If a "solved" problem like email actually brought the net down... for a while. How do you get patches for a sendmail program without using the internet?
So really it comes down to asking if people can have an expectation of privacy.
I'm going to say it will probably be no. Police are allowed to monitor cordless phone conversations without a warrant, even though people might not realize that those conversations are being broadcast.
In a similar way, if I sit downstream for you on a cable line or LAN, I most certainly can listen to conversations without doing too much intrusion. Maybe people will begin to pay attention to security if this keeps happening.
Specter (R-Utah) was and is very much in favor of lifting copyright restrictions on file sharing and music. And I'm sure you'll find plenty of democrats who voted for the DMCA.
The whole "Democrats are for the little guy" would fly farther if Hollywood didn't like Democrats so much, and vice versa.
Never vote one party simply because you think that will help. Actually look at the candidates.
I just got a Compaq laptop, and while screwing around with it, ended up nuking the original XP that it came with. (I had repartitioned the drives incorrectly, and then had installed GRUB.)
Anyway, when I used the recover CDs, it removed the partions that I had made, but didn't touch the MBR!!
I would have thought that this would be a major oversite of a recovery disk-- isn't that the most famous place for a virus to sleep?
(Alls well now, I'm dual booting with no difficulty...)
I've been hunting for the past hour on microsoft's site and on google looking for the XP EULA, and I can't find it. I'm awaiting delivery of a laptop with XP on it, and I want to see if I need to delete everything or not when it shows up.
Sure, the very obvious bugs might be caught, but there's a catch:
If software has bugs that are easy to see, and come up frequently, it's perceived as "buggy", and few people will download it, and fewer eyes will look for the bugs.
If a bug is harder to reproduce, it probably doesn't come up very often, and not everyone will spend time looking for the bug. Hell, I'm in a research environment, I know how to code, and my KDE print daemon crashes everyday. I don't care enough to try to submit bug traq reports, or look at why it's crashing, I'm just going to hope that it magically gets fixed in the next release.
Exactly how many people here have actually hunted through source code to find the one little bug that annoys them, like the fact that Konq occasionally ignores links, or that it's javascript interpretter is far from perfect. Source code is not the magic bullet, but I'll admit it's better than nothing.
Here is a laptop for $1000. Just one. The first I could find.
At our school, the $100/semester is expicity towards equipping the computing centers in the University, and making sure that they have enough systems for everybody. The support staff is paid for by other monies.
Now, I'm not going to attempt to solve the "Low-bidder" or "Bad-spec" jobs-- these are completely different and impossible problems that happen any time money is spent. And as for "being obsolete" -- what does that matter? If a laptop is useless in 4 years, who care? Exactly what can you buy that will not be obsolete in 4 years?
And as for people and maintenance, I'm not saying that the University needs to run an exchange program, just have the students deal with the manufacturer and have a waranty. if they break it, they pay for the next one. No big deal.
Yes, there are problems, but it should be looked at as a solution, even if it means all the cushy IT people might have to have there hours cut.
Let's assume that the fee is setup like ours is here: $100/semester.
Now, I'm sure a large school could get a deal with dell/compaq/hp for say $800 - 4x$200. If the student leaves early, they either turn their laptop in or they pay the rest for it. Otherwise, it's theirs.
Syncronize all of your files with a version encrypted with a randomly generated key. Not instant, but they all become effectively "erased". (Unless your using some type of versioning system)
Seems fairly trivial to include the state name as part of the "nationalization" of the DL. Also, most states , as far as I know, do not use the SSN for the DL. Not in cleartext anyway.
Then do the loop with pointers to functions, since you can just as easily implement it in a procedural language.
In fact, you can carry around any information want, call it whatever you want, and then do it however you want to. (Although this has more to do with weak typing than procedural).
OOP is a mindset. Another poster mentioned that you can write OO style in C. He's right, you can. About your example, sure, you can add up the flowrates across dozens of different types of pipes. And if you want "constnat fast", then set up a function table.
I will also point out, that these function tables of yours must always be used. So EVERY call has two references. Now we kind of need to look at what the common case is, and determine if this "add across all pipes" is what we do most often. If it isn't, then we have a lot of pain all the time for not much return.
And it could also be the case, as it is in most instances, that only one procedure is needed, since we hand in a finite element mesh to begin with. We might also note that there are two cases:
1) We have a butt ton of procedures, and thus a butt ton of calls. Most likely in loops. Now we have to pay for your constant lookup. And I still have a bunch of functions that don't make sense to apply to every object.
2) We have very few calls. In this case we have very few procedure, and thus don't have to worry about maintaining across groups of functions.
Something that has gotten lost in this argument is that often in engineering, you don't have functions that apply across large groups of objects. You simply take data, put it through some matrix operations, render it, and then spit it out. Thus you don't need "render class" and "matrix class" and any other class to get things done. Procedures in engineering are often to break up one long piece of code into smaller chunks, but they aren't called randomly or otherwise.
For the C++ version, I have to do a lookup of what the object is, and then a lookup of where the function is that is being called, and then run the code. So, we're probably looking at 2 memory references, and an indirect jump.
For the switch case, I have one memory refernce to determine type, and then a second refernece into a jump table. Thus I have 1 memory reference and 1 indirect jump, plus the original call to the function.
So they seem to be very similar in the number of jumps and lookups required. Thus I don't think it would even matter if there were 1000's of cases (but my function would be very long and messy).
And if the true question is to write faster, you'd have to look at your programmer's style, which is the same for maintaining. If I spread by objects around 50 different files in some haphazard method, then it really doesn't matter which language you're coding in, they both suck for maintanance.
Yes, since it sets a precedent. Legal maneuvers have always worked on precedent, even when they are only slightly related. Here is a law that has been upheld that says certain emails are not allowed. What's next? My guess is emailing anything that relates to pornography to people that might be underage.
And don't forget about banning encrypted emails that might allow terrorists to communicate.
Or methods of producing viruses or circumventing copyright schemes.
And if you're not allowed to email them, why should you be allowed to have any of these things on your computer?
It's just a step down the slippery slope, and while we might like the result, we really shouldn't like the bottom of the pit.
No, police are not monitoring, yet. However, if he legislators pass an anti-terrorism bill that outlaws encrypted email transmissions, then the next step will be police monitoring, and they have the anti-spam bills precedent of banning certain types of speech as precedent.
with procedural programming it's hard, since every procedure needs to know how to deal with every type of pipe at compile time.
No, only the procedures that you need to use need to be modified. If you don't need them, you don't change them. And changing procedures vs. writing new ones is an argument of a different sort, and can go either way as far as ease of use/ease of change.
You haven't saved any time in either case, and if you are using special cases, you can easily write macro wrappers (for example, making a round pipe from an oval pipe by
#define ROUNDCALC(r) ovalcalc(r,r)
)
So, yes, there are certain places OOP is a nice way to do things, but I think tying yourself to that mindset is a bad plan.
Slashdot Mirror
You don't need to worry about someone determining your scheme and starting to hunt through your ports using the naming scheme.
/etc/init.d/sendmail stop
apt-get update
apt-get install sendmail
/etc/init.d/sendmail start
Can't do an apt-get if the network is flooded with mail messages, can you?
If a "solved" problem like email actually brought the net down... for a while. How do you get patches for a sendmail program without using the internet?
Yes I was. Thanks. (I think it's because Spector was on Fox News Sunday.)
So really it comes down to asking if people can have an expectation of privacy.
I'm going to say it will probably be no. Police are allowed to monitor cordless phone conversations without a warrant, even though people might not realize that those conversations are being broadcast.
In a similar way, if I sit downstream for you on a cable line or LAN, I most certainly can listen to conversations without doing too much intrusion. Maybe people will begin to pay attention to security if this keeps happening.
Specter (R-Utah) was and is very much in favor of lifting copyright restrictions on file sharing and music. And I'm sure you'll find plenty of democrats who voted for the DMCA.
The whole "Democrats are for the little guy" would fly farther if Hollywood didn't like Democrats so much, and vice versa.
Never vote one party simply because you think that will help. Actually look at the candidates.
I just got a Compaq laptop, and while screwing around with it, ended up nuking the original XP that it came with. (I had repartitioned the drives incorrectly, and then had installed GRUB.)
Anyway, when I used the recover CDs, it removed the partions that I had made, but didn't touch the MBR!!
I would have thought that this would be a major oversite of a recovery disk-- isn't that the most famous place for a virus to sleep?
(Alls well now, I'm dual booting with no difficulty...)
Like this one.
I only read the first half... it's a very bizarre story.
Before badmouthing C
I was not bad mouthing C, as much as I was joking about the list of programming languages being looked at.
However, I'm still of the belief that C++ really gives you nothing over C, except another bulet on your resume.
Why did he look at C++ twice, I'm not sure it deserves a first look, let alone two.
Thanks (I'm not running Windows currently.)
I've been hunting for the past hour on microsoft's site and on google looking for the XP EULA, and I can't find it. I'm awaiting delivery of a laptop with XP on it, and I want to see if I need to delete everything or not when it shows up.
Where the hell did microsoft hide the EULA?
Sure, the very obvious bugs might be caught, but there's a catch:
If software has bugs that are easy to see, and come up frequently, it's perceived as "buggy", and few people will download it, and fewer eyes will look for the bugs.
If a bug is harder to reproduce, it probably doesn't come up very often, and not everyone will spend time looking for the bug. Hell, I'm in a research environment, I know how to code, and my KDE print daemon crashes everyday. I don't care enough to try to submit bug traq reports, or look at why it's crashing, I'm just going to hope that it magically gets fixed in the next release.
Exactly how many people here have actually hunted through source code to find the one little bug that annoys them, like the fact that Konq occasionally ignores links, or that it's javascript interpretter is far from perfect. Source code is not the magic bullet, but I'll admit it's better than nothing.
Interesting to note that there is no mention of the 'we didn't violate Russian law' argument
Except perhaps in the opening paragraph of the CNN headline:
arguing the law is vague, too broad and shouldn't apply to a foreign company.
Here is a laptop for $1000. Just one. The first I could find.
At our school, the $100/semester is expicity towards equipping the computing centers in the University, and making sure that they have enough systems for everybody. The support staff is paid for by other monies.
Now, I'm not going to attempt to solve the "Low-bidder" or "Bad-spec" jobs-- these are completely different and impossible problems that happen any time money is spent. And as for "being obsolete" -- what does that matter? If a laptop is useless in 4 years, who care? Exactly what can you buy that will not be obsolete in 4 years?
And as for people and maintenance, I'm not saying that the University needs to run an exchange program, just have the students deal with the manufacturer and have a waranty. if they break it, they pay for the next one. No big deal.
Yes, there are problems, but it should be looked at as a solution, even if it means all the cushy IT people might have to have there hours cut.
Let's assume that the fee is setup like ours is here: $100/semester.
Now, I'm sure a large school could get a deal with dell/compaq/hp for say $800 - 4x$200. If the student leaves early, they either turn their laptop in or they pay the rest for it. Otherwise, it's theirs.
It's not possible to hibernate a single process.
Wow, so the fact that its been done here is just a red herring?
Does Virtual Memory mean anything to you?
If you go to the screenshots, I'd swear that was Goku. It seems that DragonBall Z would have a bigger claim to infringement.
And I must admit, it does look really stupid.
Syncronize all of your files with a version encrypted with a randomly generated key. Not instant, but they all become effectively "erased". (Unless your using some type of versioning system)
Seems fairly trivial to include the state name as part of the "nationalization" of the DL. Also, most states , as far as I know, do not use the SSN for the DL. Not in cleartext anyway.
Then do the loop with pointers to functions, since you can just as easily implement it in a procedural language.
In fact, you can carry around any information want, call it whatever you want, and then do it however you want to. (Although this has more to do with weak typing than procedural).
OOP is a mindset. Another poster mentioned that you can write OO style in C. He's right, you can. About your example, sure, you can add up the flowrates across dozens of different types of pipes. And if you want "constnat fast", then set up a function table.
I will also point out, that these function tables of yours must always be used. So EVERY call has two references. Now we kind of need to look at what the common case is, and determine if this "add across all pipes" is what we do most often. If it isn't, then we have a lot of pain all the time for not much return.
And it could also be the case, as it is in most instances, that only one procedure is needed, since we hand in a finite element mesh to begin with. We might also note that there are two cases:
1) We have a butt ton of procedures, and thus a butt ton of calls. Most likely in loops. Now we have to pay for your constant lookup. And I still have a bunch of functions that don't make sense to apply to every object.
2) We have very few calls. In this case we have very few procedure, and thus don't have to worry about maintaining across groups of functions.
Something that has gotten lost in this argument is that often in engineering, you don't have functions that apply across large groups of objects. You simply take data, put it through some matrix operations, render it, and then spit it out. Thus you don't need "render class" and "matrix class" and any other class to get things done. Procedures in engineering are often to break up one long piece of code into smaller chunks, but they aren't called randomly or otherwise.
Well, let's compare.
For the C++ version, I have to do a lookup of what the object is, and then a lookup of where the function is that is being called, and then run the code. So, we're probably looking at 2 memory references, and an indirect jump.
For the switch case, I have one memory refernce to determine type, and then a second refernece into a jump table. Thus I have 1 memory reference and 1 indirect jump, plus the original call to the function.
So they seem to be very similar in the number of jumps and lookups required. Thus I don't think it would even matter if there were 1000's of cases (but my function would be very long and messy).
And if the true question is to write faster, you'd have to look at your programmer's style, which is the same for maintaining. If I spread by objects around 50 different files in some haphazard method, then it really doesn't matter which language you're coding in, they both suck for maintanance.
Does this still seem problamatic?
Yes, since it sets a precedent. Legal maneuvers have always worked on precedent, even when they are only slightly related. Here is a law that has been upheld that says certain emails are not allowed. What's next? My guess is emailing anything that relates to pornography to people that might be underage.
And don't forget about banning encrypted emails that might allow terrorists to communicate.
Or methods of producing viruses or circumventing copyright schemes.
And if you're not allowed to email them, why should you be allowed to have any of these things on your computer?
It's just a step down the slippery slope, and while we might like the result, we really shouldn't like the bottom of the pit.
No, police are not monitoring, yet. However, if he legislators pass an anti-terrorism bill that outlaws encrypted email transmissions, then the next step will be police monitoring, and they have the anti-spam bills precedent of banning certain types of speech as precedent.
with procedural programming it's hard, since every procedure needs to know how to deal with every type of pipe at compile time.
No, only the procedures that you need to use need to be modified. If you don't need them, you don't change them. And changing procedures vs. writing new ones is an argument of a different sort, and can go either way as far as ease of use/ease of change.
You haven't saved any time in either case, and if you are using special cases, you can easily write macro wrappers (for example, making a round pipe from an oval pipe by
#define ROUNDCALC(r) ovalcalc(r,r)
)
So, yes, there are certain places OOP is a nice way to do things, but I think tying yourself to that mindset is a bad plan.