You seem to be unaware of the great fiscal burden of transporting zombies over the Canada-US border post 9/11. It was much easier (and cheaper) for the first few films.
Of course, back then, the US had plenty of zombies with free time to make movies, and as a consequence did not have to import many. Now we have elected all of them to public office.
However, having played resident evil (it was necessary research for a biology class), I assure you that zombie dogs rarely have their damaged tissue repaired, which is a detail the article seems to overlook.
How about RTFS(ummary)? The summary clearly asks about other software products or even hardware.
Also, I think that the point was that Netscape had a "complete" product when launching. And I think that the argument of the article only applies if the software is self patching. This is automatically true for a web service, but it could be true for something like a browser (although maybe not over a modem).
I'd mod you up as interesting. Made me think about some others:
How old was Nintendo before they made the NES? Nobody cared about the company that made playing cards and crappy plastic puzzles, but then... blammo!
For that matter, how old was Micro-Soft (or however they used to write it) before Windows (ok, not 25, but still...)? I mean, a company brand based on Solitaire has to mean something.
And how old was the SCO brand before the current company started suing everyone and making everyone hate them?
Some brands exist for a while before being important/infamous.
Actually, the specific thing I wanted to do was create a live slideshow with the ability to dump new photos into the slideshow mix by attaching a camera to the computer. I could automatically download the photos from the camera and bring them into iPhoto for a slideshow. However, iPhoto has limited capabilities with choosing the transitions via script, and no control to force a given image as the next image in the slideshow. PowerPoint can do everything except import a photo (which makes it useless for live updates) -- although the image importing is a known bug. There is also the slideshow application (used by spotlight). It also lacks transition selection via script.
It is possible to do everything by scripting quicktime, but that wasn't really the point of my post. I just wish that Apple opened their applications to their own scripting interface in a way that was actually useful.
Seriously, though, I would really like Apple to make iWork into a complete product. It isn't just missing an excel and access replacement. It is also missing key "Apple" functionality: applescript capability.
While keynote 2 does expose an applescript dictionary, it is completely useless. Things you should be able to do (but can't) via script:
1. Create a new document (slideshow) 2. Add a slide 3. Edit the slide 4. Set the transition effect
OK, so basically anything useful. The sad part is that Microsoft PowerPoint has an almost useful applescript integration. I say "almost" because the bindings for creating image slides is broken (you get a nice interpreter error when you try to create an image from a file).
Heh. I just tried it myself. If it helps, if you can append.csv to the end of the filename, it opens calc, even if the file itself is tab delimited (it asks for the delimeter anyway if it is comma delimited).
The thing that (still) gets me is how long it takes to plot datasets with a few thousand points. gnuplot does it fine, but oocalc is terrible at it.
Re:Must all write-ups include a BS controversy?
on
AMD Quad Cores, Oh My
·
· Score: 1
It's particularly BS because most of the programs I run *are* multi-threaded. So they already have the ability to hyperthread. It's not even a chicken-and-the-egg problem as some have commented. It helps most applications right now. Ok, so grep isn't threaded, but runing ps | grep foo does run as two processes and is helped by a dual core (or hyperthreading, or multiple CPUs).
1. Reversible computing is Turing complete. In fact, this is a standard part of the intro chapters of quantum computing books (as all quantum gates are reversible -- unitary operators). You have to pad extra zeroes to your input to use as scratch paper for the reversible computation (so, your intuition about useless data is correct). However, there is a general technique for reusing the scratch padding (but nobody actually shows the steps in their papers anymore because it is so common). You can find it in a book on quantum computing. Unfortunately, I dropped the course, and don't remember the specifics (and though I can see the book, it is too far to reach).
2. As far as using the Turing model, there is one for reversible computing. Nobody I know uses it when discussing reversible computing. Most people use a circuit model (complexity of the "algorithm" is length of the circuit). Someone obviously did the comparison at some point, but turing machines are painful to work with (very verbose). In general, reversing a computation to recapture the scratch work requires double the "run time" -- you just CNOT the bits you want to save out of the answer part and reverse the whole computation (minus the bits you just saved). What remains is the input plus scratch plus the output. OK, I guess I do remember the specifics.
Does your university have a policy on challenging a course? I know all of the UC schools allow you to challenge any course that isn't the prerequisite for a course you have already taken.
As an undergrad, I only knew one student that made use of this rule. However, he was driven to graduate in three years with a double major in CS and physics (he succeeded).
I am a mac user. I own a dual G5, a powerbook, and a mac mini. I also own 2 linux machines and one (currently dead, recovery disk in hand) windows machine.
I wouldn't consider putting linux on my powerbook because I would lose the instant-on feature that makes the powerbook such a convenient laptop (I'd also lose MS Office, which is a requirement for work, and Open Office still doesn't properly open most word documents I try due to the presence of arrows and other non-alphabet characters).
I wouldn't put linux on my dual G5 because I couldn't use a number of my commercial applications (Lightwave, Flash, iTunes), as well as devices (don't know if my printer, scanner, camera, iPod, airport, etc. have linux drivers).
I might consider linux on the mini, but my girlfriend uses that computer and I have no real reason to switch... so I probably won't.
As far as considering linux: I probably wouldn't replace my linux boxes with macs. I use them as headless servers, and they compile and run the programs I want without hassle. Minor corruptions are easy to fix (manually repairing a corrupted NetInfo database is a pain).
Additionally, I can't drop my windows machine, even if I would like to. If the occasional side job comes in with "we need you to modify our windows device driver to do Y", it is much easier to do that with a windows development machine than without.
BTW, I know two science geeks who run linux on their powerbooks. One is even running GNUstep / WindowMaker, which seems really backwards to me.
As far as I can tell, C++ is rather standardized (nowhere near Java though...).
Funny you should say that as C++ is an ISO standard and Java is not a standard at all. In fact, the Java Language specification can't really be called a standard definition because it defines Java's lexical grammar in terms of Java classes:
Section 3.8 of the java language specification:
A "Java letter" is a character for which the method Character.isJavaIdentifierStart returns true. A "Java letter-or-digit" is a character for which the method Character.isJavaIdentifierPart returns true.
You could iterate through all characters (not that you can really do that for all Unicode character sets... but anyway...), and Sun could change their mind and say, "Naw, that's no longer a valid letter." Changing Character's static methods would keep the spec correct but change the lexical rules of the language.
OK, your meaning was actually obvious, just ranting...
Incidentally, the Win32 APIs *are* a standard. ECMA standardized them many years ago (yes, they have changed since). None of the GUI libraries for linux are standard (I may be full of shit, X may be recognized by a standards body). Of course, this just means that ECMA really is a "Standards for sale" organization.
Re:How Israeli Companies Are Succeeding...
on
Business Under Fire
·
· Score: 1
There is a German base on US soil: Hollowman SFB in Alamagordo, New Mexico. OK, so it is a single squadron on a base shared with something like 6 US squadrons, but the base operations are largely paid for by the German government (the base was scheduled to close before the Germans came). So the locals were actually fairly grateful for the "Sleaze Ball"s coming to town. That place would have been a ghost town if the base closed. Now there is a Wallmart and an Applebee's (although, sadly, those are the two biggest local attractions).
The Israeli government is also in talks to house a squadron on US soil.
Oh, and the purpose of the base is not for strategic importance. It's that Germany is roughly the same size as new mexico with ten times the population. They have nowhere to drop bombs for live fire practice other than the ocean. The fact that the US military has roughly the same ammount of live fire ranges as the total land mass of many European contries is not to be underestimated.
Anyway, that's my foreign policy factoid of the day.
Regarding Education: How can it be overrated if the majority says it is worthless?
Are you comparing the actual results against the claims made by college recruiters? Do you hold all marketing claims to the same criteria?
Ok, the rest of this is serious...
Regarding Programming: Yes, you can program quite well without a college degree. And once you have experience, few(er) employers will care that you have one or not.
Even most successful programmers with degrees learned most of their programming outside of school. Part of this is because most successful programmers have an interest in programming, and perhaps didn't wait until college to learn programming, or were first exposed to programming in college and then developed a thirst for it.
I have worked with a number of programmers with and without degrees (and the degrees varied wildly in discipline), and have personally found that formal education, years of experience, etc. are all bad indicators of how good a programmer will be. If you haven't, read Peopleware. The authors actually conducted a study (as opposed to my heresay, or worse, authors like Yourdon who make claims based on personal observation as if they were based on extensive objective research) on programmer productivity and basically determined that the best indicator of a programmer's capability is to measure the capabilities of a coworker. All the other factors seem to have no correllation with performance. Good people work together. Of course, that isn't the criteria used by all HR people, it just happens to be the criteria supported by empirical research.
Back to education: Now, that said, depending on the metric you use to measure success determines whether or not a college education is valuable. Certainly if you are happy with programming and have the experience to stay employed, then you should just be happy with being happy. If you desire advancement in the corporate hierarchy, then college might help for the value of the paper (but then again, certification programs might be a better value for you). If you value education for the sake of education, then college will be valuable for you, and it will be more valuable than alternatives.
There are a few times when education for the paper still causes college to be a worthwhile investment. Government contractors have their rates audited. They have to justify pay. For companies like this (and other large companies that aren't contractors), they may have a formula for determining pay. In that case, the formula may weigh a college degree as 3 years of experience. If the company will pay for you to go to school, then you can get a degree in 5 years while working and net a pay increase equal to 8 years of experience (3 for the paper, 5 for actual experience). But I would argue that the actual education you get from that is only worth whatever interest it provides you (the rest of the value is in the paper degree, not the education).
Cheers, and as long as you recognize the importance of communication and inter-personal skills, then you should be fine as a professional programmer for quite some time.
Hmmm... Mabye I should have stated that I'm not positive that the ok-work visas actually are legit (they may be no-work student visas). The only things I know is that the people claimed to have student visas (which was why they had to sign up for a city college course), and that they certainly had jobs (at various retailers). I would be surprised if the retailers were breaking the law. I live in Santa Barbara, which has a glut of college graduates holding down waiter and retail positions just to keep living here, so it isn't like there's a labor shortage. Heck, even if there were, there are plenty of students looking for retail jobs as well.
As far as your fiance goes, he is lucky. My buddy's wife had a visa for medical school. Their repatriation supercedes family reunification (the med-school visas are the only one that do this AFAIK). She quit her residency one day before it was completed so that she could stay in the US. Of course, she is still an MD, and licensed to practice anywhere in the US other than California, but it is still a pisser that she had to quit. The AMA is a more powerful lobby than most, I guess.
How does anyone keep the visa program clear in their head?
What H-1B visas go to universities? Is that to hire professors? Do you have a link to any information about them not counting toward the quota?
Also, if anyone knows, why is it that some student visas (specifically, every Indian and Chinese student I've met) don't allow them to work and some (specifically, every Eastern European "student" I've met) do? The only differences I see are that the no-work student visas seem to be going to young, single graduate students, while the ok-work student visas seem to go to families that enroll in a single city college course. From what I have seen, the ok-work visas are all being abused by people who want to come to the US for work and can't find another way in. and the no-work visas are all being used in the only way they can be used -- for an education.
OK, I gleefully admit that I am too lazy to type ten words into google when I can simply post my questions to slashdot.:)
It depends on what you mean by large. If you mean 100,000 LOC, yes. If you mean 1,000,000 LOC or more, no. I was the project lead for a Navy flight scheduling system. It was a little over 200k LOC when we hit some major growing pains. The project changed database backend technology. We made that migration in a day (used Perl code to rewrite our code then manually inspected the output and fixed what remained or got butchered by the perl). We changed GUI library components (we used a grid in many places). That change we did piecemeal over a period of a week.
That said, eventually the project leads from all our projects sat down and debated the merits of rewriting the program. We had a lot of assumptions in our system that were essentially hard coded magic numbers. We built schedules for one and two seat aircraft only. We only understood the concept of pilot and WSO. We only could schedule one mission per sortie. Anyway, we had a customer that wanted almost every assumption we made undone. So, we were debating:
1. Rewrite the system from scratch with the lessons learned. 2. Write a new system for the new customer and maintain two codebases. 3. Refactor away.
This time, we chose to do the rewrite. I designed the new database for the rewrite and left the project (due to a personality conflict with the head of the marketing department). The project went over budget by a factor of three or so (and the budget was quite large compared to the first version). Three of the major developers left the company before they hit a release, two of them passing the torch of project lead as they left.
I'm not going to say that the rewrite wasn't necessary. Some rewrites are: moving from a client-server app to a web app (requirement for most systems used by the navy) is one reason. Bad code is just not a very good reason. I'm not even sure bad architecture is a good reason. Most of the design patterns from the GoF book are based around mitigating architecture problems. First, drop in a proxy object, then fix either end of the proxy, then, fix the proxy.
I should stress that the above was not under my current employer.
Anyway, that's my experience. I've designed and built multiple scheduling applications. I've also worked on a payroll system, on a couple hardware control systems, and a single embedded system. None of those ever grew beyond the million line mark, though (the embedded system was probably less than a couple thousand).
Nothing Spolsky says in his essay would have prevented Firefox, nor the better Mozilla codebase. He simply says not to rewrite from scratch. He never says anything about refactoring or improving the existing codebase. Version 2 may not have any code in common with Version 1, but throughout the development process there were feature improvements, architectural improvements, etc. The point is that by starting with a working version 1, even an ugly version 1, if the decision was made to release early, it would have been possible. Once you have something running, don't throw it away.
Of course, there is an old adage, "All absolute statements are wrong, including this one."
I don't mean to debate the accuracy of what he said, just that the interpretation you have is different than my interpretation. However, I do know that my productivity is higher when I modify a working program than when I start over. If the architecture is *really* bad, I could see where it might actually be beneficial to start over, but I think programmers have a tendency to overestimate how unworkable the current system is when the chance to rewrite from scratch appears.
I am a bit confused. You seem to think that it is impossible to be both a student and carry a job. I used to work for a company that regularly hired undergraduate students part time. They were cheap labor and we were willing to work around their hours. We certainly gave them work that is better than helpdesk tasks.
A friend of mine, Murat Karaorman, now head of the CS department of the College of Creative Studies at UCSB, started the program as a graduate student. He was a student (PhD), an instructor (Lecturer), and he worked full time at Panasonic. Add to that the fact that he had two teenage daughters.
I currently know quite a few people who are simultaneously students and professionals. In the end, they typically have more doors opened for them than either the Uni-only or the crap-job-ladder-only individuals.
however, there is one number that will change this discssion, that is 4GBPS, or the average speed between your video card and your monitor.
once we hit that, all bets are off.
Not truly the only factor. While bandwidth has been improving, latency has not. Even when I have a decent connection as far as throughput, a bad ping really makes certain applications dog (think FPS games here).
Further, assuming we continue to use an X client-server model (which seems a practical assumption for cheap clients), applications that use drag and drop will particularly suffer entirely from latency issues.
Have you ever X forwarded Mozilla's mail client across 10 hops but with high bandwidth? Clicking and keyboard works fairly well (there is a slight lag, but after the lag refreshes occur very fast). However, if you want to drag a list of emails to another folder, the client gets bogged down to the point of being useless (at which point you thank the developers for having functionally redundant context menus) due to the large number of bi-directional messaging. Perhaps this is a poor example, saturating the bandwidth produces similar results. Oh well, CounterStrike ping has greater importance than email responsiveness anyway.
This is also s stinging issue for grid computing, not just CounterStrike. Designing around latency is not trivial.
Phone sex operator is fun? Probably not without hazards. Reminds me...
One day I go to the doctor for an injury. The doctor comes back after reviewing the X-Rays and says, "I have some bad news for you. You have tennis elbow and masturbater's wrist."
I say, "But that's impossible. I don't play tennis!"
Kickass. I was aware of the IBM compiler for the Power5, but I wasn't aware that they had finally released it for OS X on G5. Thanks. I am downloading the trial right now. I am concerned about one thing: the compiler gives no indication that it can generate 64bit binaries for the G5. Also, the difference in benchmarks between g++ and XL is not nearly as big as the difference between g++ and Intel's C++ compiler.
Slashdot Mirror
You seem to be unaware of the great fiscal burden of transporting zombies over the Canada-US border post 9/11. It was much easier (and cheaper) for the first few films.
Of course, back then, the US had plenty of zombies with free time to make movies, and as a consequence did not have to import many. Now we have elected all of them to public office.
However, having played resident evil (it was necessary research for a biology class), I assure you that zombie dogs rarely have their damaged tissue repaired, which is a detail the article seems to overlook.
Either you don't know anyone in the medical profession, or you have above average PHBs :)
Most specialists get asked about things that "seem" related (at least in the mind of the PHB or, more often, neighbor) regularly.
...or recycle computer components for Dell (anyone have the link from theregister.co.uk handy?)
How about RTFS(ummary)? The summary clearly asks about other software products or even hardware.
Also, I think that the point was that Netscape had a "complete" product when launching. And I think that the argument of the article only applies if the software is self patching. This is automatically true for a web service, but it could be true for something like a browser (although maybe not over a modem).
True, but nothing compared to the NES. And then the gameboy was a brand on its own.
I'd mod you up as interesting. Made me think about some others:
... blammo!
How old was Nintendo before they made the NES? Nobody cared about the company that made playing cards and crappy plastic puzzles, but then
For that matter, how old was Micro-Soft (or however they used to write it) before Windows (ok, not 25, but still...)? I mean, a company brand based on Solitaire has to mean something.
And how old was the SCO brand before the current company started suing everyone and making everyone hate them?
Some brands exist for a while before being important/infamous.
Actually, the specific thing I wanted to do was create a live slideshow with the ability to dump new photos into the slideshow mix by attaching a camera to the computer. I could automatically download the photos from the camera and bring them into iPhoto for a slideshow. However, iPhoto has limited capabilities with choosing the transitions via script, and no control to force a given image as the next image in the slideshow. PowerPoint can do everything except import a photo (which makes it useless for live updates) -- although the image importing is a known bug. There is also the slideshow application (used by spotlight). It also lacks transition selection via script.
It is possible to do everything by scripting quicktime, but that wasn't really the point of my post. I just wish that Apple opened their applications to their own scripting interface in a way that was actually useful.
Nahh, they'd call it "Tables."
Seriously, though, I would really like Apple to make iWork into a complete product. It isn't just missing an excel and access replacement. It is also missing key "Apple" functionality: applescript capability.
While keynote 2 does expose an applescript dictionary, it is completely useless. Things you should be able to do (but can't) via script:
1. Create a new document (slideshow)
2. Add a slide
3. Edit the slide
4. Set the transition effect
OK, so basically anything useful. The sad part is that Microsoft PowerPoint has an almost useful applescript integration. I say "almost" because the bindings for creating image slides is broken (you get a nice interpreter error when you try to create an image from a file).
AppleWorks did have decent scripting capability.
Heh. I just tried it myself. If it helps, if you can append .csv to the end of the filename, it opens calc, even if the file itself is tab delimited (it asks for the delimeter anyway if it is comma delimited).
The thing that (still) gets me is how long it takes to plot datasets with a few thousand points. gnuplot does it fine, but oocalc is terrible at it.
It's particularly BS because most of the programs I run *are* multi-threaded. So they already have the ability to hyperthread. It's not even a chicken-and-the-egg problem as some have commented. It helps most applications right now. Ok, so grep isn't threaded, but runing ps | grep foo does run as two processes and is helped by a dual core (or hyperthreading, or multiple CPUs).
1. Reversible computing is Turing complete. In fact, this is a standard part of the intro chapters of quantum computing books (as all quantum gates are reversible -- unitary operators). You have to pad extra zeroes to your input to use as scratch paper for the reversible computation (so, your intuition about useless data is correct). However, there is a general technique for reusing the scratch padding (but nobody actually shows the steps in their papers anymore because it is so common). You can find it in a book on quantum computing. Unfortunately, I dropped the course, and don't remember the specifics (and though I can see the book, it is too far to reach).
2. As far as using the Turing model, there is one for reversible computing. Nobody I know uses it when discussing reversible computing. Most people use a circuit model (complexity of the "algorithm" is length of the circuit). Someone obviously did the comparison at some point, but turing machines are painful to work with (very verbose). In general, reversing a computation to recapture the scratch work requires double the "run time" -- you just CNOT the bits you want to save out of the answer part and reverse the whole computation (minus the bits you just saved). What remains is the input plus scratch plus the output. OK, I guess I do remember the specifics.
RE: Accelerated program
Does your university have a policy on challenging a course? I know all of the UC schools allow you to challenge any course that isn't the prerequisite for a course you have already taken.
As an undergrad, I only knew one student that made use of this rule. However, he was driven to graduate in three years with a double major in CS and physics (he succeeded).
I am a mac user. I own a dual G5, a powerbook, and a mac mini. I also own 2 linux machines and one (currently dead, recovery disk in hand) windows machine.
I wouldn't consider putting linux on my powerbook because I would lose the instant-on feature that makes the powerbook such a convenient laptop (I'd also lose MS Office, which is a requirement for work, and Open Office still doesn't properly open most word documents I try due to the presence of arrows and other non-alphabet characters).
I wouldn't put linux on my dual G5 because I couldn't use a number of my commercial applications (Lightwave, Flash, iTunes), as well as devices (don't know if my printer, scanner, camera, iPod, airport, etc. have linux drivers).
I might consider linux on the mini, but my girlfriend uses that computer and I have no real reason to switch... so I probably won't.
As far as considering linux: I probably wouldn't replace my linux boxes with macs. I use them as headless servers, and they compile and run the programs I want without hassle. Minor corruptions are easy to fix (manually repairing a corrupted NetInfo database is a pain).
Additionally, I can't drop my windows machine, even if I would like to. If the occasional side job comes in with "we need you to modify our windows device driver to do Y", it is much easier to do that with a windows development machine than without.
BTW, I know two science geeks who run linux on their powerbooks. One is even running GNUstep / WindowMaker, which seems really backwards to me.
As far as I can tell, C++ is rather standardized (nowhere near Java though...).
Funny you should say that as C++ is an ISO standard and Java is not a standard at all. In fact, the Java Language specification can't really be called a standard definition because it defines Java's lexical grammar in terms of Java classes:
Section 3.8 of the java language specification:
A "Java letter" is a character for which the method Character.isJavaIdentifierStart returns true. A "Java letter-or-digit" is a character for which the method Character.isJavaIdentifierPart returns true.
You could iterate through all characters (not that you can really do that for all Unicode character sets... but anyway...), and Sun could change their mind and say, "Naw, that's no longer a valid letter." Changing Character's static methods would keep the spec correct but change the lexical rules of the language.
OK, your meaning was actually obvious, just ranting...
Incidentally, the Win32 APIs *are* a standard. ECMA standardized them many years ago (yes, they have changed since). None of the GUI libraries for linux are standard (I may be full of shit, X may be recognized by a standards body). Of course, this just means that ECMA really is a "Standards for sale" organization.
There is a German base on US soil: Hollowman SFB in Alamagordo, New Mexico. OK, so it is a single squadron on a base shared with something like 6 US squadrons, but the base operations are largely paid for by the German government (the base was scheduled to close before the Germans came). So the locals were actually fairly grateful for the "Sleaze Ball"s coming to town. That place would have been a ghost town if the base closed. Now there is a Wallmart and an Applebee's (although, sadly, those are the two biggest local attractions).
The Israeli government is also in talks to house a squadron on US soil.
Oh, and the purpose of the base is not for strategic importance. It's that Germany is roughly the same size as new mexico with ten times the population. They have nowhere to drop bombs for live fire practice other than the ocean. The fact that the US military has roughly the same ammount of live fire ranges as the total land mass of many European contries is not to be underestimated.
Anyway, that's my foreign policy factoid of the day.
Regarding Education:
How can it be overrated if the majority says it is worthless?
Are you comparing the actual results against the claims made by college recruiters? Do you hold all marketing claims to the same criteria?
Ok, the rest of this is serious...
Regarding Programming:
Yes, you can program quite well without a college degree. And once you have experience, few(er) employers will care that you have one or not.
Even most successful programmers with degrees learned most of their programming outside of school. Part of this is because most successful programmers have an interest in programming, and perhaps didn't wait until college to learn programming, or were first exposed to programming in college and then developed a thirst for it.
I have worked with a number of programmers with and without degrees (and the degrees varied wildly in discipline), and have personally found that formal education, years of experience, etc. are all bad indicators of how good a programmer will be. If you haven't, read Peopleware. The authors actually conducted a study (as opposed to my heresay, or worse, authors like Yourdon who make claims based on personal observation as if they were based on extensive objective research) on programmer productivity and basically determined that the best indicator of a programmer's capability is to measure the capabilities of a coworker. All the other factors seem to have no correllation with performance. Good people work together. Of course, that isn't the criteria used by all HR people, it just happens to be the criteria supported by empirical research.
Back to education:
Now, that said, depending on the metric you use to measure success determines whether or not a college education is valuable. Certainly if you are happy with programming and have the experience to stay employed, then you should just be happy with being happy. If you desire advancement in the corporate hierarchy, then college might help for the value of the paper (but then again, certification programs might be a better value for you). If you value education for the sake of education, then college will be valuable for you, and it will be more valuable than alternatives.
There are a few times when education for the paper still causes college to be a worthwhile investment. Government contractors have their rates audited. They have to justify pay. For companies like this (and other large companies that aren't contractors), they may have a formula for determining pay. In that case, the formula may weigh a college degree as 3 years of experience. If the company will pay for you to go to school, then you can get a degree in 5 years while working and net a pay increase equal to 8 years of experience (3 for the paper, 5 for actual experience). But I would argue that the actual education you get from that is only worth whatever interest it provides you (the rest of the value is in the paper degree, not the education).
Cheers, and as long as you recognize the importance of communication and inter-personal skills, then you should be fine as a professional programmer for quite some time.
I just ran a few java apps. I use Apple's JVM, so maybe the complaints are for people who like to compile their own VM?
Hmmm... Mabye I should have stated that I'm not positive that the ok-work visas actually are legit (they may be no-work student visas). The only things I know is that the people claimed to have student visas (which was why they had to sign up for a city college course), and that they certainly had jobs (at various retailers). I would be surprised if the retailers were breaking the law. I live in Santa Barbara, which has a glut of college graduates holding down waiter and retail positions just to keep living here, so it isn't like there's a labor shortage. Heck, even if there were, there are plenty of students looking for retail jobs as well.
As far as your fiance goes, he is lucky. My buddy's wife had a visa for medical school. Their repatriation supercedes family reunification (the med-school visas are the only one that do this AFAIK). She quit her residency one day before it was completed so that she could stay in the US. Of course, she is still an MD, and licensed to practice anywhere in the US other than California, but it is still a pisser that she had to quit. The AMA is a more powerful lobby than most, I guess.
How does anyone keep the visa program clear in their head?
:)
What H-1B visas go to universities? Is that to hire professors? Do you have a link to any information about them not counting toward the quota?
Also, if anyone knows, why is it that some student visas (specifically, every Indian and Chinese student I've met) don't allow them to work and some (specifically, every Eastern European "student" I've met) do? The only differences I see are that the no-work student visas seem to be going to young, single graduate students, while the ok-work student visas seem to go to families that enroll in a single city college course. From what I have seen, the ok-work visas are all being abused by people who want to come to the US for work and can't find another way in. and the no-work visas are all being used in the only way they can be used -- for an education.
OK, I gleefully admit that I am too lazy to type ten words into google when I can simply post my questions to slashdot.
It depends on what you mean by large. If you mean 100,000 LOC, yes. If you mean 1,000,000 LOC or more, no. I was the project lead for a Navy flight scheduling system. It was a little over 200k LOC when we hit some major growing pains. The project changed database backend technology. We made that migration in a day (used Perl code to rewrite our code then manually inspected the output and fixed what remained or got butchered by the perl). We changed GUI library components (we used a grid in many places). That change we did piecemeal over a period of a week.
That said, eventually the project leads from all our projects sat down and debated the merits of rewriting the program. We had a lot of assumptions in our system that were essentially hard coded magic numbers. We built schedules for one and two seat aircraft only. We only understood the concept of pilot and WSO. We only could schedule one mission per sortie. Anyway, we had a customer that wanted almost every assumption we made undone. So, we were debating:
1. Rewrite the system from scratch with the lessons learned.
2. Write a new system for the new customer and maintain two codebases.
3. Refactor away.
This time, we chose to do the rewrite. I designed the new database for the rewrite and left the project (due to a personality conflict with the head of the marketing department). The project went over budget by a factor of three or so (and the budget was quite large compared to the first version). Three of the major developers left the company before they hit a release, two of them passing the torch of project lead as they left.
I'm not going to say that the rewrite wasn't necessary. Some rewrites are: moving from a client-server app to a web app (requirement for most systems used by the navy) is one reason. Bad code is just not a very good reason. I'm not even sure bad architecture is a good reason. Most of the design patterns from the GoF book are based around mitigating architecture problems. First, drop in a proxy object, then fix either end of the proxy, then, fix the proxy.
I should stress that the above was not under my current employer.
Anyway, that's my experience. I've designed and built multiple scheduling applications. I've also worked on a payroll system, on a couple hardware control systems, and a single embedded system. None of those ever grew beyond the million line mark, though (the embedded system was probably less than a couple thousand).
Nothing Spolsky says in his essay would have prevented Firefox, nor the better Mozilla codebase. He simply says not to rewrite from scratch. He never says anything about refactoring or improving the existing codebase. Version 2 may not have any code in common with Version 1, but throughout the development process there were feature improvements, architectural improvements, etc. The point is that by starting with a working version 1, even an ugly version 1, if the decision was made to release early, it would have been possible. Once you have something running, don't throw it away.
Of course, there is an old adage, "All absolute statements are wrong, including this one."
I don't mean to debate the accuracy of what he said, just that the interpretation you have is different than my interpretation. However, I do know that my productivity is higher when I modify a working program than when I start over. If the architecture is *really* bad, I could see where it might actually be beneficial to start over, but I think programmers have a tendency to overestimate how unworkable the current system is when the chance to rewrite from scratch appears.
I am a bit confused. You seem to think that it is impossible to be both a student and carry a job. I used to work for a company that regularly hired undergraduate students part time. They were cheap labor and we were willing to work around their hours. We certainly gave them work that is better than helpdesk tasks.
A friend of mine, Murat Karaorman, now head of the CS department of the College of Creative Studies at UCSB, started the program as a graduate student. He was a student (PhD), an instructor (Lecturer), and he worked full time at Panasonic. Add to that the fact that he had two teenage daughters.
I currently know quite a few people who are simultaneously students and professionals. In the end, they typically have more doors opened for them than either the Uni-only or the crap-job-ladder-only individuals.
however, there is one number that will change this discssion, that is 4GBPS, or the average speed between your video card and your monitor.
once we hit that, all bets are off.
Not truly the only factor. While bandwidth has been improving, latency has not. Even when I have a decent connection as far as throughput, a bad ping really makes certain applications dog (think FPS games here).
Further, assuming we continue to use an X client-server model (which seems a practical assumption for cheap clients), applications that use drag and drop will particularly suffer entirely from latency issues.
Have you ever X forwarded Mozilla's mail client across 10 hops but with high bandwidth? Clicking and keyboard works fairly well (there is a slight lag, but after the lag refreshes occur very fast). However, if you want to drag a list of emails to another folder, the client gets bogged down to the point of being useless (at which point you thank the developers for having functionally redundant context menus) due to the large number of bi-directional messaging. Perhaps this is a poor example, saturating the bandwidth produces similar results. Oh well, CounterStrike ping has greater importance than email responsiveness anyway.
This is also s stinging issue for grid computing, not just CounterStrike. Designing around latency is not trivial.
Phone sex operator is fun? Probably not without hazards. Reminds me...
One day I go to the doctor for an injury. The doctor comes back after reviewing the X-Rays and says, "I have some bad news for you. You have tennis elbow and masturbater's wrist."
I say, "But that's impossible. I don't play tennis!"
OK, old joke...
Kickass. I was aware of the IBM compiler for the Power5, but I wasn't aware that they had finally released it for OS X on G5. Thanks. I am downloading the trial right now. I am concerned about one thing: the compiler gives no indication that it can generate 64bit binaries for the G5. Also, the difference in benchmarks between g++ and XL is not nearly as big as the difference between g++ and Intel's C++ compiler.