...as opposed to the non-editable, non-text configuration files that Unix systems are famous for?
These are the people who named the editor "ed". Don't overthink it.
Re:This recipe is missing
on
MySQL Cookbook
·
· Score: 1
+4, Interesting? the recipe is linked in your post. change the defaults.
A snippet from the post in question:
There's no way (that I can find) to completely turn off non-transactional tables. As I understand it, if I forget to tell it when creating a table to make it transactional, it's silently not. If a transaction involves even a single non-transactional table, the whole thing is non-transactional. This makes me nervous.
Maybe it's just me, but I think that putting all your eggs in one basket is not smart. All it would take would be on critical vulnerability to be discovered and all of a sudden a potential attacker can get to all of your keys. Not good if you ask me.
I disagree. Right now, we're putting all our eggs in a bunch of half-assed baskets woven from tissue paper and lunchmeat. I'd much rather trust one well-audited, well-engineered solution than the 100 home rolled ones we have to trust now.
KDE does this now with KWallet (although without the spiffy kernel-level protections the author claims that Windows supports). If I'm writing a KDE application, I don't have to worry about getting password storage right - some other folks who know a whole lot more about the problem have already taken care of it for me.
I think this is good in the same way that using libc's strncmp is better than writing your own. Sure, there might be some undiscovered flaw lurking that's just waiting to open our systems to the world, and an environment of heterogeneous strncmp implementations would keep a successful attack from owning everything that links to libc. And yet, I have a lot of faith that the libc version is much better than anything I'm likely to come up with on my own.
Finally, if an error in strncmp were to be discovered, an upgrade of one library file would fix every dynamically linked program on my system. If each of those programs used their own, then each one would have to be audited to make sure they weren't broken in a similar way. In the same way, an upgrade to KWallet helps every program that uses it. Other programs have to hope that new vulnerabilities are specific to KWallet's own code and not a more general problem.
The Unix way is to build a tool that does one thing supremely well, then trust it. I think this is a prime candidate for the same treatment.
By the way, I'm only using KWallet as an example because I'm familiar with it; I'd be even more interested in Theo de Raadt getting a wild hair and writing OpenSecureStore some weekend.
At least I hope it didn't get removed just because the driver was "old" and "buggy"...
That's the best - perhaps the only - reason for removing drivers. It's not so much the age, as that that the kernel hardware abstraction layer evolves over time, and some of those old drivers are written against a hardware model that just doesn't exist anymore. If someone wanted to update the ftape driver so that it actually worked on the latest HAL, I'm sure someone would accept it back into the kernel. The fact that no one has done so is rather telling.
If you are using a hard-drive based player, however, then the number of disk reads is directly proportional to your disk rate, and your battery life is inversely proportional.
Are you sure about that? The platters will be spinning anyway in either case, and both involve long streaming reads of contiguous data. Would it really make a difference if a drive spent 2% of its available time reading FLAC blocks compared to 1% of its time reading an MP3?
You are right, but just a small nitpick--CDs aren't lossless. All digital audio is lossy, CDs are just so good that we've decided to use them as the standard.
Wrong. CDs are no more lossy (and usually far less so) than any other "normal" recording medium. They're made from a stream of discreet level readings instead of a perfectly smooth analog curve, but that's not the same as saying that they're lossy.
I disagree wholeheartedly. Did you notice the part that he's giving speeches for free? The gist of the rules seemed to be that if he's doing this, he doesn't want to spend his own money to provide a free service, or to be unreasonably uncomfortable in transit to the location.
Prima donnas have rules like "there must be exactly 23 brown M&Ms" and "[star] only drinks imported goat milk". Rules like "I don't want to spend 12 hours in economy class flights" and "I don't want to lose money on the deal" sound pretty kind.
I'm not trying to be a politician, though, so I don't have to impress anyone with my cool demeanor. Besides, media hot works at least as well as media cool, and I'm just not a "cool kid". I have to be a shit-hot individual instead.
Nothing wrong with that. Of course, it's quite appropriate to a thread defending ESR.
That was the first I heard of these. At first glance, though, they look pretty reasonable. He said he'll even sleep on someone's couch if that's all the lodging someone can arrange. What part of his list did you take issue with?
I don't have a strong opinion either way on your other items, but I didn't see anything on this one that would make me think poorly of him.
The system is fucking stupid because it still leaves you in dependency hell. I don't really know how people can find the stomach to dispute that. It's like when some program bluescreens windows and people make excuses for it. "Well the program did such and such"... fuck you!
Know what I like about you, Martin? We can never tell what you're really thinking. You bear the inscrutability of the Orient, the stolid face of The Gambler. Very few will ever discern your true feelings from the subtle hints you leave behind.
Additionally, the cellphone makers are leasing public property (the airwaves) and building a fence around them to keep the public out (unless you buy a key / plan from them)
Ignoring the rest, is there a reason they shouldn't be able to restrict access to something they're leasing? I mean, if you lease a house, should everyone else get to use it?
I know what you're getting at, but your wording there is a bit problematic.
Perl 5's requirement for taking explicit references to aggregates confuses/frustrates me.
You meant "references stored in scalars", just as I said. I'd agree with the rest of your statement, too: I'm confused by the idea of a modern high-level language being dependent on explicit reference manipulation for common tasks, and frustrated that I have to do that in Perl but not in Python or Ruby (or Java, for that matter).
I don't hate Perl and I'm competent with it, but if I wanted to deal with that level of bookkeeping, I'd probably just write C and be done with it.
Is it just me, or is it possible to create perfectly legible code in Perl if you use good technique, just like in any other language?
No. It's handling of complex data structures is syntactically nightmarish. In Python, a list of dicts (Perl would call them hashes) looks exactly like a list of hashes:
Voila. That's it. You now know how to write any list, a nested list, a dict, a dict of lists, or any other combination you can think of. Furthermore, you can pass any of those as an argument to a function, or return any of them from a function - with no extra syntax whatsoever:
def foo(bar): print bar; return [[4, 5], [6, 7]]
print foo({'foo': [1, 2, 3]})
# Result: # {'foo': [1, 2, 3]} # [[4, 5], [6, 7]]
Now, for the exact opposite, read the perldoc page for "perlref" and see how Perl likes you to pass around data structures. Hint: you have to cast everything to the "scalar" datatype before you can pass it in or out of a function.
These million-and-one special cases are why I abandoned Perl for Python. Yes, it's theoretically possible to write clean Perl code, but the language has so many artificial barriers to doing so that it's practically impossible. The best you can reasonably hope for is that your group, project, or employer chooses a set of coding standards that tells you how to handle every non-standard gotcha that arises so that you'll recognize them when (not if) you see them.
Holy crap a revolution! I get my radio for FREEEEEE!
I hate those stupid ads - are they a Clearchannel specialty? - about not paying for something you can get for free. Well, where I live, I have the choice of Clearchannel Pop, CC Rock, CC Country 1-5, and CC "greatest hits of 80s, 90s, and now". So, I opted for Sirius to hear the great stuff I can't hear over the airwaves here.
I'm too cheap to ever pay for anything I could get for free. I don't drink bottled water, but I do pay for satellite radio.
But good UI design isn't primarily about giving you what you like, it's about giving you what makes you and other users work most effectively, and that frequently differs from what you like.
Yes, but when 90% of people asked say, for example, that they despise the Gnome file dialog, maybe it's time to step back and reevaluate the decision. Yes, in theory, Gnome's choices might be better than what people think they want. But when the result is massively unpopular, that should be a huge warning sign. They can't just keep saying "try it some more and you'll start to like it" forever and hope that eventually everyone decides they were right.
Well, apparently enough people like Gnome that distributions that have made it the primary desktop choice prosper and that lots of people contribute a lot of time and effort to it.
I'll admit that it may be more appropriate for new users, and therefore a reasonable choice for distros that are aiming for wide migration away from Windows. I just honestly don't see that lasting forever.
If you yourself really don't like it, just don't use it; there are plenty of alternatives.
This is true. Perhaps Gnome should be asking themselves why people are migrating to them en masse as they begin to get comfortable on Linux.
If people like Linus and you want to help, learn something about UI design first. Then, you can either contribute suggestions for specific improvements justified based on accepted UI design criteria, or you can participate in user testing. Your and Linus's opinions, on the other hand, are pretty much worthless.
That's right, kids: unless you become an expert on how other people might use a computer, you are unqualified to have an opinion on how you would like to use it. And never forget that even if you hate a design decision, all of your friends hate it, and everyone on the Internet seems to hate it, your collective opinions are worthless compared to that of a person who once read somewhere that it was a really good idea.
Oops I was under the impression that the linux drivers were just the ported win32 ones and they had an abstraction layer to support multiple platforms.
They are. Well, more accurately, the drivers are identical - neither is the "native" version. There's no such thing as the "win32 ones", since they're written against the same HAL.
Ergo, that board is officially dropped on all platforms as of the new 97xx series of drivers. Don't blame Linux; they've cut off support universally.
No, it's not. The "supported" page in the README is out of date. Whenever I tried to use the latest drivers, X failed to start with an "unsupported card" message in Xorg.0.log.
From the site, it would appear that the card will still work. From the driver itself, it doesn't.
What is so bad about including the proprietary drivers.
It breeds complacency. My home desktop has an old GeForce MX 400 card which still works perfectly well. It renders my 1600x1200 desktop cleanly and quickly, and basically does what I want it to. I don't have a strong need for OpenGL but do like to play games occasionally (eg Quake or Second Life) and although it's not fast, it worked perfectly.
Note I said "worked". Nvidia has officially deprecated my card, so no new drivers will ever support it. New kernel with an incompatible ABI? I can't upgrade to it. Security vulnerability? I can't get the fix. Basically, I can either keep using my system in its current state forever, or buy a new card purely for the driver upgrade.
Yes, I know my card is old and slow by today's standards. But if it works for me and I'm happy with it, why should I have to replace it? Given that my motherboard has an old Via chipset that Nvidia only supports in AGP 2x mode and that new cards are all but impossible to get working (I've tried), I'm looking at a complete system upgrade just to get a new driver.
With a Free driver, in the worst case situation I could at least attempt to fix new problems on my own as they arise. With closed drivers, I have no control whatsoever. I like Free software for philosophical reasons, but it also has huge practical advantages. This is one of them.
Today, it is entirely up to Microsoft and if you dare wander outside their edicts and trigger their damned "tiltbits", you are fucked.
In other words, you have to play nicely like everyone else and you don't like it? Please forgive my total lack of sympathy for you having to make your kernel code behave well.
Re:Opposite way of thinking?
on
PHP 5 in Practice
·
· Score: 2, Insightful
Is that really true? I find myself more the opposite where I know how to solve a problem theoretically but I need to know the exact syntax (and sometimes even the libraries/classes already available) to allow me to do what I want. Is this because I'm only a recent (2 yrs) CS grad?
Yes, but I think it also speaks to the level of challenge to your comfort you've been faced with. There are three major roadblocks programmers run into:
Language syntax. This should mostly cure itself over time.
Algorithm choice. Again, experience tends to take care of this, regardless of the language you're working with today.
Best practice. Unless you become one of the experts in your field, you should be asking your self how those people have solved the same problem you're working on. The best algorithms and perfect syntax won't help you apply a bad solution to the problem at hand.
Cookbooks mitigate the last of these. Sometimes you just want to be told how to best solve problem X - the syntax and algorithms will come naturally afterward.
Slashdot Mirror
...as opposed to the non-editable, non-text configuration files that Unix systems are famous for?
These are the people who named the editor "ed". Don't overthink it.
A snippet from the post in question:
Where's the recipe for fixing that problem?
Exactly, just like Apache has been a horrible smoking crater of wormy infection compared to the far less widely used IIS. Oh, wait...
I'm not saying that IIS is awful (anymore), but it's clear that marketshare does not automatically mean a higher rate of compromise.
I disagree. Right now, we're putting all our eggs in a bunch of half-assed baskets woven from tissue paper and lunchmeat. I'd much rather trust one well-audited, well-engineered solution than the 100 home rolled ones we have to trust now.
KDE does this now with KWallet (although without the spiffy kernel-level protections the author claims that Windows supports). If I'm writing a KDE application, I don't have to worry about getting password storage right - some other folks who know a whole lot more about the problem have already taken care of it for me.
I think this is good in the same way that using libc's strncmp is better than writing your own. Sure, there might be some undiscovered flaw lurking that's just waiting to open our systems to the world, and an environment of heterogeneous strncmp implementations would keep a successful attack from owning everything that links to libc. And yet, I have a lot of faith that the libc version is much better than anything I'm likely to come up with on my own.
Finally, if an error in strncmp were to be discovered, an upgrade of one library file would fix every dynamically linked program on my system. If each of those programs used their own, then each one would have to be audited to make sure they weren't broken in a similar way. In the same way, an upgrade to KWallet helps every program that uses it. Other programs have to hope that new vulnerabilities are specific to KWallet's own code and not a more general problem.
The Unix way is to build a tool that does one thing supremely well, then trust it. I think this is a prime candidate for the same treatment.
By the way, I'm only using KWallet as an example because I'm familiar with it; I'd be even more interested in Theo de Raadt getting a wild hair and writing OpenSecureStore some weekend.
That's the best - perhaps the only - reason for removing drivers. It's not so much the age, as that that the kernel hardware abstraction layer evolves over time, and some of those old drivers are written against a hardware model that just doesn't exist anymore. If someone wanted to update the ftape driver so that it actually worked on the latest HAL, I'm sure someone would accept it back into the kernel. The fact that no one has done so is rather telling.
Are you sure about that? The platters will be spinning anyway in either case, and both involve long streaming reads of contiguous data. Would it really make a difference if a drive spent 2% of its available time reading FLAC blocks compared to 1% of its time reading an MP3?
Wrong. CDs are no more lossy (and usually far less so) than any other "normal" recording medium. They're made from a stream of discreet level readings instead of a perfectly smooth analog curve, but that's not the same as saying that they're lossy.
I disagree wholeheartedly. Did you notice the part that he's giving speeches for free? The gist of the rules seemed to be that if he's doing this, he doesn't want to spend his own money to provide a free service, or to be unreasonably uncomfortable in transit to the location.
Prima donnas have rules like "there must be exactly 23 brown M&Ms" and "[star] only drinks imported goat milk". Rules like "I don't want to spend 12 hours in economy class flights" and "I don't want to lose money on the deal" sound pretty kind.
Nothing wrong with that. Of course, it's quite appropriate to a thread defending ESR.
That was the first I heard of these. At first glance, though, they look pretty reasonable. He said he'll even sleep on someone's couch if that's all the lodging someone can arrange. What part of his list did you take issue with?
I don't have a strong opinion either way on your other items, but I didn't see anything on this one that would make me think poorly of him.
Know what I like about you, Martin? We can never tell what you're really thinking. You bear the inscrutability of the Orient, the stolid face of The Gambler. Very few will ever discern your true feelings from the subtle hints you leave behind.
Ignoring the rest, is there a reason they shouldn't be able to restrict access to something they're leasing? I mean, if you lease a house, should everyone else get to use it?
I know what you're getting at, but your wording there is a bit problematic.
"Wiiwoaded"?
You meant "references stored in scalars", just as I said. I'd agree with the rest of your statement, too: I'm confused by the idea of a modern high-level language being dependent on explicit reference manipulation for common tasks, and frustrated that I have to do that in Perl but not in Python or Ruby (or Java, for that matter).
I don't hate Perl and I'm competent with it, but if I wanted to deal with that level of bookkeeping, I'd probably just write C and be done with it.
No. It's handling of complex data structures is syntactically nightmarish. In Python, a list of dicts (Perl would call them hashes) looks exactly like a list of hashes:
Voila. That's it. You now know how to write any list, a nested list, a dict, a dict of lists, or any other combination you can think of. Furthermore, you can pass any of those as an argument to a function, or return any of them from a function - with no extra syntax whatsoever:
Now, for the exact opposite, read the perldoc page for "perlref" and see how Perl likes you to pass around data structures. Hint: you have to cast everything to the "scalar" datatype before you can pass it in or out of a function.
These million-and-one special cases are why I abandoned Perl for Python. Yes, it's theoretically possible to write clean Perl code, but the language has so many artificial barriers to doing so that it's practically impossible. The best you can reasonably hope for is that your group, project, or employer chooses a set of coding standards that tells you how to handle every non-standard gotcha that arises so that you'll recognize them when (not if) you see them.
I hate those stupid ads - are they a Clearchannel specialty? - about not paying for something you can get for free. Well, where I live, I have the choice of Clearchannel Pop, CC Rock, CC Country 1-5, and CC "greatest hits of 80s, 90s, and now". So, I opted for Sirius to hear the great stuff I can't hear over the airwaves here.
I'm too cheap to ever pay for anything I could get for free. I don't drink bottled water, but I do pay for satellite radio.
Yes, but when 90% of people asked say, for example, that they despise the Gnome file dialog, maybe it's time to step back and reevaluate the decision. Yes, in theory, Gnome's choices might be better than what people think they want. But when the result is massively unpopular, that should be a huge warning sign. They can't just keep saying "try it some more and you'll start to like it" forever and hope that eventually everyone decides they were right.
That's right, kids: unless you become an expert on how other people might use a computer, you are unqualified to have an opinion on how you would like to use it. And never forget that even if you hate a design decision, all of your friends hate it, and everyone on the Internet seems to hate it, your collective opinions are worthless compared to that of a person who once read somewhere that it was a really good idea.
No. Stephen Hawking's love child, perhaps, but not 'shopped.
They are. Well, more accurately, the drivers are identical - neither is the "native" version. There's no such thing as the "win32 ones", since they're written against the same HAL.
Ergo, that board is officially dropped on all platforms as of the new 97xx series of drivers. Don't blame Linux; they've cut off support universally.
Oh, and I'm not the only one who had that problem. The new 97xx drivers have officially obsoleted my card, and by extension, my entire system.
No, it's not. The "supported" page in the README is out of date. Whenever I tried to use the latest drivers, X failed to start with an "unsupported card" message in Xorg.0.log.
From the site, it would appear that the card will still work. From the driver itself, it doesn't.
It breeds complacency. My home desktop has an old GeForce MX 400 card which still works perfectly well. It renders my 1600x1200 desktop cleanly and quickly, and basically does what I want it to. I don't have a strong need for OpenGL but do like to play games occasionally (eg Quake or Second Life) and although it's not fast, it worked perfectly.
Note I said "worked". Nvidia has officially deprecated my card, so no new drivers will ever support it. New kernel with an incompatible ABI? I can't upgrade to it. Security vulnerability? I can't get the fix. Basically, I can either keep using my system in its current state forever, or buy a new card purely for the driver upgrade.
Yes, I know my card is old and slow by today's standards. But if it works for me and I'm happy with it, why should I have to replace it? Given that my motherboard has an old Via chipset that Nvidia only supports in AGP 2x mode and that new cards are all but impossible to get working (I've tried), I'm looking at a complete system upgrade just to get a new driver.
With a Free driver, in the worst case situation I could at least attempt to fix new problems on my own as they arise. With closed drivers, I have no control whatsoever. I like Free software for philosophical reasons, but it also has huge practical advantages. This is one of them.
In other words, you have to play nicely like everyone else and you don't like it? Please forgive my total lack of sympathy for you having to make your kernel code behave well.
Yes, but I think it also speaks to the level of challenge to your comfort you've been faced with. There are three major roadblocks programmers run into:
Cookbooks mitigate the last of these. Sometimes you just want to be told how to best solve problem X - the syntax and algorithms will come naturally afterward.