...and nobody blames the real culprits: whichever idiots designed the Flash player, browser and OS so that an advertisement on any random, untrusted website can download things to your machine without your permission. It's 2007. If we have to have things like the Flash player, why on earth doesn't it run in a sandbox with no access to your files? Why doesn't every single window popped up by Flash have a mandatory button 'go away and don't ever show me popups from this Flash file again'?
My father in law showed me how he uses dowsing rods. He takes hefty copper wire (about 8 gauge or so,) cuts it into two pieces each about half a meter long,
Clearly he'd get much better results with Monster cable rather than unbranded copper wire.
There could be trouble between Google and Sun, according to someone. Google's Dalvik has advantages, according to some people, but it also has disadvantages. According to a developer not working for Google or Sun, it is possible that Google didn't pick Dalvik for technical reasons, although we don't know. There could be trouble for Google, say some people, because of 'intellectual property' Google may or may not have used, although we don't know what that 'intellectual property' might be. Stefano Mazzocchi says he doesn't know what Sun would do, but he is curious. Nobody working for Google would comment. Nobody working for Sun would comment.
Next week: why some people say that Microsoft could be in unexpected difficulties if it launches its own Linux distribution, which many observers have seen as likely, although others disagree.
um $5 billion I meant. Probably wasting all of that money is not feasible but still a company can 'spend' money by paying inflated prices to its subsidiaries, etc.
That's a recipe for corruption and mandated stupidity; a company could bid to build $5T 'worth' of infrastructure and then artificially inflate costs to meet that level by paying exorbitant salaries to the boss's nephew, etc. The wireless spectrum is a scarce resource, and one way to allocate it (not the only way, but one that seems to work well for other scarce resources such as land) is for companies to pay money for the part they want. This at least has the advantage that it's free of political considerations; ten dollars is more than nine dollars, no question about it, and it doesn't make a difference if the nine dollars was offered by someone who has good connections in the White House, or is well-known on TV, etc.
Ah, Ghostbusters. I remember that on the Spectrum. The weird crackling noise at the start which someone explained to me was meant to be a speech sample. Driving around the map running over ghosts so that later (in what now seems a masterpiece of boring gameplay) you could suck them up with your ghost vacuum. The boxing glove on a spring which jumped up to catch ghosts (was that in the movie?). The rather anticlimactic Marshmallow Alert when all you had to do was put down ghost bait and you could run him over on the map screen. And what was the point of the game anyway? It ends when the ectoplasm count reaches 999, but should you try to slow that down or speed it up by letting ghosts reach Zuul so you can get to the end quicker?
Some exercise professionals recommend that instead of sitting on a chair at work, you sit on a Swiss ball. Then your core muscles have to work moderately hard to keep you stable. I haven't tried it but it's said to be fairly tiring to do it all day long.
If you prefer to just burn some energy, there's always the desk treadmill.
Yes, configure your database to listen only for connections on localhost. If you are moronic enough to leave it listening to the whole world and choose obvious passwords (or put the password directly in your program code which may then get leaked), then I guess a firewall would help, but it would be better just not to be stupid in the first place.
I agree with most of the points you mentioned, I just disagree with the tone taken by the article which is 'OMG database server with no firewall!'. If it's securely configured, it doesn't need one. And if it's badly set up, then you should just apply some common sense and fix it... sadly this doesn't generate enough money for firewall vendors.
I also don't like the way it refers to 'without protection' as if security were some thing to be applied afterwards like a coat of paint. I guess it's psychologically appealing to have some button in a control panel that says 'enable protection' and you don't have to think about it.
If you were writing to a file then yes, it is important to check that the write was successful -- after all, you might have run out of disk space or something. But then you would be using fprintf() rather than printf().
Not necessarily, e.g. gzip <in >out. You can't really assume that your standard output is unimportant or that failures to write it can be ignored. Of course if your application is just generating chatter messages perhaps it doesn't matter if they get lost; but those should probably go to stderr anyway.
My point (which, I'm pleased to note, managed not to escape at least one person with mod points) was that most of the time, STDOUT and STDERR are the same device. So if printf() goes tits-up, it's highly unlikely that any attempt to display an error message will work either. (I suppose you could exit non-zero, but who checks exit codes in shell scripts?)
Well, quite. But do the right thing and at least the user will have the possibility to write robust scripts calling your program. If they choose to write flaky shell scripts that plough on ignoring errors, that's their choice, but don't force it on them.
Myself I tend to write perl scripts more than shell scripts and always check the exit status (if only to die with an error message).
If you were talking about a genuinely separate network - air wall security - then I'd say you are often right. But I don't really see the difference between having a database server on a publicly accessible network with tight login security, versus putting it on a so-called 'private network' and then setting up VPNs and tunnels so that it's publicly accessible after all. That just moves all the authentication into the VPN. It's an advantage if you don't trust your database server to behave securely, but it would be better just to have a sensible authorization mechanism such as Kerberos on the database.
So what? Do you really think the number of characters written is important, interesting, or vital to security in some way? More importantly, what will you do if you find that not all characters were written?
Do *something*, anything to flag the error and alert the user - do not just continue blindly and pretend all is well. Few bugs are more annoying than the one that writes a zero-length output file and then tells you your data has been saved.
In state machine construction, there are two choices to make for invalid events: ignore or can't happen.
This may be true, but failure to printf() is not an invalid event. It's an entirely possible outcome of calling printf(), one that is documented in the manual page. You need to design your state machine so that there is an appropriate state and transitions for the 'output failed' case.
Failure to write a file is not something that 'cannot happen'. And usually, it is not something you want to 'ignore'.
There are cases where it is perfectly valid to ignore the return event of a printf.
Indeed, but it's a good rule of thumb to always check it and do *something* to handle errors, unless you specifically know that output failures aren't cared about here.
You can't correct it, but you should at least notify the user rather than continuing blindly. For example, if you are writing to an output file with printf() and the write fails, you shouldn't go on to tell the user that the file was saved successfully.
For 'almost impossible' conditions, dying immediately with an error message is maybe not ideal, but still a hundred times better than silently ignoring the error and reporting success.
Why on earth should a database server need a firewall? Last time I looked, DBMSes required a login with a username and password before giving any access. I hope that the days of default passwords like scott/tiger are long gone, and if not, you should get a more secure database rather than masking the problem with a firewall (which does nothing to protect against internal attacks).
Hopefully the DBMS supports SSL or other encrypted connections so outsiders can't eavesdrop or hijack sessions.
'Too expensive' is obviously a personal judgement. What's too expensive for me is probably not too expensive for a millionaire who enjoys spending money. If you choose not to buy something, very often the reason is that it is too expensive: not good enough to be worth spending that money on, given the amount of money you have to spend and the other things you want to buy. This is not in any way an attack on Apple or a claim that the fantastic multi-touch technology didn't require a great deal of research, development and manufacturing skill.
Re:Yet ANOTHER sound server?
on
Fedora 8 Released
·
· Score: 2, Funny
Is anyone collecting a list of bad free software project names? My favourite is 'RabbitWare Linux' which was, alas, never released.
I shouldn't have to compete with every twit in every 3rd world country willing to work for the cost of a soda a day.
Maybe you think that, but the person paying the wages probably thinks otherwise. Since they are paying the bills, their opinion counts. You are, of course, free to start a company employing only workers on Western salaries if you believe that's how it should be. Let whichever provides the best service win.
Well, quite. If Java doesn't have a good-quality, free implementation then I'll dump Java and use something else instead. gcj and gij are heroic efforts but they were always trying to catch up to a semi-proprietary standard.
'closed source' is an inane term. I don't think anyone from gcj or gij was describing Sun's Java as 'closed source'. It's non-free, which is what matters. Merely being able to look at the source code doesn't mean you have freedom to use, share and change the software.
Why do you want to design a system to 'protect people who do something new'? Surely the system should be run to promote progress in science and the useful arts, as specified in the US Constitution?
Maybe it does promote progress to have patents on software, but it's not a foregone conclusion; study some of the arguments (there may also be a good site arguing in favour of swpats, but I don't know of one) and decide what works best in the public interest, rather than just assuming that any measure in favour of 'inventors' is going to help the public.
But from an individual site's point of view, it does make a big difference to have your MTA drop incoming connections immediately on getting an invalid address, rather than accept the mail and send back a soft bounce. Lots of spam is sent to random.address@known.site in the hope of getting somewhere. While accepting these messages ties up the spammer's resources, it also ties up your machine's resources.
Slashdot Mirror
...and nobody blames the real culprits: whichever idiots designed the Flash player, browser and OS so that an advertisement on any random, untrusted website can download things to your machine without your permission. It's 2007. If we have to have things like the Flash player, why on earth doesn't it run in a sandbox with no access to your files? Why doesn't every single window popped up by Flash have a mandatory button 'go away and don't ever show me popups from this Flash file again'?
Has gcc been ported to a GPU yet? Can you compile kernels (or Gentoo) on your video card?
There could be trouble between Google and Sun, according to someone. Google's Dalvik has advantages, according to some people, but it also has disadvantages. According to a developer not working for Google or Sun, it is possible that Google didn't pick Dalvik for technical reasons, although we don't know. There could be trouble for Google, say some people, because of 'intellectual property' Google may or may not have used, although we don't know what that 'intellectual property' might be. Stefano Mazzocchi says he doesn't know what Sun would do, but he is curious. Nobody working for Google would comment. Nobody working for Sun would comment.
Next week: why some people say that Microsoft could be in unexpected difficulties if it launches its own Linux distribution, which many observers have seen as likely, although others disagree.
um $5 billion I meant. Probably wasting all of that money is not feasible but still a company can 'spend' money by paying inflated prices to its subsidiaries, etc.
That's a recipe for corruption and mandated stupidity; a company could bid to build $5T 'worth' of infrastructure and then artificially inflate costs to meet that level by paying exorbitant salaries to the boss's nephew, etc. The wireless spectrum is a scarce resource, and one way to allocate it (not the only way, but one that seems to work well for other scarce resources such as land) is for companies to pay money for the part they want. This at least has the advantage that it's free of political considerations; ten dollars is more than nine dollars, no question about it, and it doesn't make a difference if the nine dollars was offered by someone who has good connections in the White House, or is well-known on TV, etc.
Ah, Ghostbusters. I remember that on the Spectrum. The weird crackling noise at the start which someone explained to me was meant to be a speech sample. Driving around the map running over ghosts so that later (in what now seems a masterpiece of boring gameplay) you could suck them up with your ghost vacuum. The boxing glove on a spring which jumped up to catch ghosts (was that in the movie?). The rather anticlimactic Marshmallow Alert when all you had to do was put down ghost bait and you could run him over on the map screen. And what was the point of the game anyway? It ends when the ectoplasm count reaches 999, but should you try to slow that down or speed it up by letting ghosts reach Zuul so you can get to the end quicker?
Good theme music though.
Some exercise professionals recommend that instead of sitting on a chair at work, you sit on a Swiss ball. Then your core muscles have to work moderately hard to keep you stable. I haven't tried it but it's said to be fairly tiring to do it all day long.
If you prefer to just burn some energy, there's always the desk treadmill.
Yes, configure your database to listen only for connections on localhost. If you are moronic enough to leave it listening to the whole world and choose obvious passwords (or put the password directly in your program code which may then get leaked), then I guess a firewall would help, but it would be better just not to be stupid in the first place.
I agree with most of the points you mentioned, I just disagree with the tone taken by the article which is 'OMG database server with no firewall!'. If it's securely configured, it doesn't need one. And if it's badly set up, then you should just apply some common sense and fix it... sadly this doesn't generate enough money for firewall vendors.
I also don't like the way it refers to 'without protection' as if security were some thing to be applied afterwards like a coat of paint. I guess it's psychologically appealing to have some button in a control panel that says 'enable protection' and you don't have to think about it.
Not necessarily, e.g. gzip <in >out. You can't really assume that your standard output is unimportant or that failures to write it can be ignored. Of course if your application is just generating chatter messages perhaps it doesn't matter if they get lost; but those should probably go to stderr anyway.Well, quite. But do the right thing and at least the user will have the possibility to write robust scripts calling your program. If they choose to write flaky shell scripts that plough on ignoring errors, that's their choice, but don't force it on them.
Myself I tend to write perl scripts more than shell scripts and always check the exit status (if only to die with an error message).
If you were talking about a genuinely separate network - air wall security - then I'd say you are often right. But I don't really see the difference between having a database server on a publicly accessible network with tight login security, versus putting it on a so-called 'private network' and then setting up VPNs and tunnels so that it's publicly accessible after all. That just moves all the authentication into the VPN. It's an advantage if you don't trust your database server to behave securely, but it would be better just to have a sensible authorization mechanism such as Kerberos on the database.
Failure to write a file is not something that 'cannot happen'. And usually, it is not something you want to 'ignore'.Indeed, but it's a good rule of thumb to always check it and do *something* to handle errors, unless you specifically know that output failures aren't cared about here.
You can't correct it, but you should at least notify the user rather than continuing blindly. For example, if you are writing to an output file with printf() and the write fails, you shouldn't go on to tell the user that the file was saved successfully.
For 'almost impossible' conditions, dying immediately with an error message is maybe not ideal, but still a hundred times better than silently ignoring the error and reporting success.
Why on earth should a database server need a firewall? Last time I looked, DBMSes required a login with a username and password before giving any access. I hope that the days of default passwords like scott/tiger are long gone, and if not, you should get a more secure database rather than masking the problem with a firewall (which does nothing to protect against internal attacks).
Hopefully the DBMS supports SSL or other encrypted connections so outsiders can't eavesdrop or hijack sessions.
Um... check the parent of my post. That's the comment spam. Nothing to do with OpenSolaris, Sun etc.
Is this the first comment spam on Slashdot? It's strange, I haven't seen it before.
Why, yes, I _was_ born on February 30th, you insensitive clod!
'Too expensive' is obviously a personal judgement. What's too expensive for me is probably not too expensive for a millionaire who enjoys spending money. If you choose not to buy something, very often the reason is that it is too expensive: not good enough to be worth spending that money on, given the amount of money you have to spend and the other things you want to buy. This is not in any way an attack on Apple or a claim that the fantastic multi-touch technology didn't require a great deal of research, development and manufacturing skill.
Is anyone collecting a list of bad free software project names? My favourite is 'RabbitWare Linux' which was, alas, never released.
I think that calling it 'not open source' is fine but the term 'closed source' sounds lame (IMHO) and is just likely to confuse people.
Well, quite. If Java doesn't have a good-quality, free implementation then I'll dump Java and use something else instead. gcj and gij are heroic efforts but they were always trying to catch up to a semi-proprietary standard.
'closed source' is an inane term. I don't think anyone from gcj or gij was describing Sun's Java as 'closed source'. It's non-free, which is what matters. Merely being able to look at the source code doesn't mean you have freedom to use, share and change the software.
Why do you want to design a system to 'protect people who do something new'? Surely the system should be run to promote progress in science and the useful arts, as specified in the US Constitution?
Maybe it does promote progress to have patents on software, but it's not a foregone conclusion; study some of the arguments (there may also be a good site arguing in favour of swpats, but I don't know of one) and decide what works best in the public interest, rather than just assuming that any measure in favour of 'inventors' is going to help the public.
But from an individual site's point of view, it does make a big difference to have your MTA drop incoming connections immediately on getting an invalid address, rather than accept the mail and send back a soft bounce. Lots of spam is sent to random.address@known.site in the hope of getting somewhere. While accepting these messages ties up the spammer's resources, it also ties up your machine's resources.