the attendant pain that comes with it.
Well, it varies from basic tech support, to systems administration, to writing code, to architecting systems, to network maintainance, dealing with performance, security, ...
What you are thinking of is probably business IT, which SUCKS.
Perhaps the people who worked on this sort stuff:
http://en.wikipedia.org/wiki/Bell_Labs#Discoveries_and_Developments
A few of them are even at one of those search engine companies.
The idea of profit is not evil. However, the idea of unrestricted, increasing growth in profits forever is. Make a reasonable profit. But don't expect the percentage growth to remain the same on and on.
The ISP probably buys 10 Mbit/s of transit, but sells it to 100 users at 1Mit/s each (1 1:10 oversubscription). If users wanted less oversubscription, those plans would be available (typically business class ADSL, or T1/E1 and above).
This would be reasonable for usage patterns which were previously considered normal (web surfing, email, some chat) over asymmetric links, with a large bandwidth bias in favour of the subscriber.
This is one of the reasons why ISPs forbid end-users from running servers (which have a bandwidth usage bias in the other direction).
It isn't necessary that you use only 1/4th of the bandwidth, it is that you use full bandwidth only 1/4th of the total time, and be nice to your neighbours. What we have on the other hand is a culture which encourages grabbing shared resources for yourself all the time instead of being nice to your neighbours. ISPs don't encourage the "be nice to your neighbours" viewpoint either.
In a civilised world, file-sharing wouldn't even be a problem because it would be dealt with by social measures (but we would also have world peace and no spam). In this scenario, the file-sharers are the abusers of the last-mile bandwidth commons, and cause trouble for the remaining users (just like spammers cause trouble for the rest of the Internet).
Distributed file sharing technologies re-define normal, throwing their existing calculations off kilter.