That would only work until people start doing what they basically do now and confirming that the comments are correct without actually taking the time to verify that fact. (Subtly incorrect is often much worse than obviously incorrect.) Granted, the system would remind the programmer to check, but a rushed glance at the comment will likely result in "yeah, it's fine" when it's not.
Last night, after posting the grandparent, I was thinking about that very problem, and thought that [for big shops with the resources to pull it off] extreme programming might be the answer here: The code and the comments are flagged as dirty after any changes made by the primary programmer, and can only be flagged back to clean after they have been cleared by the secondary programmer.
Again, it would require a huge investment up front [say $100,000 per programmer, and $100,000 for the ultra-cool database-driven document-management content-versioning system with the kitchen sink], but, down the road, it might offer some significant savings, in the form of more stable, more maintainable, and less bug-ridden code.
GRANDPARENT: Especially if you change the code and now the comments are wrong
PARENT: You're incompetent if you don't change the comments to match the code. You're equally incompetent if you come across incorrect comments and leave them in. You're supposed to the job, so do it...
PARENT: As Fred Brooks said, "There is no silver bullet."
A database backend would go a long ways towards providing a silver bullet, i.e. if instead of writing your code to an ASCII text file, you were writing to a document management system that kept doubly linked associations between the lines of code and the comments associated with those lines of code, and if code/comment pairs had dirty bits, so that if you changed one [e.g. the code], then the dirty bit wouldn't get changed to clean until you verified that the other [e.g. the comment] was correct, then that would go a long way towards solving the problem.
I think we are still in the infancy of code/documentation/database integration, however.
However, don't try to sell the line that one has a "right" to something that they didn't produce. That is Communism, and not only does it not work practically, it's ethically and morally unjustifiable as well.
Ethics? Morality? PShaw - get with times, man.
If you see something you want, then seize it for yourself while the seizing's good.
You can also purchase external (and internal, I'm sure) capture devices that capture direct to.dv format.
Do you have any links? I'd be especially interested in a [barebones OEM] card that didn't cost like a gazillion dollars, where "a gazillion" is anything more than about, oh, say $49.95.
PS: In our case, the input signal would be Analog CCTV.
It's just unfortunate that these cards don't also support DV compression. MPEG is nice and all but sometimes when capturing from a camcorder or vhs, you want to edit the resulting video. MPEG is not ideal for this. Granted, DV capture devices do exist but none to my knowledge have a tuner.
1) In our lab, we've been using an analog CCTV signal [which gets us 60 frames per second, versus maybe 15 frames per second for digital cameras]. The ATI TV cards can pick up the signal, but some of the Hauppauge hardware can't. Does anyone know the name of the standard used by an analog CCTV signal, and what specs a card a needs to meet so as to make sure that it will "see" an analog CCTV signal?
2) Digital Video over Firewire [IEEE 1394] is supposed to have a "direct to disk" feature, so that the intermediate "signal -> MPEG" compression layer is not necessary. Does anyone know of a "direct to disk" solution for analog CCTV signals?
3) Finally, how does the Leadtek WinFast hardware compare to the hardware in this review?
I really appreciate all the work that scientists do to protect, understand and appreciate deadly virii in the name of humankind, but for the love of god, people, don't be foolheardy with these biological agents.
Your average, garden variety "scientist" is an utter and complete moron.
I know - I've been working on a big LabVIEW database thing, and, trust me, these people couldn't find their way out of a paper bag if you gave them a flashlight. [And you wouldn't believe the potential safety catastrophes I encountered - time and time again I was warning them about things that could catch on fire and things that could explode; quite frankly, I'm surprised they haven't blown themselves to smithereens by now.]
The fellows who frequent Slashdot typically come from a compsci/physics/math kind of a background, and would be rather shocked to see what passes for a "scientist" these days - you don't need an IQ much higher than a doorknob to get a job working in one of these biology labs. And if the inherent absence of rigor in the "biology" profession weren't bad enough in its own right, the government magnifies the problem almost infinitely by pouring billions of dollars of grant money into the "biology" cesspool every year.
Frankly, I'm at the point where I'd take the word of a good auto mechanic before I'd trust anything a "scientist" had to say.
Today, after lots of plant closings, the UAW has realized that they need to work together with the company to find solutions that build the business as a whole while maintaining a fair cut for them.
Try telling that to General Motors, which is required by its UAW contracts to pay employees for NOT working [hence GM is forced to sell cars at a loss, under "Zero Percent Financing" schemes, just to keep their assembly lines running - i.e. they would lose even more money if the assembly lines were idle, because they would still be responsible for paying the same wages as if the assembly lines were running].
To see what this has done to GM, search the recent news headlines for general+motors+junk.
That doesn't mean organized labor is inherently bad.
Au contraire, organized ANYTHING is bad, and organized labor is particularly bad.
Back in the day, Microsoft used to write a nifty little Java "compiler" called jvc.exe [to differentiate itself from Sun's javac.exe] - this was back before Sun sued them over their Java implementation.
Anyway, jvc.exe in combination with Microsoft's java.exe had the uncanny ability to spot empty loops like that and jump write over them.
[Which, no doubt, did not bring pleasure to the ghost of poor ol' Alan Turing...]
Eric Schmidt, formerly of Novell, now of Google, is a real political animal:
Elton John helps raise money for Gore
September 20, 2000
ATHERTON, Calif. (Reuters) - Flamboyant rock star Elton John, making his first foray into American politics after three decades of performing in the United States, endorsed Vice President Al Gore at a ritzy Silicon Valley fund-raiser... The fund-raiser, at the home of Novell Corp. Chief Executive Eric Schmidt, raised $3.25 million for the Democratic National Committee...
Gore television network debuts in August
April 5, 2005
San Francisco, CA (UPI) -- The television network headed by former U.S. Vice President Al Gore debuts Aug. 1 and will feature short-form content rather than traditional long form news... Current has partnered with Internet search engine Google, executives said, to produce news updates on topics being searched on the Internet...
The question is, which makes more financial sense?
From TFA:
Advances in server technology may give data center managers a little room to breathe. Innovations from chip technology to server construction are improving hardware's capability to deal with high-density environments. According to Gordon Haff, analyst with Nashua, N.H.-based Illuminata, the days of using traditional server cooling methods are over. Companies going forward with dense server environments will need to use every little piece to get closer to its goals. Upshot: A cooler servers start with the processor.
First of all, I wouldn't call the idea of packing more and more blades in a hotter and hotter core some kind of an "advance", but we'll let that one slide.
The obvious answer here is to move the CPUs farther apart. But moving them further apart requires more square footage, and square footage is astronomical in places like San Jose and Manhattan. Therefore the following solution starts to make a whole lot of sense:
Homegrown IT outsourcing options sprout up across rural America.
Welcome to the ever-so-nascent world of rural IT sourcing. Both large and small companies are tapping into a highly skilled but often underemployed IT workforce in lower-cost rural areas--frequently as an alternative to shipping work overseas.
I can't imagine a community college computer science cirriculum of any kind. One of the first computer science courses you have to take is Discrete Math, covering basic boolean logic and set theory, and the university students I was with whined enough as it is; I suspect a community college would simply revolt.
I have Mod Points at the moment, but modding you up would be useless; you'd just get modded back down [as I expect I will be shortly], and, in the process, I'd have lost my chance to reply to you [if I understand Mod Points correctly].
Anyway, the situation is much, much worse than you imagine: A poster later in this thread mentions a Massachusetts "college" that got rid of their "Linear Algebra" requirement because the students couldn't cut the mustard [and I imagine the "Linear Algebra" in question wasn't a whole lot more complicated than multiplying a couple of matrices].
My experience was even more pathetic than that, however: In the first week of teaching a course in "Intermediate C/C++", I tried to impress upon the students the importance of data typing, and of chosing a data type that was appropriate to the problem at hand. I tried to teach them about things like additive and multiplicative overflow [e.g. if you're really serious about your mathematics, then you've got to consider the possibility that adding two positive numbers could give you a negative number, or that multiplying two numbers could give you an NaN], and about granularity in floating point numbers [abcissas and exponents, and how e.g. 32-bit floats lose begin to lose their integer granularity at i = 16M].
The result? The students went absolutely ballistic, stormed the dean's office ["This isn't computer science! This is MATH!!!"], and damned near got me fired.
Well, I hung in there, and finished out the course. For their final project [with several weeks advance notice], I asked them to write a program that would dissect very large files into a series of smaller files [or "chunks", as I called them], each capable of fitting onto a 1.44MB floppy disk, and then reassemble the large files from the little chunks [this was about eight or nine years ago; similar commercial programs now exist to do this sort of thing, such as e.g. WinRAR, which, as I understand it, is very popular with pr0n downloaders].
I figured something like this might take them about a day or so - maybe a solid eight hours on a Saturday afternoon/evening, or four hours on two consecutive Saturday afternoons - but that it was not completely unreasonable, given that they had several weeks to work on it.
The result? No student in the class turned in a working program. Or, as a certain [formerly] rotund radio personality would say: "Zip, Zero, Nada". It was just completely beyond their abilities to even begin to undertake.
Furthermore, this was not the only community college course I taught [although it was certainly the most "advanced"], and I would say that, in the maybe 18 months to 2 years that I was hanging around the community college system, I NEVER SAW A SINGLE "STUDENT" WHO WAS EVEN REMOTELY CAPABLE OF WRITING A WORTHWHILE COMPUTER PROGRAM IN A LANGUAGE LIKE C, OR EVEN REMOTELY CAPABLE OF ADMINISTERING SOMETHING MISSION CRITICAL, LIKE AN RDBMS DATABASE.
I realize that what follows is a profoundly un-PC thing to say, but community college students are morons; their IQs just aren't high enough to do this sort of thing [on average - and yes, I know that any time you take a population of several million, there will always be a few bright bulbs way out at the far end of the bell curve, and that one of those exceptional lights just might be a lone Slashdot reader who stumbles upon this comment]. And [what's possible worse]: Even if they did have the requisite IQ, they don't have the "fire in the belly" that a person needs in order to tackle a complicated problem and see a potential solution through to its completion.
And I'd go even one step further than that: Having taught at major
You wanna supply the formatted and unformatted versions, so that I can compare the two?
PS: Slashdot does allow you to use this little typesetting [or markup, or "formatting"] algorithm popularized by Tim Berners-Lee.
PPS: Believe it or not, WINWORD.EXE allows you to click on "File | Save As... | Save as type | Web Page (*.htm; *.html)" so as to save your data's formatting within the confines of that very same algorithm.
This at least gives us the right to our own data back, since we can then convert it to a more useable format...
That microsoft opens up their format for reading, and specifies parts of it, makes it possible to write software to convert this data to a open format, or index it and such. Therefor, we can still save in MS format, but have much-less tie in.
You seem to be under the impression that ".DOC" documents use something other than eight bit ASCII characters to store data. Try this: Open up WINWORD.EXE, type in "abcdefg", save the file as "abcdefg.DOC", then open up "abcdefg.DOC" with NOTEPAD.EXE.
Guess what? NOTEPAD.EXE will show you that your data, the string "abcdefg" is there in the file just as it ought to be.
There is no loss of data when using WINWORD.EXE; rather, there is a gain of typesetting [or markup, or "formatting"] structure that other typesetting [or markup, or "formatting"] programs might not be able to understand.
Microsoft owns the rights to their own proprietary typesetting [or markup, or "formatting"] algorithms, but they make absolutely no claims whatsoever on the underlying data that those algorithms act upon.
If you don't like Microsoft's typesetting algorithms, then use Corel's [WordPerfect], or IBM's [Lotus Word Pro], or Apple's [iWork], or hell, even Donald Knuth's.
And after you've tried those other proprietary algorithms, ask yourself whether Microsoft's proprietary typesetting algorithms failed to offer you any value for your money.
Besides, even if none of what I've said is true, you can still always take your ".DOC" documents, open them in WINWORD.EXE, and click on "File | Save As... | Save as type | Text Only (*.txt)" and never have to deal with Microsoft for the remainder of the life of your data.
But doesn't the CLR merely turn the byte-code into machine code? Maybe I'm oversimplifying things here, but I would assume that those could be compiled as well, the same way any #INCLUDE statement would do it. The benefit to this would be able to compile the same source using a mono compiler, and get the benefits of Mono, but where it isn't needed, just compile it to native lang. As long as things don't fork (big if here), we should be fine.
I think you're talking about doing away with the virtual machines, and having syntactically similar languages that "compile" straight to machine language executables rather than to "byte code" virtual machine classes. There are several companies that offer commercial products that do something like that for Java, although I don't know whether anyone has done it yet for C#.
My point, though, was that while it's relatively trivial to create the compiler to do this, the compiler, in and of itself, doesn't get you much: It's the feature-rich library [not the "compile once, run anywhere" nonsense] that gives the virtual machine paradigm its power in the marketplace. Without the feature-rich libraries, you get maybe the ability to write simple command line "Hello World!" kinds of programs, but not much more.
So it seems to me that you'd need to go to Sun and ask, "Hey, will you port all your feature-rich Java libraries to my Java-esque language that compiles to machine language executables?" Or to Microsoft, and ask "Hey, will you port all your feature-rich.NET libraries to my C#-esque language that compiles to machine language executables?"
To which I imagine they'd respond: "Go fly a kite."
First, release a C# compiler for linux. If you can get developers to user C# that compiles to native code so that it's not dependant on the CLR, then you can introduce the cross-platform aspects of it.
While I agree that it would be nice to do away with the overhead consumed by the virtual machine, I was always under the impression that [with maybe a few exceptions*] the creation of a compiler for a language is relatively trivial.
The thing that makes these virtual machine languages so powerful [and the thing that makes them so appealing to the marketplace] is not that nonsense about "compile once, run anywhere" [which, as I understand it, has largely proved to be a fraud], but rather the feature-richness of their libraries: So much stuff has been written into the libraries that the application programmer's job devolves into not much more than providing a little glue to hold together the specific library calls that his project needs [i.e. it's foolish to reinvent the kitchen sink if it's already been invented for you].
*One instance where I've heard that the compiler end of the thing is rather impressive is javac's handling of multithreading logic.
Put System swapfile, temporary internet files and temp folders on a Flash memory RAID array
I didn't think that it was possible to separate "temporary internet files" from the rest of the profile.
I know you can use things like roaming profiles to move the entire profile directory [essentially the value that lives at %USERPROFILE%] to a network share, but again, that tends to move the entire profile there, not just parts of it:
Application Data
Cookies
Desktop
Favorites
Local Settings
My Documents
NetHood
PrintHood
Recent
SendTo
Start Menu
And that was why people got out of the habit of using roaming profiles [about five or six years ago]: Because roaming profiles forced GBs and GBs of useless, nonsense data to live at the server, it was damned near impossible to pull it all back to the local [client] machine each time you logged on.
But if there is a way to select some of the directories to live in one place, and others to live in another place, then PLEASE TELL ME HOW TO DO IT!!!
I'd like nothing better than to be able to keep most of that stuff on the server [the remote machine], but keep the really GB-intensive nonsense stuff, such as
Local Settings\Temporary Internet Files
on the local machine.
That's a pretty big problem.
on
The PC Is Not Dead
·
· Score: 2, Informative
Only problem is when the main servers go down you're killing not just one user but a whole organization.
Uh, that's a pretty big problem.
But then again, a single point of failure usually is.
Step 3. Make the business requirements clear, valid, non-conflicting, and technically viable
My experience has been that this step always gets done in reverse: The programmer assesses the situation, decides what the "business requirements" are, and then, when he's finished creating the solution, that solution becomes the "business requirement", or even the "business model" itself.
I.e. the programmer is the only one on the team with an IQ sufficient to imagine and then create a business structure; the management team exists solely for the purpose of enforcing [on the "employees"] the structure created by the programmer.
Step 4. Give the developers a good development environment (hardware, software, and surroundings)
This is really key: Don't go telling people that you're embarking on some big new project that's got all these great funding guarantees from the suits who live higher up the food chain, and then try to nickel and dime them to death on the things they need to get the job done:
"Sorry, we don't have any money in the budget to get you a keyboard with keys that don't stick..."
"Sorry, we don't have any money in the budget to get you a chair that isn't broken and doesn't cause your back to ache incessantly every day you sit in it..."
"Sorry, we don't have any money in the budget to purchase you a dual [or triple] head video card and two [or three] monitors so that you can see your work on screen..."
"Sorry, we don't have any money in the budget to purchase the Developer's Kit. Here, use this 30-day trial evaluation package they sent us, and just remember that at the end of every month, you need to set the date on your computer back to January 1, 2003..."
"Sorry, we don't have any money in the budget to give you an office. Here, sit at this bench along with the secretarial staff and share this computer with the receptionist..."
I mean, if the money isn't there, then don't undertake the fsck-ing project in the first place.
And sure as hell don't go jet-setting all over the damned globe to your precious "conferences" while your technical staff are being denied the hardware they need to get the damned job done.
Slashdot Mirror
That would only work until people start doing what they basically do now and confirming that the comments are correct without actually taking the time to verify that fact. (Subtly incorrect is often much worse than obviously incorrect.) Granted, the system would remind the programmer to check, but a rushed glance at the comment will likely result in "yeah, it's fine" when it's not.
Last night, after posting the grandparent, I was thinking about that very problem, and thought that [for big shops with the resources to pull it off] extreme programming might be the answer here: The code and the comments are flagged as dirty after any changes made by the primary programmer, and can only be flagged back to clean after they have been cleared by the secondary programmer.
Again, it would require a huge investment up front [say $100,000 per programmer, and $100,000 for the ultra-cool database-driven document-management content-versioning system with the kitchen sink], but, down the road, it might offer some significant savings, in the form of more stable, more maintainable, and less bug-ridden code.
GRANDPARENT: Especially if you change the code and now the comments are wrong
PARENT: You're incompetent if you don't change the comments to match the code. You're equally incompetent if you come across incorrect comments and leave them in. You're supposed to the job, so do it...
PARENT: As Fred Brooks said, "There is no silver bullet."
A database backend would go a long ways towards providing a silver bullet, i.e. if instead of writing your code to an ASCII text file, you were writing to a document management system that kept doubly linked associations between the lines of code and the comments associated with those lines of code, and if code/comment pairs had dirty bits, so that if you changed one [e.g. the code], then the dirty bit wouldn't get changed to clean until you verified that the other [e.g. the comment] was correct, then that would go a long way towards solving the problem.
I think we are still in the infancy of code/documentation/database integration, however.
However, don't try to sell the line that one has a "right" to something that they didn't produce. That is Communism, and not only does it not work practically, it's ethically and morally unjustifiable as well.
Ethics? Morality? PShaw - get with times, man.
If you see something you want, then seize it for yourself while the seizing's good.
You're welcome.
THANKS!!!
I have no idea why different capture cards would have a problem with it. I would think that any video capture card would work.
My impression is that analog CCTV doesn't have a "channel" [e.g. WCBS New York Channel 2, WNBC New York Channel 4, WABC New York Channel 7, etc].
It seems as though the default Hauppauge software package can't see analog signals unless they have a "channel".
You can also purchase external (and internal, I'm sure) capture devices that capture direct to
Do you have any links? I'd be especially interested in a [barebones OEM] card that didn't cost like a gazillion dollars, where "a gazillion" is anything more than about, oh, say $49.95.
PS: In our case, the input signal would be Analog CCTV.
It's just unfortunate that these cards don't also support DV compression. MPEG is nice and all but sometimes when capturing from a camcorder or vhs, you want to edit the resulting video. MPEG is not ideal for this. Granted, DV capture devices do exist but none to my knowledge have a tuner.
1) In our lab, we've been using an analog CCTV signal [which gets us 60 frames per second, versus maybe 15 frames per second for digital cameras]. The ATI TV cards can pick up the signal, but some of the Hauppauge hardware can't. Does anyone know the name of the standard used by an analog CCTV signal, and what specs a card a needs to meet so as to make sure that it will "see" an analog CCTV signal?
2) Digital Video over Firewire [IEEE 1394] is supposed to have a "direct to disk" feature, so that the intermediate "signal -> MPEG" compression layer is not necessary. Does anyone know of a "direct to disk" solution for analog CCTV signals?
3) Finally, how does the Leadtek WinFast hardware compare to the hardware in this review?
Thanks!
I really appreciate all the work that scientists do to protect, understand and appreciate deadly virii in the name of humankind, but for the love of god, people, don't be foolheardy with these biological agents.
Your average, garden variety "scientist" is an utter and complete moron.
I know - I've been working on a big LabVIEW database thing, and, trust me, these people couldn't find their way out of a paper bag if you gave them a flashlight. [And you wouldn't believe the potential safety catastrophes I encountered - time and time again I was warning them about things that could catch on fire and things that could explode; quite frankly, I'm surprised they haven't blown themselves to smithereens by now.]
The fellows who frequent Slashdot typically come from a compsci/physics/math kind of a background, and would be rather shocked to see what passes for a "scientist" these days - you don't need an IQ much higher than a doorknob to get a job working in one of these biology labs. And if the inherent absence of rigor in the "biology" profession weren't bad enough in its own right, the government magnifies the problem almost infinitely by pouring billions of dollars of grant money into the "biology" cesspool every year.
Frankly, I'm at the point where I'd take the word of a good auto mechanic before I'd trust anything a "scientist" had to say.
Today, after lots of plant closings, the UAW has realized that they need to work together with the company to find solutions that build the business as a whole while maintaining a fair cut for them.
Try telling that to General Motors, which is required by its UAW contracts to pay employees for NOT working [hence GM is forced to sell cars at a loss, under "Zero Percent Financing" schemes, just to keep their assembly lines running - i.e. they would lose even more money if the assembly lines were idle, because they would still be responsible for paying the same wages as if the assembly lines were running].
To see what this has done to GM, search the recent news headlines for general+motors+junk.
That doesn't mean organized labor is inherently bad.
Au contraire, organized ANYTHING is bad, and organized labor is particularly bad.
I'm running IE v6.0.2800.1106, on Windows 2000 Advanced Server, with Service Pack 4, and all critical updates.
The smiley-face hello world page renders just fine for me.
Maybe it's a flaw in XP.
Back in the day, the old MS-DOS editing program, EDIT.COM, ran a polling loop that would drive the CPU up to 100%.
The Intel guys used to recommend it as a way to stress test your system.
Back in the day, Microsoft used to write a nifty little Java "compiler" called jvc.exe [to differentiate itself from Sun's javac.exe] - this was back before Sun sued them over their Java implementation.
Anyway, jvc.exe in combination with Microsoft's java.exe had the uncanny ability to spot empty loops like that and jump write over them.
[Which, no doubt, did not bring pleasure to the ghost of poor ol' Alan Turing...]
Eric Schmidt, formerly of Novell, now of Google, is a real political animal:
The question is, which makes more financial sense?
From TFA:
First of all, I wouldn't call the idea of packing more and more blades in a hotter and hotter core some kind of an "advance", but we'll let that one slide.The obvious answer here is to move the CPUs farther apart. But moving them further apart requires more square footage, and square footage is astronomical in places like San Jose and Manhattan. Therefore the following solution starts to make a whole lot of sense:
And in the middle of nowhere, square footage is essentially free.No, it failed because people want REAL computers, and REAL computers are CHEAP.
No, it failed because poor people don't have any money to spend. [Which is what certain hoity-toity types would call a "tautology".]
You can't squeeze water out of a rock.
Each business if different but getting one running for 90K is just about impossible. That won't even cover two people for six months.
In Soviet Russia, real men don't take salaries from their startups - they burn through their life's savings.
And when their life savings are spent, they max out all their credit cards.
And when they've maxed out all their credit cards, they auction off their mint condition copy of Amazing Fantasy #15.
I can't imagine a community college computer science cirriculum of any kind. One of the first computer science courses you have to take is Discrete Math, covering basic boolean logic and set theory, and the university students I was with whined enough as it is; I suspect a community college would simply revolt.
I have Mod Points at the moment, but modding you up would be useless; you'd just get modded back down [as I expect I will be shortly], and, in the process, I'd have lost my chance to reply to you [if I understand Mod Points correctly].
Anyway, the situation is much, much worse than you imagine: A poster later in this thread mentions a Massachusetts "college" that got rid of their "Linear Algebra" requirement because the students couldn't cut the mustard [and I imagine the "Linear Algebra" in question wasn't a whole lot more complicated than multiplying a couple of matrices].
My experience was even more pathetic than that, however: In the first week of teaching a course in "Intermediate C/C++", I tried to impress upon the students the importance of data typing, and of chosing a data type that was appropriate to the problem at hand. I tried to teach them about things like additive and multiplicative overflow [e.g. if you're really serious about your mathematics, then you've got to consider the possibility that adding two positive numbers could give you a negative number, or that multiplying two numbers could give you an NaN], and about granularity in floating point numbers [abcissas and exponents, and how e.g. 32-bit floats lose begin to lose their integer granularity at i = 16M].
The result? The students went absolutely ballistic, stormed the dean's office ["This isn't computer science! This is MATH!!!"], and damned near got me fired.
Well, I hung in there, and finished out the course. For their final project [with several weeks advance notice], I asked them to write a program that would dissect very large files into a series of smaller files [or "chunks", as I called them], each capable of fitting onto a 1.44MB floppy disk, and then reassemble the large files from the little chunks [this was about eight or nine years ago; similar commercial programs now exist to do this sort of thing, such as e.g. WinRAR, which, as I understand it, is very popular with pr0n downloaders].
I figured something like this might take them about a day or so - maybe a solid eight hours on a Saturday afternoon/evening, or four hours on two consecutive Saturday afternoons - but that it was not completely unreasonable, given that they had several weeks to work on it.
The result? No student in the class turned in a working program. Or, as a certain [formerly] rotund radio personality would say: "Zip, Zero, Nada". It was just completely beyond their abilities to even begin to undertake.
Furthermore, this was not the only community college course I taught [although it was certainly the most "advanced"], and I would say that, in the maybe 18 months to 2 years that I was hanging around the community college system, I NEVER SAW A SINGLE "STUDENT" WHO WAS EVEN REMOTELY CAPABLE OF WRITING A WORTHWHILE COMPUTER PROGRAM IN A LANGUAGE LIKE C, OR EVEN REMOTELY CAPABLE OF ADMINISTERING SOMETHING MISSION CRITICAL, LIKE AN RDBMS DATABASE.
I realize that what follows is a profoundly un-PC thing to say, but community college students are morons; their IQs just aren't high enough to do this sort of thing [on average - and yes, I know that any time you take a population of several million, there will always be a few bright bulbs way out at the far end of the bell curve, and that one of those exceptional lights just might be a lone Slashdot reader who stumbles upon this comment]. And [what's possible worse]: Even if they did have the requisite IQ, they don't have the "fire in the belly" that a person needs in order to tackle a complicated problem and see a potential solution through to its completion.
And I'd go even one step further than that: Having taught at major
You wanna supply the formatted and unformatted versions, so that I can compare the two?
PS: Slashdot does allow you to use this little typesetting [or markup, or "formatting"] algorithm popularized by Tim Berners-Lee.
PPS: Believe it or not, WINWORD.EXE allows you to click on "File | Save As... | Save as type | Web Page (*.htm; *.html)" so as to save your data's formatting within the confines of that very same algorithm.
This at least gives us the right to our own data back, since we can then convert it to a more useable format...
That microsoft opens up their format for reading, and specifies parts of it, makes it possible to write software to convert this data to a open format, or index it and such. Therefor, we can still save in MS format, but have much-less tie in.
You seem to be under the impression that ".DOC" documents use something other than eight bit ASCII characters to store data. Try this: Open up WINWORD.EXE, type in "abcdefg", save the file as "abcdefg.DOC", then open up "abcdefg.DOC" with NOTEPAD.EXE.
Guess what? NOTEPAD.EXE will show you that your data, the string "abcdefg" is there in the file just as it ought to be.
There is no loss of data when using WINWORD.EXE; rather, there is a gain of typesetting [or markup, or "formatting"] structure that other typesetting [or markup, or "formatting"] programs might not be able to understand.
Microsoft owns the rights to their own proprietary typesetting [or markup, or "formatting"] algorithms, but they make absolutely no claims whatsoever on the underlying data that those algorithms act upon.
If you don't like Microsoft's typesetting algorithms, then use Corel's [WordPerfect], or IBM's [Lotus Word Pro], or Apple's [iWork], or hell, even Donald Knuth's.
And after you've tried those other proprietary algorithms, ask yourself whether Microsoft's proprietary typesetting algorithms failed to offer you any value for your money.
Besides, even if none of what I've said is true, you can still always take your ".DOC" documents, open them in WINWORD.EXE, and click on "File | Save As... | Save as type | Text Only (*.txt)" and never have to deal with Microsoft for the remainder of the life of your data.
But doesn't the CLR merely turn the byte-code into machine code? Maybe I'm oversimplifying things here, but I would assume that those could be compiled as well, the same way any #INCLUDE statement would do it. The benefit to this would be able to compile the same source using a mono compiler, and get the benefits of Mono, but where it isn't needed, just compile it to native lang. As long as things don't fork (big if here), we should be fine.
I think you're talking about doing away with the virtual machines, and having syntactically similar languages that "compile" straight to machine language executables rather than to "byte code" virtual machine classes. There are several companies that offer commercial products that do something like that for Java, although I don't know whether anyone has done it yet for C#.
My point, though, was that while it's relatively trivial to create the compiler to do this, the compiler, in and of itself, doesn't get you much: It's the feature-rich library [not the "compile once, run anywhere" nonsense] that gives the virtual machine paradigm its power in the marketplace. Without the feature-rich libraries, you get maybe the ability to write simple command line "Hello World!" kinds of programs, but not much more.
So it seems to me that you'd need to go to Sun and ask, "Hey, will you port all your feature-rich Java libraries to my Java-esque language that compiles to machine language executables?" Or to Microsoft, and ask "Hey, will you port all your feature-rich .NET libraries to my C#-esque language that compiles to machine language executables?"
To which I imagine they'd respond: "Go fly a kite."
First, release a C# compiler for linux. If you can get developers to user C# that compiles to native code so that it's not dependant on the CLR, then you can introduce the cross-platform aspects of it.
While I agree that it would be nice to do away with the overhead consumed by the virtual machine, I was always under the impression that [with maybe a few exceptions*] the creation of a compiler for a language is relatively trivial.
The thing that makes these virtual machine languages so powerful [and the thing that makes them so appealing to the marketplace] is not that nonsense about "compile once, run anywhere" [which, as I understand it, has largely proved to be a fraud], but rather the feature-richness of their libraries: So much stuff has been written into the libraries that the application programmer's job devolves into not much more than providing a little glue to hold together the specific library calls that his project needs [i.e. it's foolish to reinvent the kitchen sink if it's already been invented for you].
*One instance where I've heard that the compiler end of the thing is rather impressive is javac's handling of multithreading logic.
Put System swapfile, temporary internet files and temp folders on a Flash memory RAID array
I didn't think that it was possible to separate "temporary internet files" from the rest of the profile.
I know you can use things like roaming profiles to move the entire profile directory [essentially the value that lives at %USERPROFILE%] to a network share, but again, that tends to move the entire profile there, not just parts of it:
And that was why people got out of the habit of using roaming profiles [about five or six years ago]: Because roaming profiles forced GBs and GBs of useless, nonsense data to live at the server, it was damned near impossible to pull it all back to the local [client] machine each time you logged on.But if there is a way to select some of the directories to live in one place, and others to live in another place, then PLEASE TELL ME HOW TO DO IT!!!
I'd like nothing better than to be able to keep most of that stuff on the server [the remote machine], but keep the really GB-intensive nonsense stuff, such as
on the local machine.Only problem is when the main servers go down you're killing not just one user but a whole organization.
Uh, that's a pretty big problem.
But then again, a single point of failure usually is.
Although the fact that I work for Oracle can possibly have something to do with that...
Cheater.
Step 3. Make the business requirements clear, valid, non-conflicting, and technically viable
My experience has been that this step always gets done in reverse: The programmer assesses the situation, decides what the "business requirements" are, and then, when he's finished creating the solution, that solution becomes the "business requirement", or even the "business model" itself.
I.e. the programmer is the only one on the team with an IQ sufficient to imagine and then create a business structure; the management team exists solely for the purpose of enforcing [on the "employees"] the structure created by the programmer.
Step 4. Give the developers a good development environment (hardware, software, and surroundings)
This is really key: Don't go telling people that you're embarking on some big new project that's got all these great funding guarantees from the suits who live higher up the food chain, and then try to nickel and dime them to death on the things they need to get the job done:
I mean, if the money isn't there, then don't undertake the fsck-ing project in the first place.And sure as hell don't go jet-setting all over the damned globe to your precious "conferences" while your technical staff are being denied the hardware they need to get the damned job done.