Since java was introduced I have had faster and faster computers, yet java doesn't run any faster than it used to.
That may be true, but speed isn't the only measurement of a "good" system. Ease of development & maintenance, lack of buffer overflows, security, reasonably standard APIs, quantity of available developers, platform independence are all legitimate requirements.
As for scalability, I have issues about using app servers & J2EE where a decent database will generally suffice, but that's beyond the language.
I think that document refers to new (connect by features) rather than (new connect by) features.
I've been happily using it on version 8. The connect_by_root & sys_connect or whatever it is may well be introduced in 10g. Certainly I could have killed for the ORDER SIBLINGS BY clause in 8 but it's only available in 9i+.
Point of order - CONNECT BY has been around for ages, possibly since 6 or even 5.
In the real world it can be very useful, although a bit non-relational in the true sense.
I've heard that too, although I find it hard to swallow. Considering how cheap & easy it is to copy a CD, once you've produced the master I'd imagine it's pretty much a cash cow. When your're knocking them out at $10 or whatever that's a huge mark up even with overheads. Each concert obviously has an associated cost that won't go away with time.
I must admit, I wonder if these guys & their record companies need to re-evaluate the whole industry. Once, recordings were the main way people heard an artist and the live concerts were something special. Now the recording needs to be seen as pretty much advertising for the concerts, T shirts etc.
Perhaps the good times are over and in the same way as programming is now commodity, so is making music. You can make a living - just not a particularly spectacular one.
Absoultely!
As an ex-assembly programmer in business applications (circa '87) I can say in this arena it's now virtually useless. I've certainly moved off it and have never looked back.
Since then, more factors than speed contribute to a system's success, e.g.
1) Maintainablity - #1 for a reason. This stuff will be around for years with an ever-changing staff to look after it.
2) Extensibility - nothing stays the same and a good design will reflect this. There is often very little point in spending 100 hours optimizing a few specific cases that will be gone in a year.
3)IO. Most of the stuff I write goes to a database. No amount of hand coded assembler will make my program go faster once I'm waiting for a database call to complete, which is can be several orders of magnitude longer than processing the result set.
3.5) In the above cases, making use of a simple, safe threading model built into the language is better than attempting to write it faster in assembler.
4)Use the database properly. Naieve database access is responsible for 95% of performance issues I've recently come across (application level joins/aggregation for example).
5) An optimised or appropriate algorithm is often way better than an optimised implementation. For example, I've seen code iterating through some sort of list to find a value rather than using a map in the first place.
6) In house, you can be pragmatic. If you have a well desgined system that underperforms it may well be cheaper to spend $1000 on a new box than a week of someone's time to rewrite or optimize in assembler, not to mention run it through the battery of regression tests, backout scenarios and all that other crap.
7)Cross platform.
Anyway - not to slag off assembly totally, but in the world of mainstream applications you're 99.9% correct with your 99.9%.
A variation on the ending is that he calls out "197" and the whole place erupts in howling laughter for about 20 minutes. When he asks what was so funny about #197 he's told that nobody had heard it before.
I wasn't suggesting it was at all effective. Just wondering what the Linux version of such superstitious behaviour would be. ls ls ls try again...?
never mind.
I'll second that. What I also noticed that in the internet cafes the 'solution' to a crappy connection that kept dropping out was to right-click the desktop and choose 'Refresh' from the menu a couple of times then try again. Fuck knows what the Linux equivalent to this would be. Dead chicken waving?
You're right. Depending on where in the world you go, paper can have advantages. In certain parts of the world it can be great to have your precious notes in an essentially valueless form. Even if you don't find yourself bargaining with a mugger for your possesions, just leaving stuff in hotels can be hazardous but a notepad rarely proves tempting. Even just having a quite drink can draw unwanted attention if you're clicking away at something.
That said, I wish I'd had the facilities to record sounds - markets, rainforests, etc. as well as pictures. The ability to upload piccies & stuff to home regularly would be a great boon too - not that it's happened to me yet, but losing your holiday films is pretty crap.
Well said. I would also point out that as far as I know, most decent RDBMSs allow you to write your SPs in Java. Not quite as efficient as PL/SQL or whatever but less clunky and more portable should the port be required.
Different FUD for different reasons. Just because I can see it in one place doesn't mean I can't possibly think it can happen elsewhere. My 2 points being
1) Governments can increase their votes by creating fear and then making a big, high profile show of "addressing it"
2) This action is nothing but one of these shows.
There may be many other others playing similar games, but when the government uses this strategy it can also appear legitimate while really only spending the taxpayers' money in order to get votes.
I think "scaring the piss outta everyone" is a well used device that shakey governments employ to increase their chances of re-election. It's pretty clear that a determined terrorist can do what they want to do if they put their mind to it, even if everyone is asked if they packed their own bags. This is just FUD, and expensive & inconvinient FUD at that.
Very good way to show the world that linux is more mainstream every day.
Sorry to rain on your parade, but not much of the world would consider a C++ compiler mainstream. (In fact I doubt most of the world know or care what it is).
I would suggest that your experience indicates that internet access belongs in the library. Arguments against having it in the classroom are still valid as far as I can tell.
But the question is, is it really worth it? Considering that pilots are just "along for the ride" almost all of the time, what's the cost?
From various sources:
~$250K p/a
Max 100 hours per month
=~$208 per hour
Assuming a 1/2 full plane that's about 1$ per passenger on the larger ones. I'd pay that.
1. An adult male human.
2. A human regardless of sex or age; a person.
3. A human or an adult male human belonging to a specific occupation, group, nationality, or other category. Often used in combination: a milkman; a congressman; a freeman.
4. The human race; mankind: man's quest for peace.
Good enough for me, but I've never been that PC anyway.
Slashdot Mirror
I M IN MOMS BSMNT. LOL.
That's what I thought too. Also, the over enthusiasic use of ! rang a few warning bells for me. Oh well.
Or The journalists that work at the outfit the link went to. Did you notice it took 3 of them to write that article? Talk about overstaffed.
As for scalability, I have issues about using app servers & J2EE where a decent database will generally suffice, but that's beyond the language.
I think that document refers to new (connect by features) rather than (new connect by) features.
I've been happily using it on version 8. The connect_by_root & sys_connect or whatever it is may well be introduced in 10g. Certainly I could have killed for the ORDER SIBLINGS BY clause in 8 but it's only available in 9i+.
Point of order - CONNECT BY has been around for ages, possibly since 6 or even 5.
In the real world it can be very useful, although a bit non-relational in the true sense.
I've heard that too, although I find it hard to swallow. Considering how cheap & easy it is to copy a CD, once you've produced the master I'd imagine it's pretty much a cash cow. When your're knocking them out at $10 or whatever that's a huge mark up even with overheads. Each concert obviously has an associated cost that won't go away with time.
I must admit, I wonder if these guys & their record companies need to re-evaluate the whole industry. Once, recordings were the main way people heard an artist and the live concerts were something special. Now the recording needs to be seen as pretty much advertising for the concerts, T shirts etc.
Perhaps the good times are over and in the same way as programming is now commodity, so is making music. You can make a living - just not a particularly spectacular one.
Absoultely!
As an ex-assembly programmer in business applications (circa '87) I can say in this arena it's now virtually useless. I've certainly moved off it and have never looked back.
Since then, more factors than speed contribute to a system's success, e.g.
1) Maintainablity - #1 for a reason. This stuff will be around for years with an ever-changing staff to look after it.
2) Extensibility - nothing stays the same and a good design will reflect this. There is often very little point in spending 100 hours optimizing a few specific cases that will be gone in a year.
3)IO. Most of the stuff I write goes to a database. No amount of hand coded assembler will make my program go faster once I'm waiting for a database call to complete, which is can be several orders of magnitude longer than processing the result set.
3.5) In the above cases, making use of a simple, safe threading model built into the language is better than attempting to write it faster in assembler.
4)Use the database properly. Naieve database access is responsible for 95% of performance issues I've recently come across (application level joins/aggregation for example).
5) An optimised or appropriate algorithm is often way better than an optimised implementation. For example, I've seen code iterating through some sort of list to find a value rather than using a map in the first place.
6) In house, you can be pragmatic. If you have a well desgined system that underperforms it may well be cheaper to spend $1000 on a new box than a week of someone's time to rewrite or optimize in assembler, not to mention run it through the battery of regression tests, backout scenarios and all that other crap.
7)Cross platform.
Anyway - not to slag off assembly totally, but in the world of mainstream applications you're 99.9% correct with your 99.9%.
A variation on the ending is that he calls out "197" and the whole place erupts in howling laughter for about 20 minutes. When he asks what was so funny about #197 he's told that nobody had heard it before.
I wasn't suggesting it was at all effective. Just wondering what the Linux version of such superstitious behaviour would be.
ls
ls
ls
try again...?
never mind.
I'll second that.
What I also noticed that in the internet cafes the 'solution' to a crappy connection that kept dropping out was to right-click the desktop and choose 'Refresh' from the menu a couple of times then try again.
Fuck knows what the Linux equivalent to this would be. Dead chicken waving?
You're right. Depending on where in the world you go, paper can have advantages. In certain parts of the world it can be great to have your precious notes in an essentially valueless form. Even if you don't find yourself bargaining with a mugger for your possesions, just leaving stuff in hotels can be hazardous but a notepad rarely proves tempting. Even just having a quite drink can draw unwanted attention if you're clicking away at something.
That said, I wish I'd had the facilities to record sounds - markets, rainforests, etc. as well as pictures. The ability to upload piccies & stuff to home regularly would be a great boon too - not that it's happened to me yet, but losing your holiday films is pretty crap.
Well said. I would also point out that as far as I know, most decent RDBMSs allow you to write your SPs in Java. Not quite as efficient as PL/SQL or whatever but less clunky and more portable should the port be required.
Different FUD for different reasons. Just because I can see it in one place doesn't mean I can't possibly think it can happen elsewhere. My 2 points being
1) Governments can increase their votes by creating fear and then making a big, high profile show of "addressing it"
2) This action is nothing but one of these shows.
There may be many other others playing similar games, but when the government uses this strategy it can also appear legitimate while really only spending the taxpayers' money in order to get votes.
I think "scaring the piss outta everyone" is a well used device that shakey governments employ to increase their chances of re-election. It's pretty clear that a determined terrorist can do what they want to do if they put their mind to it, even if everyone is asked if they packed their own bags. This is just FUD, and expensive & inconvinient FUD at that.
Well said, but too late for my mod points for what it's worth.
I would suggest that your experience indicates that internet access belongs in the library. Arguments against having it in the classroom are still valid as far as I can tell.
I think by the time it emerged from marketing it would be a "stochastic process".
~$250K p/a
Max 100 hours per month
=~$208 per hour
Assuming a 1/2 full plane that's about 1$ per passenger on the larger ones. I'd pay that.
Good enough for me, but I've never been that PC anyway.
Exactly. For the price of a couple of beers it has to be worth it. Who the fuck worries over $2?