It is no longer possible to purchase anything close to the minimum processor required to "get into modern computing". Processors go off the market far before they're obsolete. And the only real thing driving processor upgrades is Windows bloat and video games. Windows bloat can be easily avoided using a real operating system. Hopefully someday the console manufacturers will get their act together and I can stop trying to play games on my workstation.
Transmeta and Apple have the right approach: increasing CPU power is no longer innovation. Consumer chip companies should spend more time looking at the quality of clockcycles (ie: what you can do with them and the user experience that results). Of course there will always be a modest market for monster server processors...
Yeah! It's damn well time that grad and post-grad students take their place amongst the heroic archtypes! I'm tired of watching movies where I can't relate because they take place outdoors. And what's with heros who have these things called "lives", come on, lets have some realism. My friends may feign boredom, but they're actually jealous of my dangerous and exciting research, right? Well okay, so it's not dangerous...
Testing, whether field, regression, or otherwise can only prove the presence of bugs and security holes. The only kind of testing that can prove the absence of bugs is absolute coverage, which is computationally infeasable in even the most trivial programs. If you actually want to prove things, you need to use formal methods, there's really no other way.
Never mind "n years without a security hole", imagine an OS that could say "security holes can only occur under violation of the laws of logic"!
Please don't mix presentation style with semantic markup. Your sentence is not the content of an accent, but should simply be presented on aural browsers using a particular accent. A simplistic example of the correct way to mark up your data would be:
<span style="accent: scottish">I'll have you know it's the eigthth most widely circulated paper.</span>
However following the standards set by the working drafts for XHTML2 you should instead give the span entity a unique id and set its style using an ID-selector in a page-wide stylesheet. You also may wish to provide alternative accents if the "scottish" one is not available on the user's system.
As a CompSci grad student, I don't fully trust something until I understand what it's doing under the hood or see at least semi-formal justification. If Python wants to be a serious teaching language, it either needs massive industry use (Java), simple implementation (Scheme & C), or very elegant semantics (Scheme & Haskell) -- right now it is nowhere close to having any of these.
As a grad student who may someday be teaching introductory CS courses and thinks the current C-or-Java approach is lacking, I think quite a lot about pedagogical languages. As much as I may enjoy programming in Python, I have serious doubts about its suitability in this area. Lets evaluate the languages you mention:
C/ASM: rough learning curve and not practical for upper-year courses with OO (neither C++ nor Objective-C seems suited for teaching) but the students can grasp what's really going on and learn about low-level goodies.
Java: not elegantly designed and too complex to even come close to mastering in the first year but prized in industry (or at least students think so).
Perl: Yeah, right.
Lisp/Scheme: elegant implementation and yet immensely useful for teaching theoretical concepts, however I think the extreme dynamic typing is starting to show its age.
Python: not simple, not theoretically interesting, not especially elegant, not used in industry. Python seems to fall just short of every criteria. Not really short, but too short.
I think Python needs two things before I'd reconsider it as a teaching language:
A bit more exposure in industry so students don't say "what the fuck is Python?" in the first lecture.
Idealised Python: A more elegant, minimalist, simplistic version to start the students out on and which could conceivably be implemented in a half-year course.
For some reason gaming companies got the idea that the only popular games are those that are maximally realistic. As a result, they consistently sacrifice gameplay for gee-whiz graphics. This leaves people like me, who aren't willing to a pay a graphics tax to play games, happily stuck with CounterStrike. When will the industry get a clue?!
Could someone fill me in on how this economics of the TV->DVD market work? I just can't understand why Fox would leave a show in their vault when an intern could have it burned to DVDs and ready to sell in a week or two.
Millennium was so loved that one of the major papers in Vancouver published fanfiction to finish off the series after it was cancelled. If Fox had have released the DVDs right away, they would have sold tonnes. Now I'd like to think it'd at least be worth their while -- Millennium is one of my favourite shows of all time and its short run means the full set should be affordable.
Why the hell don't the consoles have mouses and keyboards yet?! I was really impressed with the controls the first time I played Halo, until I realised that all they had done was manage to be almost as good as my PC. Now that some consoles use USB, there isn't even a technological reason to resist.
I'm tired of upgrading my desktop to play a video game once in a while. But the games I do like all require rich input devices. I want a console to play CounterStrike -- is that so much to ask?!
I don't know what this flashy thing wiggling in the water is, but it certainly looks tastier than work...
This is your only good point Apple are certainly masters of the PowerPC. However as ESR is fond to remind us, very little of the Linux experience is actually from the kernel: you could easily build a free desktop on Darwin.
I think applications perhaps benefit from open source even more than systems. Non-programmers can be much more active participants in the bug fixing process and the ease of evaluating competing applicatiosn should lead to faster evolution.
I agree that Linus' approach doesn't scale, however that's what keeps Linux modular and non-monolithic.
By now everyone who uses a computer should know better than to compare version numbers between programs. Unless, of course, you want to compare MacOS 2.x with Linux, Apache, etc.
My efficiency increases when moving from a non-customised Mac or Windows GUI to my Sawfish/Gnome desktop as much as moving from DOS to ksh. I use my computer enough that I'm willing to put in the time to learning new features. If someone else is using it, they can use a simpler (or their own!) environment.
Therefore, the programming becomes not, writing the code, but specifying what code has to do in some way...
Rather than "specifying", how about we use the word "declaring"?
You feed in the rules of the game in a special language...
Lets invent a term: "special" languages used for the purpose of "declaring" behaviour could be called..."declarative languages".
Still, in some cases, where the problem is easy to state, you should have a solution program in just a few minutes; whereas now it could take hours or days.
So these "declarative languages" will be high-level and good for rapid prototyping?
Oh yeah, and the idea probably works for parallel and quantum computers too.
The easiest way to make it parallisable would be to only allow expressions with no side effects, so passive function application would have to be the primary means of computation.
You know, this rings a bell; I'm sure I've heard of something like this before...?
I'm really happy with my Celeron 466s on a BP-6. but lately roll-your-own SMP has been taking a turn for the corporate.:( Where are the dual Durons roundups?! Quantity over quality is the only way to go.
I attended a presentation yesterday for one of my Master's colleagues. Her thesis topic is implementing a buffer optimising technique in PostgreSQL. She claims that although there is extensive theoretical backing for the algorithm, it has never been implemented.
Clearly this will be a major boon for PostgreSQL. Why did she choose that as her platform? Because she can't get access to the source of other DBMSs, of course! (Actually her research group has close enough ties to IBM that she probably could have got DB2, but I'd assume that she's also favouring the smaller source size.)
The best databases of today are commercial, but the best ones of tomorrow will be OS. Just as academia leapfrogged over industry to make Haskell, the really big ideas will appear in OS first.
Re:to the tune of "if your'e happy & you know
on
Strike on Iraq
·
· Score: 1
I'm sorry, but that module doesn't appear to have anything to do with security. It is about software failure, which is simply engineering failure science applied to software. While I do not dispute that such topics are an important part of a Software Engineer's curriculum, I believe that the course discussed in the story is fundamentally different.
If you don't believe me, ask yourself: in the numerous case studies performed in your "module", did any of them feature intentional failure? Did any buffer overflows occur to compromise a system or were they accidents as a result of poor design?
Um, why do you think North Korea started "squealing" in the first place? If a cowboy in charge of the largest military in the world put me on his list of targets (the "Axis of Evil") and then proceeded to invade the first country on that list, I would sure as hell start building some weapons of deterence. The surest way to avoid a forced "regime change" is to have nukes within distance of Seoul, Beijing, and Tokyo.
And it's not like North Korea broke any international laws (unlike, possibly, the US): the Nuclear Non-Proliferation Treaty includes "the right to withdraw from the Treaty if it decides that extraordinary events...have jeopardized the supreme interests of its country." (Article X.1)
Remember when Jobs demanded that jobsforpresident.org be shutdown? It wasn't because he didn't want to be President, it's because he didn't want to blow his load too early! He wants to be a darkhorse candidate, so he needs to come in late. Gore isn't really there to be on Apple's BoD, he's there to advise Steve!
But who will his running mate be? It seems like Woz has been out of the spotlight too long. Who else would fit?
4. Al Gore knows government. True, but he knew the last government. That's a problem. He knew a lot of key decisons makers and might have been able to leverage those. Now there is a new government, and many decisions of scale and size are made by political apointee's and their sub-ordinates. This means he in fact has very little leverage for Apple now that GWBush is in office.
Presidents come and go, but mandarins are in it for the long haul. And we all know who's actually making the decisions...
You make the assumption that there will be a point in the future where they will need to be able to solve such problems on their own. As computers become more obquitious this assumption must be questioned. Socrates argued that writing is dangerous, how is your stance different?
Incidentally, I taught myself crude numerical methods in highschool by abusing my calculator in multiple choice tests when I didn't know the correct method. I may know those techniques less, but I learned different, perhaps more valuable ones, in the meanwhile.
The Software Practices Lab at UBC has been doing all sorts of AOP research lately. Of particular interest:
A study to determine whether AOP is actually useful.
Implementation of all the GoF Design Patterns in both vanilla Java and AspectJ.
They're also working on AspectC and AspectSmalltalk. It's my understanding that UBC is one of the major world centers in AOP research -- do you think I should do a PhD there?
The problem is that employers, for whatever reason, want all sorts of information that they don't need. What's needed are some digital certificates to replace all these numbers.
Need to know if I can legally work in the country? Here's my certificate from HRDC. Need to give me a paycheck? Here's a unique deposit number from my bank. Need to do a credit check? Here's a certificate from my bank. etc.
I think the Segway is a lot less effort, especially when going up hills. In particular, ptorrone, who has done a lot of posting in this story, explains that he doesn't have a shower at work and frequently needs to meet with clients first thing in the morning.
Last summer I biked to work every day and got a T-shirt soaked in sweat for my trouble. As much as I appreciated the exercise, I would prefer using electric in the morning and work my ass off in the afternoon.
Slashdot Mirror
It is no longer possible to purchase anything close to the minimum processor required to "get into modern computing". Processors go off the market far before they're obsolete. And the only real thing driving processor upgrades is Windows bloat and video games. Windows bloat can be easily avoided using a real operating system. Hopefully someday the console manufacturers will get their act together and I can stop trying to play games on my workstation.
Transmeta and Apple have the right approach: increasing CPU power is no longer innovation. Consumer chip companies should spend more time looking at the quality of clockcycles (ie: what you can do with them and the user experience that results). Of course there will always be a modest market for monster server processors...
Yeah! It's damn well time that grad and post-grad students take their place amongst the heroic archtypes! I'm tired of watching movies where I can't relate because they take place outdoors. And what's with heros who have these things called "lives", come on, lets have some realism. My friends may feign boredom, but they're actually jealous of my dangerous and exciting research, right? Well okay, so it's not dangerous...
Remember, these men do research for you!
Testing, whether field, regression, or otherwise can only prove the presence of bugs and security holes. The only kind of testing that can prove the absence of bugs is absolute coverage, which is computationally infeasable in even the most trivial programs. If you actually want to prove things, you need to use formal methods, there's really no other way.
Never mind "n years without a security hole", imagine an OS that could say "security holes can only occur under violation of the laws of logic"!
like the fish?
You have a good point. Marking it up as a paragraph or sentence would be more semantically correct.
Please don't mix presentation style with semantic markup. Your sentence is not the content of an accent, but should simply be presented on aural browsers using a particular accent. A simplistic example of the correct way to mark up your data would be:
<span style="accent: scottish">I'll have you know it's the eigthth most widely circulated paper.</span>
However following the standards set by the working drafts for XHTML2 you should instead give the span entity a unique id and set its style using an ID-selector in a page-wide stylesheet. You also may wish to provide alternative accents if the "scottish" one is not available on the user's system.
As a CompSci grad student, I don't fully trust something until I understand what it's doing under the hood or see at least semi-formal justification. If Python wants to be a serious teaching language, it either needs massive industry use (Java), simple implementation (Scheme & C), or very elegant semantics (Scheme & Haskell) -- right now it is nowhere close to having any of these.
As a grad student who may someday be teaching introductory CS courses and thinks the current C-or-Java approach is lacking, I think quite a lot about pedagogical languages. As much as I may enjoy programming in Python, I have serious doubts about its suitability in this area. Lets evaluate the languages you mention:
I think Python needs two things before I'd reconsider it as a teaching language:
For some reason gaming companies got the idea that the only popular games are those that are maximally realistic. As a result, they consistently sacrifice gameplay for gee-whiz graphics. This leaves people like me, who aren't willing to a pay a graphics tax to play games, happily stuck with CounterStrike. When will the industry get a clue?!
Could someone fill me in on how this economics of the TV->DVD market work? I just can't understand why Fox would leave a show in their vault when an intern could have it burned to DVDs and ready to sell in a week or two.
Millennium was so loved that one of the major papers in Vancouver published fanfiction to finish off the series after it was cancelled. If Fox had have released the DVDs right away, they would have sold tonnes. Now I'd like to think it'd at least be worth their while -- Millennium is one of my favourite shows of all time and its short run means the full set should be affordable.
Please mod the parent up -- keep the dream alive!
Why the hell don't the consoles have mouses and keyboards yet?! I was really impressed with the controls the first time I played Halo, until I realised that all they had done was manage to be almost as good as my PC. Now that some consoles use USB, there isn't even a technological reason to resist.
I'm tired of upgrading my desktop to play a video game once in a while. But the games I do like all require rich input devices. I want a console to play CounterStrike -- is that so much to ask?!
I don't know what this flashy thing wiggling in the water is, but it certainly looks tastier than work...
E: Invalid operation fucked
Rather than "specifying", how about we use the word "declaring"?
Lets invent a term: "special" languages used for the purpose of "declaring" behaviour could be called..."declarative languages".
So these "declarative languages" will be high-level and good for rapid prototyping?
The easiest way to make it parallisable would be to only allow expressions with no side effects, so passive function application would have to be the primary means of computation.
You know, this rings a bell; I'm sure I've heard of something like this before...?
I'm really happy with my Celeron 466s on a BP-6. but lately roll-your-own SMP has been taking a turn for the corporate. :( Where are the dual Durons roundups?! Quantity over quality is the only way to go.
I attended a presentation yesterday for one of my Master's colleagues. Her thesis topic is implementing a buffer optimising technique in PostgreSQL. She claims that although there is extensive theoretical backing for the algorithm, it has never been implemented.
Clearly this will be a major boon for PostgreSQL. Why did she choose that as her platform? Because she can't get access to the source of other DBMSs, of course! (Actually her research group has close enough ties to IBM that she probably could have got DB2, but I'd assume that she's also favouring the smaller source size.)
The best databases of today are commercial, but the best ones of tomorrow will be OS. Just as academia leapfrogged over industry to make Haskell, the really big ideas will appear in OS first.
Ha ha ha ha ha ha ha!
I'm sorry, but that module doesn't appear to have anything to do with security. It is about software failure, which is simply engineering failure science applied to software. While I do not dispute that such topics are an important part of a Software Engineer's curriculum, I believe that the course discussed in the story is fundamentally different.
If you don't believe me, ask yourself: in the numerous case studies performed in your "module", did any of them feature intentional failure? Did any buffer overflows occur to compromise a system or were they accidents as a result of poor design?
Um, why do you think North Korea started "squealing" in the first place? If a cowboy in charge of the largest military in the world put me on his list of targets (the "Axis of Evil") and then proceeded to invade the first country on that list, I would sure as hell start building some weapons of deterence. The surest way to avoid a forced "regime change" is to have nukes within distance of Seoul, Beijing, and Tokyo.
And it's not like North Korea broke any international laws (unlike, possibly, the US): the Nuclear Non-Proliferation Treaty includes "the right to withdraw from the Treaty if it decides that extraordinary events...have jeopardized the supreme interests of its country." (Article X.1)
Remember when Jobs demanded that jobsforpresident.org be shutdown? It wasn't because he didn't want to be President, it's because he didn't want to blow his load too early! He wants to be a darkhorse candidate, so he needs to come in late. Gore isn't really there to be on Apple's BoD, he's there to advise Steve!
But who will his running mate be? It seems like Woz has been out of the spotlight too long. Who else would fit?
Presidents come and go, but mandarins are in it for the long haul. And we all know who's actually making the decisions...
What age did you teach her Logo? Any advice?
You make the assumption that there will be a point in the future where they will need to be able to solve such problems on their own. As computers become more obquitious this assumption must be questioned. Socrates argued that writing is dangerous, how is your stance different?
Incidentally, I taught myself crude numerical methods in highschool by abusing my calculator in multiple choice tests when I didn't know the correct method. I may know those techniques less, but I learned different, perhaps more valuable ones, in the meanwhile.
The Software Practices Lab at UBC has been doing all sorts of AOP research lately. Of particular interest:
They're also working on AspectC and AspectSmalltalk. It's my understanding that UBC is one of the major world centers in AOP research -- do you think I should do a PhD there?
The problem is that employers, for whatever reason, want all sorts of information that they don't need. What's needed are some digital certificates to replace all these numbers.
Need to know if I can legally work in the country? Here's my certificate from HRDC. Need to give me a paycheck? Here's a unique deposit number from my bank. Need to do a credit check? Here's a certificate from my bank. etc.
I think the Segway is a lot less effort, especially when going up hills. In particular, ptorrone, who has done a lot of posting in this story, explains that he doesn't have a shower at work and frequently needs to meet with clients first thing in the morning.
Last summer I biked to work every day and got a T-shirt soaked in sweat for my trouble. As much as I appreciated the exercise, I would prefer using electric in the morning and work my ass off in the afternoon.