Sometimes I wonder why smartphone owners don't buy a $50 phone to complement their smartphone. The smartphone has all the snazzy features and the $50 phone is simple enough to not suffer from showstopping bugs like non-working telephone capabilities.
Unfortunately, though, that means a raytracing video game console needs to be three times s powerful as the PS3 in order to render relatively simple scenes. For full-scale gaming... Ten times? Fifteen? Seems unlikely to happen soon unless the console has specialized realtime raytracing hardware that works more efficiently than the PS3 SPU cluster.
I'm not an American student, so I ask out of interest: What's the meaning of those three-letter codes? I often see "101" to denote an introductory course and sometimes numbers like "202" for advanced courses - and sometimes entriely different numbers for everything. What's the system behind them?
In the interest of mutual learning I'll share how we do things at my university (University of Bremen, Germany): The courses I'm aware of have all kinds of names (eg. "Operating Systems 1", "Introduction to Production Computer Science*" (abbreviated to "Production CS 1") or even "Between Science and Fiction. A Project Seminar: Computer Science in the Context of Movies" (abbreviated to "Between Science and Fiction")).
Additionally, every course has a module number identifying faculty, degree course, when you an enroll, module domain** and lastly a unique identifier. "Between Science and Fiction" was course 03-05-H-803.5g (faculty 3 (math/CS), degree course 5 (CS diploma/B.Sc./M.Sc.), main study period ("Hauptstudium"), application domain (8), identifier 03.5g; the 03 is shared with courses like an ethics course and data protection).
The module number is only relevant for enrollment and the assessment.
* In Germany we use the term "Informatik", which is roughly equivalent to "computer science", but carries slightly different connotations. Don't confuse it with "Informationswissenschaften" (information science), though.
** All CS courses give you a certain amount of ECTS points in one of the following domains: Theory, practice and application. Every student has to accumulate a certain number of points in each of these domains, additionally some in the free domain (can be filled with courses from any degree course) and some you get automatically through a mandtory two-year project.
It's not as bad as I thought; apparently I just happened to run into a screenshot from someone's rethemed Windows. Of course, IE8 still looks different from everything else, but not quite as atrociously as I originally thought.
I think it's weird how the big OS vendors can't make consistent interfaces if their lives depended on it. Apple usually test-drives new UI concepts in iTunes before putting them into the next OS X release. Windows has been a mess of application-specific look-and-feels for quite some while now. Now even IE looks completely unlike everything else.
Am I the only person who's a bit bugged by that?
(By the way, iTunes is not that bad - it still looks mostly like the other apps save for the occasional new widget and the ugly scroll bars.)
Yes, IE8 is a marked improvement. But... I don't know what the proper antonym for "goodwill" is, but the IE devs have accumulated a lot of it over the years. Their and their product's reputation is so bad that in order to be seen as anthing but clowns they need to release not just a better product but the best product. They need to decidedly out-perform at least one of Gecko, Presto and Webkit in terms of standard compliance if they want any actual respect.
Microsoft didn't just drop the ball, they tossed in into a ditch and let it rot there for a few years. When they finally remembered they had a ball in the first place the competition was already miles ahead of them. If they want to actually play the game and be taken seriously again they don't just need to move in the same direction as everyone else; they need to do that and be faster than everyone or they will always play catch-up.
Yes, I know that it's not entirely fair to expect the IE devs to do better than everyone else faster than everyone else. But that's what they need to do if they want recognition.
If all "open source supporters" had your attitude, free software wouldnt have survived the 90s.
If none had, the Linux mindshare would be a good bit smaller. Absolutely refusing to deal with a certain situation because you don't have ideal conditions is hardly going to help, regardles of what th situation is.
"Buy a compatible card and vote with your $CURRENCY_UNITs" is advice I can wholeheartedly agree with; "go back to Windows because you can't have a 100.0% Free system" is not. Some of us like (for example) having accelerated graphics with post-R200 capabilities under Linux and they wouldn't use it if they couldn't. And we recommend and use Linux even though the drivers don't have the RMS Seal of Approval.
Only pragmatists would run Free Software into the ground, but no pragmatists would mean that it remains confined to a niche forever. It's the right mix that makes everything work. Like always.
I'm envious of people who can actually use that feature. On mobile phones I rarely manage to do more than 1-3 WPM, whether the phone tries to be helpful or not. Usually more with T9 turned off.
Probably depends on region. I'm in Germany and I've never knowingly met anyone who actually uses AIM. I assumed that the service was essentially dead, ICQ being much more popular.
Overall, from my experience most people appear to be using ICQ with some using MSN. Jabber (including GTalk) is used only by the tech-savvy, AIM and YIM have zero apparent market share.
It's probably also not blessed by the admins and can get you into trouble if you use it on a company PC (= grossly violate the comapn's IT security guidelines).
I don't know how I came up with five sums, especially since I gave the correct number in an earlier post. It's four checksums; three block-level checksums and one for the whole number.
That's true, of course. Especially since most German systems were designed with the lessons learned from other countries in mind. However, it's still somewhat amazing that your gov never got around to changing the SSN, seeing as it's the One Almighty Number.
Then again, there's probably not a lobby attached to the issue, so most congressmen probably are unaware that a simple incrementing number could have any disadvantages.
It would stop clerical errors, though. If it wasn't possible to accidentally declare the wrong person dead without MUCH work (proportionally growing with the number of checksums - the German ID number uses no less than five checksums, for example) less people would end up complaining about losing their identity and for the rare cases that do happen, lots of red tape would make sense. It would be harder to obtain someone else's identity by complaining about the number.
Also, reducing reliance on a central ID number might be a good idea, also. In Germany giving away your ID number is relatively unproblematic because it's rarely used except as a particularly unsafe way of verifying your age online. Everywhere you really need the number (ie. while dealing with certain governmental agencies), they will require your ID card, which employs a number of safety measures like holographic reproduction of the photo to keep people from faking them.
Most businesses aren't even interested in the ID number much. Generally, the ID card is the document everything revolves around - and it's much easier to make a fake-proof ID card than it is to make a fake-proof integer.
The problem lies not with "alive" but rather with "you". How do you convince them that you are indeed the person declared dead and not, in fact, someone else who wants to take over the identity of the deceased?
Of course, the sensible approach would be to check the records upon complaint and verify that everything was indeed entered correctly. But since we're talking beaurocracy here they'll only do that if the complaint comes from the "deceased" themself because they can't go around correcting mistakes, actual or not, without proper identification.
The solution is not more digits, but to make social security numbers, nay ALL identifying numbers, self checksumming.
SSNs aren't even checksummed? Holy shit, that's pretty primitive for a number that can ruin a person's life if entered incorrectly.
The German ID card, for example, has a 26-character alphanumeric string that features no less than four checksums:
The first nine digits contain information about your main domicile and a serial number. The tenth digit is the checksum for them. The block ends with a single character identifying your citizenship (AFAIK it's always "D").
The next seven digits are your date of birth in the format YYMMDD and a checksum for the DOB.
The next seven digits are the expiration date for the ID card in the same format and a checksum for them.
The last digit is a checksum for all preceding digits.
That way a simple error is likely to be noticed and the software could even tell you which part was entered incorrectly.
There's a nice invention called a pocket. My backpack has one; I tend to store stuff like adapters and the remote in there so I don't even have to remember bringing them. Bad preparation is not a neccessary requirement for being an Apple user; a basic sense of style and a smug feeling of superiority are the only ones.
If you mean the MacBook Air that's so simple you can't even use Ethernet and watch a DVD at the same time without using a USB hub - yes, that doesn't really work. But then again the MBA is an exec toy and the target demographic has no qualms paying top money for a device that's not really useful.
As for VGA ports: Having them can make things worse - like I said, the 15" MBP was designed to be relatively small and thus doesn't have much room for connectors. Every bit of space that desn't already house some port has either the hard drive or the optical drive behind it or is part of the hinge. Additional ports would've been nice, but Apple preferred a small frame. Whether or not that makes sense to you is your problem - in an industrial environment where you need a serial port an Apple is out of the question while a student might prefer having a less full backpack over being able to easily interface with JTAG ports.
By the way, I chant the holy It Just Works Mantra only if things really just work. WLAN Just Works on my MBP, while attaching an external HDD that insists on using two USB ports doesn't - the USB ports are on opposing sides of the notebook and most USB Y cables aren't long enough to reach around it. Apple's not magically better, but hardware-wise they usually make pretty solid decisions, their use of Sony batteries nonwithstanding.
Actually, there already are dedicated raytracing accelerators (like SaarCOR) and NVidia has presented simple realtime raytracing on the G80 at last year's CeBIT. It's not just Intel who is interested in this.
Eclipse is good, even if somewhat quirky at times.
As for Java being designed to work in a pure Java environment: I know, but even though something is not supposed to be the primary design goal that doesn't mean you can't implement it. Like copying a file or a way to execute an external application and read its output that can be done in less than twenty lines.
So a simple mechanical adaptor that goes between the cable and the notebook is unacceptable? For how long are notebooks supposed to support old standards? The VGA connector is becoming more and more rare with most people switching to flatscreen monitors and it takes up real estate that can be used to house additional ports (in fact, the MBP would have to either sacrifice the ExpressCard slot or change the entire hinge and become at least one centimeter taller in order to be able to house that port). I think having an adaptor that doesn't even consist of anything besides two conectors and wiring is quite reasonable, especially when you get it for free with every notebook.
One could also lament that many notebooks don't directly support PS/2 mice. There is, however, a simple adaptor that is very robust and easy to use and doesn't require any special software. Everyone wins - the notebook manufacturers don't have to use real estate to support an old interface and the users can still use PS/2 mice if they want to. Unlike DVI-to-VGA adaptor, this one usually comes bundled with newer PS/2 mice because the mouse manufacturers have accepted that the PS/2 interface standard is mostly obsolete.
P.S. Gotta love those network topology diagrams. Pretty dang nice for a nonprofit.
Of course. They just made a node called "Wikipedia Design Principles" and when they were sure that the smartest minds of Wikipedia had contributed, they used it to build their network. That's why they have such a robust computer network and Monster Deluxe power strips with gilded contacts for 99.5% lower impedance in the crucial 220 Hz level.
I wished that Java had better low-level functions. It's hard to take a programming language seriously when the official way to copy a file in said language involves two File objects, a BufferedReader and a BufferedWriter. I think it's reasonably safe to assume that most OSes have a standard API for copying files that Java could call - and if that isn't an option Sun could transparently implement everything in Java.
Java is nice, but only if you work in a pure Java environment and can set up everything so yu don't need to talk to the OS much. Once you run into things like PID management or basic filesystem interactions you need to fall back to horribly obtuse solutions like the one above. In a heterogenous environment (where Java needs to talk to applications written in other languages and rewriting the other apps to be used with JNI is unfeasible) Java can be quite a pain.
Slashdot Mirror
Sometimes I wonder why smartphone owners don't buy a $50 phone to complement their smartphone. The smartphone has all the snazzy features and the $50 phone is simple enough to not suffer from showstopping bugs like non-working telephone capabilities.
Unfortunately, though, that means a raytracing video game console needs to be three times s powerful as the PS3 in order to render relatively simple scenes. For full-scale gaming... Ten times? Fifteen? Seems unlikely to happen soon unless the console has specialized realtime raytracing hardware that works more efficiently than the PS3 SPU cluster.
Check out Odin Sphere (link goes to links to the trailer). That game is just beautiful.
I'm not an American student, so I ask out of interest: What's the meaning of those three-letter codes? I often see "101" to denote an introductory course and sometimes numbers like "202" for advanced courses - and sometimes entriely different numbers for everything. What's the system behind them?
In the interest of mutual learning I'll share how we do things at my university (University of Bremen, Germany): The courses I'm aware of have all kinds of names (eg. "Operating Systems 1", "Introduction to Production Computer Science*" (abbreviated to "Production CS 1") or even "Between Science and Fiction. A Project Seminar: Computer Science in the Context of Movies" (abbreviated to "Between Science and Fiction")).
Additionally, every course has a module number identifying faculty, degree course, when you an enroll, module domain** and lastly a unique identifier. "Between Science and Fiction" was course 03-05-H-803.5g (faculty 3 (math/CS), degree course 5 (CS diploma/B.Sc./M.Sc.), main study period ("Hauptstudium"), application domain (8), identifier 03.5g; the 03 is shared with courses like an ethics course and data protection).
The module number is only relevant for enrollment and the assessment.
* In Germany we use the term "Informatik", which is roughly equivalent to "computer science", but carries slightly different connotations. Don't confuse it with "Informationswissenschaften" (information science), though.
** All CS courses give you a certain amount of ECTS points in one of the following domains: Theory, practice and application. Every student has to accumulate a certain number of points in each of these domains, additionally some in the free domain (can be filled with courses from any degree course) and some you get automatically through a mandtory two-year project.
It's not as bad as I thought; apparently I just happened to run into a screenshot from someone's rethemed Windows. Of course, IE8 still looks different from everything else, but not quite as atrociously as I originally thought.
I think it's weird how the big OS vendors can't make consistent interfaces if their lives depended on it. Apple usually test-drives new UI concepts in iTunes before putting them into the next OS X release. Windows has been a mess of application-specific look-and-feels for quite some while now. Now even IE looks completely unlike everything else.
Am I the only person who's a bit bugged by that?
(By the way, iTunes is not that bad - it still looks mostly like the other apps save for the occasional new widget and the ugly scroll bars.)
Yes, IE8 is a marked improvement. But... I don't know what the proper antonym for "goodwill" is, but the IE devs have accumulated a lot of it over the years. Their and their product's reputation is so bad that in order to be seen as anthing but clowns they need to release not just a better product but the best product. They need to decidedly out-perform at least one of Gecko, Presto and Webkit in terms of standard compliance if they want any actual respect.
Microsoft didn't just drop the ball, they tossed in into a ditch and let it rot there for a few years. When they finally remembered they had a ball in the first place the competition was already miles ahead of them. If they want to actually play the game and be taken seriously again they don't just need to move in the same direction as everyone else; they need to do that and be faster than everyone or they will always play catch-up.
Yes, I know that it's not entirely fair to expect the IE devs to do better than everyone else faster than everyone else. But that's what they need to do if they want recognition.
You mean Microsoft finally implemented the elusive file:// protocol? No way!
"Buy a compatible card and vote with your $CURRENCY_UNITs" is advice I can wholeheartedly agree with; "go back to Windows because you can't have a 100.0% Free system" is not. Some of us like (for example) having accelerated graphics with post-R200 capabilities under Linux and they wouldn't use it if they couldn't. And we recommend and use Linux even though the drivers don't have the RMS Seal of Approval.
Only pragmatists would run Free Software into the ground, but no pragmatists would mean that it remains confined to a niche forever. It's the right mix that makes everything work. Like always.
I'm envious of people who can actually use that feature. On mobile phones I rarely manage to do more than 1-3 WPM, whether the phone tries to be helpful or not. Usually more with T9 turned off.
Probably depends on region. I'm in Germany and I've never knowingly met anyone who actually uses AIM. I assumed that the service was essentially dead, ICQ being much more popular.
Overall, from my experience most people appear to be using ICQ with some using MSN. Jabber (including GTalk) is used only by the tech-savvy, AIM and YIM have zero apparent market share.
It's probably also not blessed by the admins and can get you into trouble if you use it on a company PC (= grossly violate the comapn's IT security guidelines).
I don't know how I came up with five sums, especially since I gave the correct number in an earlier post. It's four checksums; three block-level checksums and one for the whole number.
That's true, of course. Especially since most German systems were designed with the lessons learned from other countries in mind. However, it's still somewhat amazing that your gov never got around to changing the SSN, seeing as it's the One Almighty Number.
Then again, there's probably not a lobby attached to the issue, so most congressmen probably are unaware that a simple incrementing number could have any disadvantages.
It would stop clerical errors, though. If it wasn't possible to accidentally declare the wrong person dead without MUCH work (proportionally growing with the number of checksums - the German ID number uses no less than five checksums, for example) less people would end up complaining about losing their identity and for the rare cases that do happen, lots of red tape would make sense. It would be harder to obtain someone else's identity by complaining about the number.
Also, reducing reliance on a central ID number might be a good idea, also. In Germany giving away your ID number is relatively unproblematic because it's rarely used except as a particularly unsafe way of verifying your age online. Everywhere you really need the number (ie. while dealing with certain governmental agencies), they will require your ID card, which employs a number of safety measures like holographic reproduction of the photo to keep people from faking them.
Most businesses aren't even interested in the ID number much. Generally, the ID card is the document everything revolves around - and it's much easier to make a fake-proof ID card than it is to make a fake-proof integer.
The problem lies not with "alive" but rather with "you". How do you convince them that you are indeed the person declared dead and not, in fact, someone else who wants to take over the identity of the deceased?
Of course, the sensible approach would be to check the records upon complaint and verify that everything was indeed entered correctly. But since we're talking beaurocracy here they'll only do that if the complaint comes from the "deceased" themself because they can't go around correcting mistakes, actual or not, without proper identification.
The German ID card, for example, has a 26-character alphanumeric string that features no less than four checksums:
The first nine digits contain information about your main domicile and a serial number. The tenth digit is the checksum for them. The block ends with a single character identifying your citizenship (AFAIK it's always "D").
The next seven digits are your date of birth in the format YYMMDD and a checksum for the DOB.
The next seven digits are the expiration date for the ID card in the same format and a checksum for them.
The last digit is a checksum for all preceding digits.
That way a simple error is likely to be noticed and the software could even tell you which part was entered incorrectly.
There's a nice invention called a pocket. My backpack has one; I tend to store stuff like adapters and the remote in there so I don't even have to remember bringing them. Bad preparation is not a neccessary requirement for being an Apple user; a basic sense of style and a smug feeling of superiority are the only ones.
If you mean the MacBook Air that's so simple you can't even use Ethernet and watch a DVD at the same time without using a USB hub - yes, that doesn't really work. But then again the MBA is an exec toy and the target demographic has no qualms paying top money for a device that's not really useful.
As for VGA ports: Having them can make things worse - like I said, the 15" MBP was designed to be relatively small and thus doesn't have much room for connectors. Every bit of space that desn't already house some port has either the hard drive or the optical drive behind it or is part of the hinge. Additional ports would've been nice, but Apple preferred a small frame. Whether or not that makes sense to you is your problem - in an industrial environment where you need a serial port an Apple is out of the question while a student might prefer having a less full backpack over being able to easily interface with JTAG ports.
By the way, I chant the holy It Just Works Mantra only if things really just work. WLAN Just Works on my MBP, while attaching an external HDD that insists on using two USB ports doesn't - the USB ports are on opposing sides of the notebook and most USB Y cables aren't long enough to reach around it. Apple's not magically better, but hardware-wise they usually make pretty solid decisions, their use of Sony batteries nonwithstanding.
Actually, there already are dedicated raytracing accelerators (like SaarCOR) and NVidia has presented simple realtime raytracing on the G80 at last year's CeBIT. It's not just Intel who is interested in this.
Eclipse is good, even if somewhat quirky at times.
As for Java being designed to work in a pure Java environment: I know, but even though something is not supposed to be the primary design goal that doesn't mean you can't implement it. Like copying a file or a way to execute an external application and read its output that can be done in less than twenty lines.
So a simple mechanical adaptor that goes between the cable and the notebook is unacceptable? For how long are notebooks supposed to support old standards? The VGA connector is becoming more and more rare with most people switching to flatscreen monitors and it takes up real estate that can be used to house additional ports (in fact, the MBP would have to either sacrifice the ExpressCard slot or change the entire hinge and become at least one centimeter taller in order to be able to house that port). I think having an adaptor that doesn't even consist of anything besides two conectors and wiring is quite reasonable, especially when you get it for free with every notebook.
One could also lament that many notebooks don't directly support PS/2 mice. There is, however, a simple adaptor that is very robust and easy to use and doesn't require any special software. Everyone wins - the notebook manufacturers don't have to use real estate to support an old interface and the users can still use PS/2 mice if they want to. Unlike DVI-to-VGA adaptor, this one usually comes bundled with newer PS/2 mice because the mouse manufacturers have accepted that the PS/2 interface standard is mostly obsolete.
Or you just do like Apple does and ship a DVI-to-VGA adapter.
I wished that Java had better low-level functions. It's hard to take a programming language seriously when the official way to copy a file in said language involves two File objects, a BufferedReader and a BufferedWriter. I think it's reasonably safe to assume that most OSes have a standard API for copying files that Java could call - and if that isn't an option Sun could transparently implement everything in Java.
Java is nice, but only if you work in a pure Java environment and can set up everything so yu don't need to talk to the OS much. Once you run into things like PID management or basic filesystem interactions you need to fall back to horribly obtuse solutions like the one above. In a heterogenous environment (where Java needs to talk to applications written in other languages and rewriting the other apps to be used with JNI is unfeasible) Java can be quite a pain.