But all you have to do is put the iPod in one box, and then create a "connector package." When you purchase an iPod, you get one iPod box and one connector package box.
Last I checked, IEEE 1394 had less overhead than USB. What this means is that more of the transmission on IEEE 1394 is your data than some kind of header information, which translates to faster downloads to the device. Think of the cell tax when using ATM-based networks. It's a similar concept.
Besides, VHS may have won the home video tape wars, but that didn't make it better than Betamax.
Ever hear of FITL? Or IFITL? That's how they control the telco lines. Sure, they're open to competition, but they won't put the equipment necessary out there to allow for IFITL. They have it. They just need to put a different card in your "lightspeed box" and you can get much better and faster levels of DSL service.
By leaving it as FITL, you can only get IDSL (aka ISDN)... even if you live 10 feet from your central office.
In the US, many of us got saddled with FITL. Verizon, SBC, and BellSouth could easily make it IFITL, but when they do, they have to open up the lines to competition... so they don't.
If you have FITL, it doesn't matter if you're physically 10 feet from the central office. You can only get IDSL (aka ISDN).
That's what was on a sign in a coworker's office many years ago when I was working at a company that used IBM Mainframes, Solaris, and various other systems too.
It was one of my first jobs, so she explained it to me by saying that OCO meant "Object Code Only"...
Anyone notice how they seem to be all over Windows?
I'll bet the guy who used gets() is long gone, so they're still searching for each of his hidden calls to it. It's either that, or he won't admit to ever having used it.
Why is it that civil cases take so much longer than criminal ones? Even the OJ Simpson criminal case finished 16 months after arrest, and people were all up in arms about how long it was dragging on, and yet this case has been going on for two years and it hasn't even go to the court room yet!
There's a couple of major differences between civil and criminal cases.
Jury size
Criminal: 12 + alternates
Civil: 6 (8?) + alternates
Evidence requirements
Criminal: very stringent; must be clear-cut
Civil: not quite as stringent; can have minor doubts; plaintiff only needs a preponderance of evidence
Verdict requirements
Criminal: Jury must be unanimous
Civil: 2/3 majority of the Jury
Things carry on longer in civil suits during the trial phase, but don't frequently take as long in the jury deliberations because of the differences in requirements. You can present most anything you think might help your case in the civil court. In the criminal court, if it's irrelevant or has too many doubts, you don't want to bring it forward, because it makes the prosecutor look desperate.
In the case of infecting individual computers, the popularity argument, as you called it, is merely a way of saying that you're getting the best bang for your buck. You're far more likely to infect a greater number of machines. Once those machines are infected, they spread it to other machines.
In terms of market share of Apache vs IIS, the problem here is that your success rate of infecting a machine is generally going to be higher with IIS because of the discussion we've had in other threads... IIS is most likely going to be run with Administrator privileges on that machine. Apache, at least on Unix/Linux systems, runs as its own user/group, so it never has root privileges on the machine.
Instead of some guy in uniform reading your papers, looking for a particular seal, etc, he will just pull out a small scanning device, slide your license through like a credit card swipe, and then see what's going on.
As for digital pictures, I think a lot of the states, if not all of them, have switched to using some kind of digital camera now.
If there's such a big problem with the licenses being forged, why not inform the states whose licenses are forged the most and let them come up with something?
Actually, while Microsoft may have posted record profits, they've also noted that their sales of Windows have actually declined. Their entire profit increase was more due to cost cuts and sales of Halo 2.
That's what I've always liked about X. You just need to choose a window manager based on features, etc, which you then can configure certain behavioral aspects plus appearance.
As long as your window manager conforms to the necessary standards, you're free to pick and choose what you like.
Mounting a share is akin to using the "Map Network Drive" part of Windows. You can't do anything with a file from a share on Windows if you don't map the network drive first.
Getting users to mount the shares, however, is a different story. I haven't been in your situation, so I don't know if a utility already exists to "automount" SMB shares at login.
But an SMB share is not an OS, and neither is Samba. SMB is a protocol implemented by Samba to access something specific.
What you describe isn't an OS flaw. It's a flaw with the way Samba is designed. It also may have been the only way to do something like that. And if Samba is set up correctly, you should be able to write to that share if your userid matches the owning userid. That's done as part of a translation. Once I've logged in, however, I haven't had problems working on specific files I own that are on the share.
Check the Samba How-To if you're running Linux (or Solaris, or FreeBSD, etc) as your server.
Actually, it's more like the person on XP is running the process as root through sudo, or having logged in then immediately running su. You get the ID set up as a member of the Administrator group, and you have full privileges on the entire machine. The unfortunate thing about Windows XP (2K and NT, too), is that if you aren't a member of the Administrator group, there are some necessary or common functions you may not be able to do. Something like that happened on one of my contracts... until someone came around and set it up that my ID was a member of the Administrator group on that machine, there were some basic LAN connectivity things I could not accomplish.
My expertise is primarily with database and data transfer, dealing with large volumes of data (20k to 2mil plus rows per transaction).
When you're dealing with data transactions, i.e., processing data row after data row from a database table, there is only so much scaling you can do. At some point, it becomes linear processing, even if it means one process or thread is executing while the others are in a wait state.
BTW, multiplication is part of data manipulation. It's simply one facet of it.
And the answer to the question I posed is, "It depends on the architecture." On some machines, it will be faster to use a multiplication operation, while on others it will be faster to do a bit shift, though generally bit shift wins out. But that's only one thing. I've cut run times down on database programs just by rewriting the queries and making some minor changes to the programs to accomodate the new queries. I didn't hear anybody complain when 3 hours of run time went down to 20 minutes.
The only complaints he gets are from a very specific person at his shop who seems to have contempt for anyone without a CS degree and isn't a MicroZombie. My friend only needs to defend his dissertation in applied mathematics. Is he the be-all and end-all? No. Is he thinking about making things run better? Yes. Is he making them run better? Yes. He gets his work done by the deadlines and gets them working properly. He gets lots of compliments from other co-workers who end up having to maintain/upgrade his code. He is able to get things to work that others can't.
There are some things that translate well across all types of systems. There are others that can take a back seat while working on one type of system. If your stuff is distributed across several thousand machines, a lot of optimization techniques won't be noticed. If it's all centralized (single data server, etc), you'd be surprised how your response times can be improved by doing the little things, including better planning.
you'll see that all variables are padded out to machine boundaries anyway, so ordering your variables by type is just a waste of your time.
Wrong. You just explained the inefficiency of NOT ordering your variables. You padded things out to (double) word boundaries... you WASTED SPACE. When you're looking to write code for embedded, limited memory systems, you have to do these things. Once you start doing it for one, it's easily carried across to the other platforms.
There are all kinds of little things you can always do along the way to make your code run better. Which is faster: x *= 2 or x << 1? No, it doesn't make a differnece on some programs, but when you start dealing with large volumes of data, small things like choosing the faster of those two can actually start to save you noticeable amounts of time.
And if you call cutting run-time down from 4 hours to 2.5 with about 10 minutes' worth of thinking "wasted micro-optimization," I'll hire him before you any day. He's at a point where he does this automatically.
Slashdot Mirror
Last I checked, IEEE 1394 had less overhead than USB. What this means is that more of the transmission on IEEE 1394 is your data than some kind of header information, which translates to faster downloads to the device. Think of the cell tax when using ATM-based networks. It's a similar concept.
Besides, VHS may have won the home video tape wars, but that didn't make it better than Betamax.
Flexibility is a good thing.
By leaving it as FITL, you can only get IDSL (aka ISDN)... even if you live 10 feet from your central office.
If you have FITL, it doesn't matter if you're physically 10 feet from the central office. You can only get IDSL (aka ISDN).
- SBC (primarily it's PacBell portion)
- Verizon
- BellSouth
We would've long ago had a much higher penetration level, except they want to control the lines and the access.You'd have thought they would've received this during the dot-com boom or before that.
It was one of my first jobs, so she explained it to me by saying that OCO meant "Object Code Only"...
I'll bet the guy who used gets() is long gone, so they're still searching for each of his hidden calls to it. It's either that, or he won't admit to ever having used it.
View them here
There's a couple of major differences between civil and criminal cases.
Things carry on longer in civil suits during the trial phase, but don't frequently take as long in the jury deliberations because of the differences in requirements. You can present most anything you think might help your case in the civil court. In the criminal court, if it's irrelevant or has too many doubts, you don't want to bring it forward, because it makes the prosecutor look desperate.
In terms of market share of Apache vs IIS, the problem here is that your success rate of infecting a machine is generally going to be higher with IIS because of the discussion we've had in other threads... IIS is most likely going to be run with Administrator privileges on that machine. Apache, at least on Unix/Linux systems, runs as its own user/group, so it never has root privileges on the machine.
As for digital pictures, I think a lot of the states, if not all of them, have switched to using some kind of digital camera now.
If there's such a big problem with the licenses being forged, why not inform the states whose licenses are forged the most and let them come up with something?
Read about it here
As long as your window manager conforms to the necessary standards, you're free to pick and choose what you like.
Getting users to mount the shares, however, is a different story. I haven't been in your situation, so I don't know if a utility already exists to "automount" SMB shares at login.
What you describe isn't an OS flaw. It's a flaw with the way Samba is designed. It also may have been the only way to do something like that. And if Samba is set up correctly, you should be able to write to that share if your userid matches the owning userid. That's done as part of a translation. Once I've logged in, however, I haven't had problems working on specific files I own that are on the share.
Check the Samba How-To if you're running Linux (or Solaris, or FreeBSD, etc) as your server.
Poor design, IMHO.
BTW, multiplication is part of data manipulation. It's simply one facet of it.
And the answer to the question I posed is, "It depends on the architecture." On some machines, it will be faster to use a multiplication operation, while on others it will be faster to do a bit shift, though generally bit shift wins out. But that's only one thing. I've cut run times down on database programs just by rewriting the queries and making some minor changes to the programs to accomodate the new queries. I didn't hear anybody complain when 3 hours of run time went down to 20 minutes.
The only complaints he gets are from a very specific person at his shop who seems to have contempt for anyone without a CS degree and isn't a MicroZombie. My friend only needs to defend his dissertation in applied mathematics. Is he the be-all and end-all? No. Is he thinking about making things run better? Yes. Is he making them run better? Yes. He gets his work done by the deadlines and gets them working properly. He gets lots of compliments from other co-workers who end up having to maintain/upgrade his code. He is able to get things to work that others can't.
There are some things that translate well across all types of systems. There are others that can take a back seat while working on one type of system. If your stuff is distributed across several thousand machines, a lot of optimization techniques won't be noticed. If it's all centralized (single data server, etc), you'd be surprised how your response times can be improved by doing the little things, including better planning.
Wrong. You just explained the inefficiency of NOT ordering your variables. You padded things out to (double) word boundaries... you WASTED SPACE. When you're looking to write code for embedded, limited memory systems, you have to do these things. Once you start doing it for one, it's easily carried across to the other platforms.
There are all kinds of little things you can always do along the way to make your code run better. Which is faster: x *= 2 or x << 1? No, it doesn't make a differnece on some programs, but when you start dealing with large volumes of data, small things like choosing the faster of those two can actually start to save you noticeable amounts of time.
And if you call cutting run-time down from 4 hours to 2.5 with about 10 minutes' worth of thinking "wasted micro-optimization," I'll hire him before you any day. He's at a point where he does this automatically.