Think about it for a second: this is the way that "real" news agencies function in the "real world". Most news boils down to a couple of talking heads getting their soundbites propagated to a mass of news consumers. This necessarily must be the way that sites like/. do their thing, because the free software community is nothing more than a swirling mass of opinionated posts, heated discussions, and flameage. Somehow, miraculously, a little code gets written in the process -- but news regarding architectural plans and project-management plans doesn't have much code to go on -- the opinionated posts are the news in this case. If you think that a news agency isn't providing the information that you would like to consume, you are free to abstain. Better yet, provide a competing news service. Kernel Traffic is an excellent model (I prefer it to/.) but it lacks the entertainment of pedantic posts like mine explaining that journalism is nothing more than a trade: this concept of "journalistic status" is a holdover from the days when people were foolish enough to think of it as a 4th branch of government. In reality, it's just a business of shuffling tasty bits to consumers.
You could spin it that way, but I've been hoping for years that some company would compromise a bit of performance in exchange for a fanless (ergo "quiet") computer. IMO the fanless iMacs and the new cube are both changing the world for the better. Of course, I'd put Debian/PPC on one instead of MacOS X...
NeXT shipped color in 1990. The codebase was alive as long as the company was alive: right up until it was sold to Apple for US$450 million, and then it continued to live as MacOSX.
Wait a minute...the one of the cube's design goals was to be fanless, yet every picture I've seen of the Radeon card shows that it requires its own cooling fan. What gives?
Although Apple bought NeXT, the reality of the situation is that the old Apple was assimilated by NeXT soon afterward. The dual PPC machine is a rebirth of the NRW (NeXT RISC Workstation), MacOSX is NeXTstep, and now the cube is reborn. What greater evidence do you need? Note that the graphite style is a lot closer to the elegant black hardware design than it is to anything Apple offered. Fly your pirate flag, Avie.
Your observations are conveniently one-sided. Don't forget that Apple Enterprise (nee NeXT Computer) took took took from the GPL world, one example being proprietary changes to the Objective C runtime -- and it took years of whining to get them to honor the license. If your honest, you'll see greed everywhere, not just in the/. community.
I use mozilla every day, but my employer has 'standardized' on IE. Normally, this would be something that I could shrug-off, since mozilla displays everything correctly that I can throw at it. Unfortunately, intranet documents are being mounted on NT servers configured to use the "NTLM" authentication method, which neither mozilla or netscape classic can do...so I can't even get to the documents. Trying to get the web administrators to use another authentication method is like pulling healthy teeth. "Why should we? We've standardized on IE!" There is a bugzilla report about this, and it seems close to being resolved...so there's hope I guess.
While connection pooling could be used with either database engine, I would expect Postgresql to get the greater benefit, given its reputation for having greater overhead during authentication connection establishment. Pooling solves this problem completely, by factoring this latency out of each page fetch. I would be very interested in benchmarks between these two engines where pooling was used. I get the impression that MySQL gets a great reputation for performance only because people tend to use database engines in the naive way: every web hit creates a new connection.
I haven't read the slashdot code, but the number of ccncurrent users per unit of time doesn't necessarily translate to the connection load on the database -- provided that connection pooling is being used. If the nature of the system permits, thousands of users could share a mere dozen connections in a pool so long as they released connections as soon as they were done with them. This is really easy to do in java servlets where any servlet has access to some pooling class which pools the connections in static members. I don't know how I'd do this in a mod_perl environment. (Anybody care to comment?)
Anyway, the obligitory on-topic bit (lest a moderator spank me for engaging in tangents) is that I love Postgresql, and have been watching the dev team make great strides in recent past. Thank you, kind coders, for giving the world such a beautiful free database engine!
Grandma knows how to pay the boy next door to cut her lawn...she can pay him to troubleshoot her usb problem just as easily. "Here, Mrs. Wilson...it had come unplugged." "Thanks, Jimmy, here's $5."
The claim that information "wants" to be free is a charaterization of its nature. It doesn't necessarily mean that you or I should want a specific piece of information to be free. What I want may be different from what the information "wants". Consider this analogy: matter doesn't seem to "want" to be permanently formed into useful configurations. That doesn't mean we should bow down to what matter "wants". Quite the opposite: you need to make entropy your enemy if you want to make matter useful. Back to information: it's very difficult to contain, by virtue of the fact that replication has near-zero cost. That shouldn't imply that one ought never attempt to contain it.
My wife applied for a position. The offer came through email in the form of a Word document that had been created using somebody else's offer as a template. The only thing she was concerned about was the bottom-line, so she viewed the attachment using strings(1), piping it to less(1), and visually grepping for the yearly figure...what she saw was absurdely low, and she wanted to tell them to 'get bent'. A cool head prevailed...she opened it in Word later and the offer was good -- but went back to examine the innards: there was somebody else's offer complete with their name and address, buried within the saved 'undo' stack. Knowing this, I don't understand how businesses could be so careless about sending Word documents outside the company.
The 'language agnostic' thing is nothing more than a universal virtual machine. What they're saying is that JVMs can only run Java, but their runtime will run C# and anything else. They're assuming that nobody has noticed that JVMs can run python, scheme, prolog, and jeeze what else?
This will be marked 'offtopic' for being tangential, but here goes: the Model 4 rules the retrocomputing universe. I just acquired one, and even managed to find a hires board for it, and the Alcor C compiler (docs in mint condition). I'm going to equip it with 4 half-height drives in the bay, and write open-source ground-control software for this mission. Yes I am.
P.S. Anybody with Model 4 stuff to sell, drop me a line.
"Easy" is a relative term. Anything is "easy" if you already know how to do it, and if you want to reduce the number of keystrokes involved you can write a script. In general, you just copy the file to the new target name, then do a cvs add of the new file, then rm the old file & do a cvs remove on it. Finally, a cvs commit and you're done. I agree it's not as easy as it could be, but it's not something that requires frobbing the CVSROOT control files directly.
I agree with these criticisms, and that CVS is in dire need of either replacement, or a forking of the CVS codebase. The development mailing list is heavily guarded by a few people who work hard to prevent any progress in the development of CVS, so replacement is probably the answer.
One feature that I would like to see is the ability to configure the diff algorithm by file type, so that revisioned binary files (images, for instance) can be stored in the repository on a space-efficient manner. PRCS does this with the xdelta library, but I don't know if PRCS allows the use of a line-oriented-text diff algorithm for source files.
I also find the event hooks in CVS to be lacking. A generalized hook mechanism is needed that has access to both the name of the checked-in file and its contents, with both pre-commit and post-commit triggers, by file name, file type, directory, module, user, and whatever other parameter might be useful. CVS has some of this, but not enough.
Anyone who resorts to mucking-around in CVSROOT manually needs to read the FAQ.
Re:If he were to publish this article...
on
Censorship In China
·
· Score: 1
Also available on from that site:
"Tornado in a Junkyard: The Relentless Myth of Darwinism" (more info...) In this engaging page-turner, James Perloff devastates the theory of evolution. Issues tackled include the lack of transitional forms in the fossil record, the lack of evidence for "ape-men," and the amazing history of fraud that has characterized the efforts to get Darwin's theory, which he himself questioned, accepted and taught as fact. 321 pages. Paperback.
Slashdot Mirror
Think about it for a second: this is the way that "real" news agencies function in the "real world". Most news boils down to a couple of talking heads getting their soundbites propagated to a mass of news consumers. This necessarily must be the way that sites like /. do their thing, because the free software community is nothing more than a swirling mass of opinionated posts, heated discussions, and flameage. Somehow, miraculously, a little code gets written in the process -- but news regarding architectural plans and project-management plans doesn't have much code to go on -- the opinionated posts are the news in this case. If you think that a news agency isn't providing the information that you would like to consume, you are free to abstain. Better yet, provide a competing news service. Kernel Traffic is an excellent model (I prefer it to /.) but it lacks the entertainment of pedantic posts like mine explaining that journalism is nothing more than a trade: this concept of "journalistic status" is a holdover from the days when people were foolish enough to think of it as a 4th branch of government. In reality, it's just a business of shuffling tasty bits to consumers.
You could spin it that way, but I've been hoping for years that some company would compromise a bit of performance in exchange for a fanless (ergo "quiet") computer. IMO the fanless iMacs and the new cube are both changing the world for the better. Of course, I'd put Debian/PPC on one instead of MacOS X...
NeXT shipped color in 1990. The codebase was alive as long as the company was alive: right up until it was sold to Apple for US$450 million, and then it continued to live as MacOSX.
Yeah, aside from shared libraries, there's no reuse of code. snicker
Wait a minute...the one of the cube's design goals was to be fanless, yet every picture I've seen of the Radeon card shows that it requires its own cooling fan. What gives?
Although Apple bought NeXT, the reality of the situation is that the old Apple was assimilated by NeXT soon afterward. The dual PPC machine is a rebirth of the NRW (NeXT RISC Workstation), MacOSX is NeXTstep, and now the cube is reborn. What greater evidence do you need? Note that the graphite style is a lot closer to the elegant black hardware design than it is to anything Apple offered. Fly your pirate flag, Avie.
clue-4-u, u silly grammar-nazi.
Your observations are conveniently one-sided. Don't forget that Apple Enterprise (nee NeXT Computer) took took took from the GPL world, one example being proprietary changes to the Objective C runtime -- and it took years of whining to get them to honor the license. If your honest, you'll see greed everywhere, not just in the /. community.
Around my neck of the woods, VB was the Next Big Thing for years...fortunately, Java seems to be taking that title.
I use mozilla every day, but my employer has 'standardized' on IE. Normally, this would be something that I could shrug-off, since mozilla displays everything correctly that I can throw at it. Unfortunately, intranet documents are being mounted on NT servers configured to use the "NTLM" authentication method, which neither mozilla or netscape classic can do...so I can't even get to the documents. Trying to get the web administrators to use another authentication method is like pulling healthy teeth. "Why should we? We've standardized on IE!" There is a bugzilla report about this, and it seems close to being resolved...so there's hope I guess.
While connection pooling could be used with either database engine, I would expect Postgresql to get the greater benefit, given its reputation for having greater overhead during authentication connection establishment. Pooling solves this problem completely, by factoring this latency out of each page fetch. I would be very interested in benchmarks between these two engines where pooling was used. I get the impression that MySQL gets a great reputation for performance only because people tend to use database engines in the naive way: every web hit creates a new connection.
Anyway, the obligitory on-topic bit (lest a moderator spank me for engaging in tangents) is that I love Postgresql, and have been watching the dev team make great strides in recent past. Thank you, kind coders, for giving the world such a beautiful free database engine!
Yes, of course, because that is all that you understand.
Take a break, walk away from your computer. Hatred is unhealthy. Show wisdom by not engaging in it.
unzip clangref.exe
catdoc clangref.doc | less
Grandma knows how to pay the boy next door to cut her lawn...she can pay him to troubleshoot her usb problem just as easily. "Here, Mrs. Wilson...it had come unplugged." "Thanks, Jimmy, here's $5."
The claim that information "wants" to be free is a charaterization of its nature. It doesn't necessarily mean that you or I should want a specific piece of information to be free. What I want may be different from what the information "wants". Consider this analogy: matter doesn't seem to "want" to be permanently formed into useful configurations. That doesn't mean we should bow down to what matter "wants". Quite the opposite: you need to make entropy your enemy if you want to make matter useful. Back to information: it's very difficult to contain, by virtue of the fact that replication has near-zero cost. That shouldn't imply that one ought never attempt to contain it.
My wife applied for a position. The offer came through email in the form of a Word document that had been created using somebody else's offer as a template. The only thing she was concerned about was the bottom-line, so she viewed the attachment using strings(1), piping it to less(1), and visually grepping for the yearly figure...what she saw was absurdely low, and she wanted to tell them to 'get bent'. A cool head prevailed...she opened it in Word later and the offer was good -- but went back to examine the innards: there was somebody else's offer complete with their name and address, buried within the saved 'undo' stack. Knowing this, I don't understand how businesses could be so careless about sending Word documents outside the company.
The 'language agnostic' thing is nothing more than a universal virtual machine. What they're saying is that JVMs can only run Java, but their runtime will run C# and anything else. They're assuming that nobody has noticed that JVMs can run python, scheme, prolog, and jeeze what else?
There is no formal relationship between the TLD "org" and the type of legal entity known as a "non-profit organization".
P.S. Anybody with Model 4 stuff to sell, drop me a line.
"Easy" is a relative term. Anything is "easy" if you already know how to do it, and if you want to reduce the number of keystrokes involved you can write a script. In general, you just copy the file to the new target name, then do a cvs add of the new file, then rm the old file & do a cvs remove on it. Finally, a cvs commit and you're done. I agree it's not as easy as it could be, but it's not something that requires frobbing the CVSROOT control files directly.
One feature that I would like to see is the ability to configure the diff algorithm by file type, so that revisioned binary files (images, for instance) can be stored in the repository on a space-efficient manner. PRCS does this with the xdelta library, but I don't know if PRCS allows the use of a line-oriented-text diff algorithm for source files.
I also find the event hooks in CVS to be lacking. A generalized hook mechanism is needed that has access to both the name of the checked-in file and its contents, with both pre-commit and post-commit triggers, by file name, file type, directory, module, user, and whatever other parameter might be useful. CVS has some of this, but not enough.
Anyone who resorts to mucking-around in CVSROOT manually needs to read the FAQ.
"Tornado in a Junkyard: The Relentless Myth of Darwinism" (more info...) In this engaging page-turner, James Perloff devastates the theory of evolution. Issues tackled include the lack of transitional forms in the fossil record, the lack of evidence for "ape-men," and the amazing history of fraud that has characterized the efforts to get Darwin's theory, which he himself questioned, accepted and taught as fact. 321 pages. Paperback.