You mean the one they initially didn't want to fix unless you could somehow prove that it might affect you? I'd be hard pressed to think of a worse example of this magical Intel warranty support you're so proud of.
They may indeed be better now; if so, then you should have picked a case that didn't make people hate them.
...as long as you're paying cash and aren't being treated for any illness requiring health department notification. If you've got the measles, though, or paying with any insurance remotely connected to the government, then be prepared to cough up those digits (even if you're not currently choking on ladyfingers).
Yes, the only "solution" is to keep making the jar bigger. which is why they go with SHA256.
Not even close.
The ideal hash function has an equal probability of all possible results for any given input. Let's look at a non-ideal one: h(x) = 1. Now, any input results in the same hash value. Is that email signature valid? Sure: it has the expected hash! Did that ISO come through intact? Sure: the hash matches! Wouldn't be very useful, would it? In this simplified example, returning a 1024-bit hash value would be exactly as useful as returning a 1-bit hash value.
This is why MD5 is deprecated. It's not that it returns too few bits (which it does by modern standards), but that it's been shown to not return an ideal distribution of results. SHA is closer to the ideal as far as we know today, and SHA-256 adds extra bits in addition to the improved algorithm.
I bought a type M in excellent condition. I loved the feel of the keys, but the function keys are higher above the type row of the keyboard than I'm used to, and since I use Alt+Fn to switch between virtual desktops constantly, I was always having to do this weird hand-twist-and-stretch motion. Within two months it was physically painful.
No, as much as I wanted to have a ubergeek keyboard, my hands just weren't up to the task (and years of muscle memory means that I'm not willing to re-learn new type M-compatible chords). I think I'll be picking up a second Happy Hacking keyboard instead.
I won't go into exact numbers, but in our case a 10% drop in business would equate to about a 25% drop in personal income. That's probably pretty typical for single-provider practices outside the glamorous specialties. Many doctors, especially ones fresh out of med school with crippling student loans, just don't have the option of making those kinds of cuts.
This isn't really an issue for us since, as I mentioned, my wife runs nearly on time almost all the time.
Not too sure what this has to do with Apache configuration. Putting the code inside the configuration file itself isn't exactly the solution I'm looking for, more like being able to modify the config from outside of it using scripts.
I'm not sure I really see the difference. Think of "httpd.conf" as "config.pl" that gets executed by starting Apache and you're there. Either way you're dynamically generating the configuration; the difference is whether you write it out to a file first or directly altar Apache's state.
His solution - he erased every single comment from httpd.conf.
I worked for a small ISP, and we used a heavily-commented named.conf and associated zone files to keep track of configuration information, explanations for non-obvious things, etc. Since we were a small shop and worked well together, this was fine. Until we merged with another ISP. Whose admins "helpfully" slaved their BIND to ours, made it the master, and then slaved ours to theirs. Without changing the zone filenames in named.conf. I think that, had I been in the same room with them before we managed to retrieve our off-site backups, I might've had to have killed them.
One hugely useful system is to use version control on your configuration directories. Let the new guy delete at will; you can always roll back his commits and explain why you're going to beat him if he does it again.
What Rutan did is admirable and he deserves all the credit he has recieved, but the comparison's to NASA are just silly.
But you're speaking as though Rutan is finished, which doesn't seem to be the case. And even if he were to quit today, several other private groups are pursuing the same goals. There was a lot more buzz about SpaceShipOne than the recent "Hey, it didn't explode!" shuttle flight, and I imagine the budget folks at NASA wish some of that excitement was aimed in their direction.
For most of the last five decades, NASA has been the sole North American spacefaring entity. That changed last year, and I don't see any reason to believe that Burt or John Carmack or someone else can't have as much success in the next five.
My first thought was that they're trying to derail Burt Rutan by hopelessly outclassing him, even if only in advertising. We Slashdotters should be familiar enough with this strategy: "Don't buy OS X - Vista has these features that will be far better (assuming we don't drop them)!" I wouldn't be completely surprised to find that NASA mainly just wants to steal Burt's thunder in order to keep Congress from asking pesky questions like whether civilians should be able to compete in the deep space arena.
What saddens me most is that I don't really have much faith in them anymore. When I was growing up in the 70's, the folks at NASA were my heroes. They were the smartest, most determined, and best people anywhere in the world. I kind of wish I had that back, but at least private industry has given us a few new heroes to live vicariously through.
You know it's named after a catalytic process. I know it's named after a catalytic process. The rest of the world knows the guy has "kat" right there in the name of the company.
Choosing "Alphakatalytisch" might've saved the good doctor a fair amount of grief.
The first generation will be "grammar", courtesy of GNU.
Next will come "Grammar" (Gnome) and "grammatiK" (guess).
Within the following year we'll see "iWordflow", and six months past that (after a flaming rant and a two-week hackathon) everyone will upgrade to "OpenGrammar".
You'll eventually see XFGrammar for people who think Grammar and Grammatik are too bloated.
Why not describe it in a text file, but formatted as an SVG or a graphviz input file? They're easy enough to read that anyone with Notepad or better can understand and update them, but trivially compile to pretty pictures that you can give to management or hang on the wall.
sub foo { return { first => [ 1,2,3 ], second => [ 4,5,6 ] } } print foo()->{first}[2];
See, that's the sort of thing that bugs me. I understand why you have to dereference the return value, but it's annoying once you've gotten used to not having to. I'll admit that I didn't give the best example, though. I was trying to remember one of the Perl data structures that I had to look up each and every time I used it, and it looks like I picked the wrong one.
@array = @{ $hash{$key} };
That's more along the lines of the kind of stuff I had in mind. Again, I don't hate Perl. It's just that I've found other languages that map much more closely to how I think of algorithms. I always got too stuck in the syntax of Perl to really express myself.
Imagine not being able to connect to an FTP server running on Windows, only because you're using Mozilla or the FreeBSD ftp client, and such non-Microsoft products are deemed "insecure".
That's definitely worse than today's most common reason for not being able to connect to an FTP server running on Windows: it doesn't have enough free sockets left over from the warez and pr0n kiddies to open a TCP connection.
Line noise reputation aside, Perl can be _very_ readable.
I tell you what turned me away from Perl, though: the syntax for complex data structures was never clean enough that I could remember it from session to session. The python for getting a hash of arrays (in Perlspeak):
a = {'first': [1,2,3], 'second': [4,5,6]} print a['first'][2]
Perl made it possible, but Python made it easy. What about returning complex objects from functions? [1]
Again, you can do that in Perl, but it was never anywhere near that easy. I've read and written lots of Perl in my day that was well-organized and well-written, but there are certain basic operations that are just inherently ugly. Those warts simple can't be made readable because of the relatively bizarre syntax required to create them.
[1] Yes, I realize that Slashdot's broken "ecode" tags don't exactly help my case regarding Python's readability.
Hint: that's not a great way to gain credibility at the outset. Go ask your average young doctor how much they're paying in student loans, and see if the doctors' parking lot is really that different than the teachers' lot at the local high school.
Please excuse my while I shuffle out to my used Oldsmobile to go pick up my kid from (public) school. Overpaid? That'd be a nice change.
It wasn't that long ago that there was a big stink over a doctor-run web site that blacklisted malpractice plaintiffs so that doctors could deny them future coverage, regardless of who won the case.
...and as I remember, Slashdotters were widely against them being allowed to do that, too. The lesson I'd take from that data is that no matter which side of the debate the doctors are on, a lot of people will automatically assume they're at fault. The assumption being, of course, that all doctors are rich and therefore always in the wrong.
I'll remember that the next time I deal with a client that's a doctor or medical practice. I'll just be late and tell them I was troubleshooting problems that turned out to be more complex than originally described, and surely they understand since it happens all the time to them?
If the problem you were troubleshooting could potentially kill or cripple someone if you didn't fix it immediately, I imagine they'd understand. Getting fr1st p0s7 on Slashdot or overclocking a graphics card probably wouldn't get too much sympathy, though.
I'm always late to a doctor's "appointment". Why should I have to wait for them after I already made an appointment?
Congratulations, asshole: now everyone after you has to wait even longer.
My wife schedules each patient a reasonable amount of time for the problem that they're calling to see her about. On occasion, some of those problems turn out to be more complex and urgent than they expressed over the phone. Given that her alternatives are:
Spend the extra required time to treat that patient, or
Tell them, "sorry, your allotted time is up. Please see the receptionist about scheduling more next month."
Which would you pick, and why? Would you give the same answer if your problem was the one that's taking longer than expected?
Believe it or not, the vast majority of doctors would really like to stay on schedule. Given that the nature of their job is troubleshooting systems owned by users who aren't experts at explaining their problems (which anyone reading Slashdot should understand), that just isn't always possible.
In short, don't be a dick and make matters worse. A lot of the doctors I hang around with have an "n strikes" rule: screw them over n times without a legitimate excuse, and suddenly expect to find that all your appointments are at 6:30am or 7:00pm, whichever is least convenient for you. Is that really a battle you want to fight, particularly since if you weren't already sick you wouldn't be seeing them in the first place?
Another thing I thought sounded cool was the ability to cat/home/foo/music/some.mp3/raw >/dev/dsp and the mp3 would just play by using a plugin that ran it through an mp3 library.
This is exceptionally cool, and it's currently well supported by KDE's kioslaves. I much prefer that to a kernelspace solution because it means I get the same cool toys on FreeBSD as on Linux as on Solaris (and soon as on Windows).
Tech support hates them already
on
Office 12 Exposed
·
· Score: 3, Insightful
Type or ink your name below or click Select Image to select a picture to use as your signature:
I guaran-frickin'-tee our IT guy will get at least one call from a peeved user that can't 1) get Windows to recognize their inked signature or 2) get Sharpie off their LCD monitor.
I hereby propose "Strauser's Rule of UI Design":
Remember that stupid people will read your words, too. Consider the worst possible misinterpretation of anything you write, because it will always come back to you.
Slashdot Mirror
You mean the one they initially didn't want to fix unless you could somehow prove that it might affect you? I'd be hard pressed to think of a worse example of this magical Intel warranty support you're so proud of.
They may indeed be better now; if so, then you should have picked a case that didn't make people hate them.
...as long as you're paying cash and aren't being treated for any illness requiring health department notification. If you've got the measles, though, or paying with any insurance remotely connected to the government, then be prepared to cough up those digits (even if you're not currently choking on ladyfingers).
Why would you?
Not even close.
The ideal hash function has an equal probability of all possible results for any given input. Let's look at a non-ideal one: h(x) = 1. Now, any input results in the same hash value. Is that email signature valid? Sure: it has the expected hash! Did that ISO come through intact? Sure: the hash matches! Wouldn't be very useful, would it? In this simplified example, returning a 1024-bit hash value would be exactly as useful as returning a 1-bit hash value.
This is why MD5 is deprecated. It's not that it returns too few bits (which it does by modern standards), but that it's been shown to not return an ideal distribution of results. SHA is closer to the ideal as far as we know today, and SHA-256 adds extra bits in addition to the improved algorithm.
No, as much as I wanted to have a ubergeek keyboard, my hands just weren't up to the task (and years of muscle memory means that I'm not willing to re-learn new type M-compatible chords). I think I'll be picking up a second Happy Hacking keyboard instead.
This isn't really an issue for us since, as I mentioned, my wife runs nearly on time almost all the time.
I'm not sure I really see the difference. Think of "httpd.conf" as "config.pl" that gets executed by starting Apache and you're there. Either way you're dynamically generating the configuration; the difference is whether you write it out to a file first or directly altar Apache's state.
I worked for a small ISP, and we used a heavily-commented named.conf and associated zone files to keep track of configuration information, explanations for non-obvious things, etc. Since we were a small shop and worked well together, this was fine. Until we merged with another ISP. Whose admins "helpfully" slaved their BIND to ours, made it the master, and then slaved ours to theirs. Without changing the zone filenames in named.conf. I think that, had I been in the same room with them before we managed to retrieve our off-site backups, I might've had to have killed them.
One hugely useful system is to use version control on your configuration directories. Let the new guy delete at will; you can always roll back his commits and explain why you're going to beat him if he does it again.
Don't do that unless you have to. There are better ways to get the same result.
Not that it fully answers your needs, but surely someone who manages 1000s of sites would be aware of its existence?
But you're speaking as though Rutan is finished, which doesn't seem to be the case. And even if he were to quit today, several other private groups are pursuing the same goals. There was a lot more buzz about SpaceShipOne than the recent "Hey, it didn't explode!" shuttle flight, and I imagine the budget folks at NASA wish some of that excitement was aimed in their direction.
For most of the last five decades, NASA has been the sole North American spacefaring entity. That changed last year, and I don't see any reason to believe that Burt or John Carmack or someone else can't have as much success in the next five.
What saddens me most is that I don't really have much faith in them anymore. When I was growing up in the 70's, the folks at NASA were my heroes. They were the smartest, most determined, and best people anywhere in the world. I kind of wish I had that back, but at least private industry has given us a few new heroes to live vicariously through.
I posted my oh-so-clever comment about two seconds before reading the exact same thing further down. My bad.
You know it's named after a catalytic process. I know it's named after a catalytic process. The rest of the world knows the guy has "kat" right there in the name of the company.
Choosing "Alphakatalytisch" might've saved the good doctor a fair amount of grief.
Next will come "Grammar" (Gnome) and "grammatiK" (guess).
Within the following year we'll see "iWordflow", and six months past that (after a flaming rant and a two-week hackathon) everyone will upgrade to "OpenGrammar".
You'll eventually see XFGrammar for people who think Grammar and Grammatik are too bloated.
Why not describe it in a text file, but formatted as an SVG or a graphviz input file? They're easy enough to read that anyone with Notepad or better can understand and update them, but trivially compile to pretty pictures that you can give to management or hang on the wall.
See, that's the sort of thing that bugs me. I understand why you have to dereference the return value, but it's annoying once you've gotten used to not having to. I'll admit that I didn't give the best example, though. I was trying to remember one of the Perl data structures that I had to look up each and every time I used it, and it looks like I picked the wrong one.
That's more along the lines of the kind of stuff I had in mind. Again, I don't hate Perl. It's just that I've found other languages that map much more closely to how I think of algorithms. I always got too stuck in the syntax of Perl to really express myself.
That's definitely worse than today's most common reason for not being able to connect to an FTP server running on Windows: it doesn't have enough free sockets left over from the warez and pr0n kiddies to open a TCP connection.
I tell you what turned me away from Perl, though: the syntax for complex data structures was never clean enough that I could remember it from session to session. The python for getting a hash of arrays (in Perlspeak):
Perl made it possible, but Python made it easy. What about returning complex objects from functions? [1]
Again, you can do that in Perl, but it was never anywhere near that easy. I've read and written lots of Perl in my day that was well-organized and well-written, but there are certain basic operations that are just inherently ugly. Those warts simple can't be made readable because of the relatively bizarre syntax required to create them.
[1] Yes, I realize that Slashdot's broken "ecode" tags don't exactly help my case regarding Python's readability.
Hint: that's not a great way to gain credibility at the outset. Go ask your average young doctor how much they're paying in student loans, and see if the doctors' parking lot is really that different than the teachers' lot at the local high school.
Please excuse my while I shuffle out to my used Oldsmobile to go pick up my kid from (public) school. Overpaid? That'd be a nice change.
...and as I remember, Slashdotters were widely against them being allowed to do that, too. The lesson I'd take from that data is that no matter which side of the debate the doctors are on, a lot of people will automatically assume they're at fault. The assumption being, of course, that all doctors are rich and therefore always in the wrong.
If the problem you were troubleshooting could potentially kill or cripple someone if you didn't fix it immediately, I imagine they'd understand. Getting fr1st p0s7 on Slashdot or overclocking a graphics card probably wouldn't get too much sympathy, though.
Congratulations, asshole: now everyone after you has to wait even longer.
My wife schedules each patient a reasonable amount of time for the problem that they're calling to see her about. On occasion, some of those problems turn out to be more complex and urgent than they expressed over the phone. Given that her alternatives are:
- Spend the extra required time to treat that patient, or
- Tell them, "sorry, your allotted time is up. Please see the receptionist about scheduling more next month."
Which would you pick, and why? Would you give the same answer if your problem was the one that's taking longer than expected?Believe it or not, the vast majority of doctors would really like to stay on schedule. Given that the nature of their job is troubleshooting systems owned by users who aren't experts at explaining their problems (which anyone reading Slashdot should understand), that just isn't always possible.
In short, don't be a dick and make matters worse. A lot of the doctors I hang around with have an "n strikes" rule: screw them over n times without a legitimate excuse, and suddenly expect to find that all your appointments are at 6:30am or 7:00pm, whichever is least convenient for you. Is that really a battle you want to fight, particularly since if you weren't already sick you wouldn't be seeing them in the first place?
This is exceptionally cool, and it's currently well supported by KDE's kioslaves. I much prefer that to a kernelspace solution because it means I get the same cool toys on FreeBSD as on Linux as on Solaris (and soon as on Windows).
I guaran-frickin'-tee our IT guy will get at least one call from a peeved user that can't 1) get Windows to recognize their inked signature or 2) get Sharpie off their LCD monitor.
I hereby propose "Strauser's Rule of UI Design":