I don't think "steep learning curve" means what you think it means. If it's steep, it's easy to learn. Did you really meant that heavy-duty-enterprisey Java is easy to learn?
Checking an object every time it's dereferenced to make sure it hasn't been freed or worse, re-used is another - man that's a stupid problem to solve
What are you talking about? This applies neither to C++ nor to Java. Heck, there is in fact no way of checking for either directly, whether in C++ or in Java. You can hack around it in C++ by adding code to the destructor that will mark the object as "freed" in some global dictionary, but what's the point of doing that I wouldn't know. If you need such code, you're doing it wrong.
In Java it's even simpler: if you can access the object, it's there, or you've hit a garbage collector bug or a bit flip in your hardware. There's basically no way of freeing an object. You can reset the reference, but that's it. It's same as in modern C++: if you use std::weak_ptr, you may get a nullptr and you must check for it. If you use a std::shared_ptr and it's non-nullptr, the object is there, and it won't magically become nullptr unless you reset it.
Employing one of the three top consulting houses was probably what sunk it. They are big bureaucracies that mirror the government, really. Just think what those consulting houses do: they help clients who are clueless enough not to be able to do the work themselves. If the client can't do the work, you think they'll know enough to gauge consultant's performance? Nope.
It's really, really simple, in a way. The rules and their interactions form a basis of a more complex set of behaviors that emerge when you start executing them.
It's like with ants: a single ant is pretty dumb. But put them together, and you've got beautiful emergent behavior. In case of bureaucracy, you've got a bunch of "intelligent", "well meaning" simple rules. Put them together, and the emergent behavior is a pile of crap. It's like why IP internetworking won over X.25.
Both the behavior of ants and the behavior of the bureaucracies "defies reason". We're quite literally at the edge of human understanding. All that we know is that the less rules, the better it will be. Yes, so some excesses will slip through, but oh well. Right now it's all a big well meaning excess. The way government contracts are done is really just as if you paid the workers 10%, and the rest got paid to crack whores.
Can't you just buy a chain saw, a climbing harness, some rope and slowly but methodically take them down? It really isn't that hard. Been there, done that.
Yeah. I have a Sony TV and a Sony Blu-Ray player - both less than 2 years old. The crypto negotiation takes about a second, with blank screen and audible pops. On most Blu-Ray discs it happens at least twice before you get to playing the movie. With DVDs it sometimes takes place 4 times. I swear that an old CRT TV and a VCR were faster to cold-boot to a visible, playing movie, with inclusion of loading the tape, than the current generation of HD gear. It says something when a system that could, theoretically, be up and playing in 5 seconds from power-up is almost a factor of magnitude away from what the hardware allows it to do. It really takes the cake when such a system is about as "fast" as an electromechanical variant. Yeah, VCRs are nowhere near the quality of even DVDs, but still.
The stories are nothing but conversation starters. They are usually so badly edited so as to be useless for anything more than that anyway. It takes a truly amazing amount of presbyopia on the part of Dice to presume anything else. Nobody cares much about your lousy stories, slashdot.
I'm sorry to disappoint you, but who the fuck has the time to analyze the source code of a news site they simply read? Sure, if you're in web design of one sort or another you'd want to learn from the others by reverse-engineering, but that's not what any significant fraction of users is doing. I'm pretty sure slashdot could randomly insert links to bank accounts, with passwords, into the html comments, and it'd be missed for a long time.
You're mischaracterizing the issue. Narrow columns work just fine - get any paperback book and see how many characters there are per line. Compare that to the current layout on a modern wide screen. Long lines are unreadable. That doesn't mean that there should be just one column of content, though. Someone with brains can surely come up with a way of laying out a discussion in multiple columns. Sure it may not be a trivial algorithm, but that's, supposedly, what the professionals are for. Why can't slashdot innovate? Why can they only copy all the bad things others do?
Why won't you hire actual, you know, people who do usability professionally, and have them tell the designer folks what should be the constraints for design? Unfortunately, most designers demonstrably suck in the usability department. They simply don't have the training nor understanding necessary to do it right, in spite of the correct examples glaring at them from stuff that has been around for very, very long before they were born...
Widening it is actually the wrong thing to do. There are two orthogonal issues: 1. Use of the screen real estate. 2. Keeping the line lengths reasonable. The solution has been found a couple hundred years ago, just look in a newspaper. Slashdot really needs a multi-column layout.
The bigger problem is that the widescreens make a lot of web content unreadable. Long lines are horribly hard to read. If an entire paragraph takes two lines of a couple hundred characters each, filling a wide screen, then that's a problem.
Slashdot needs to switch to two-column mode on wide screens. There's a reason why books are formatted for pages that are longer in vertical direction than horizontal.
The commercial interests, big commercial interests are negatively affected by this spying. It's going to hit some bottom lines big time pretty soon. If we're to believe in the strong arm of lobbying, domestic spying should end any day now, right? Riiight:)
Where the heck did you get the idea that I think git stores diffs? All I'm saying is that Git's merge support is useless for binary files, because it doesn't know anything about them. It'd be great if I could have git show visual differences in STEP files, but it doesn't do that, so merging is at the file-level, not content-level, so it becomes no better than what subversion has.
Now pray tell how much will I pay and how long will I wait when I want to check out 20 megabytes worth of drawings at customer's site, running through a 3G wireless modem - a common scenario for me. The repo is a couple gigs large in the current revision, and has the overall size of about 100G.
git is not the end-all, be-all. Where you don't need massive distribution or you work with binary files (say, a CAD repository), git and mercurial just get in the way. They still doesn't have sparse/partial clones where you could just download selected subdirectories. That's the make-or-break feature of SVN in some of my projects. Our CAD repository is over a gigabyte. When I work on a project, I only need our parts library and the project folder. git and mercurial are great for software, but software is not the only thing version control systems are used for.
Good luck trying to get a chargeback on your card if you use it for a donation. When the word "donation" is heard by the rep, they'll end the conversation right there. Donations with credit cards are like Western Union money transfers. Once you pay, the money is gone.
Note that a kickstarter contribution is not a donation, it's a payment for service, and you can certainly ask your credit card company for a chargeback if your backed project doesn't deliver.
Slashdot Mirror
that style is too hard to discover
I'd say simply by going for the horse's mouth, as in reading Strostrup's C++ Programming Language book, will get you there 90% of the way.
I don't think "steep learning curve" means what you think it means. If it's steep, it's easy to learn. Did you really meant that heavy-duty-enterprisey Java is easy to learn?
Checking an object every time it's dereferenced to make sure it hasn't been freed or worse, re-used is another - man that's a stupid problem to solve
What are you talking about? This applies neither to C++ nor to Java. Heck, there is in fact no way of checking for either directly, whether in C++ or in Java. You can hack around it in C++ by adding code to the destructor that will mark the object as "freed" in some global dictionary, but what's the point of doing that I wouldn't know. If you need such code, you're doing it wrong.
In Java it's even simpler: if you can access the object, it's there, or you've hit a garbage collector bug or a bit flip in your hardware. There's basically no way of freeing an object. You can reset the reference, but that's it. It's same as in modern C++: if you use std::weak_ptr, you may get a nullptr and you must check for it. If you use a std::shared_ptr and it's non-nullptr, the object is there, and it won't magically become nullptr unless you reset it.
I'm pretty damn sure for another $50M you could license a tech transfer from Amazon and have your own fucking AWS, in your own datacenter.
Employing one of the three top consulting houses was probably what sunk it. They are big bureaucracies that mirror the government, really. Just think what those consulting houses do: they help clients who are clueless enough not to be able to do the work themselves. If the client can't do the work, you think they'll know enough to gauge consultant's performance? Nope.
LOL, you have no idea what it takes to scale something like facebook. Compared to it, healthcare.gov could run on one high-end server for all I care.
It's really, really simple, in a way. The rules and their interactions form a basis of a more complex set of behaviors that emerge when you start executing them.
It's like with ants: a single ant is pretty dumb. But put them together, and you've got beautiful emergent behavior. In case of bureaucracy, you've got a bunch of "intelligent", "well meaning" simple rules. Put them together, and the emergent behavior is a pile of crap. It's like why IP internetworking won over X.25.
Both the behavior of ants and the behavior of the bureaucracies "defies reason". We're quite literally at the edge of human understanding. All that we know is that the less rules, the better it will be. Yes, so some excesses will slip through, but oh well. Right now it's all a big well meaning excess. The way government contracts are done is really just as if you paid the workers 10%, and the rest got paid to crack whores.
Yeah, I can't get over the fact that many Blu-Ray discs have fucking progress bars before the menu shows up. It's like YouTube, just local.
Can't you just buy a chain saw, a climbing harness, some rope and slowly but methodically take them down? It really isn't that hard. Been there, done that.
Yeah. I have a Sony TV and a Sony Blu-Ray player - both less than 2 years old. The crypto negotiation takes about a second, with blank screen and audible pops. On most Blu-Ray discs it happens at least twice before you get to playing the movie. With DVDs it sometimes takes place 4 times. I swear that an old CRT TV and a VCR were faster to cold-boot to a visible, playing movie, with inclusion of loading the tape, than the current generation of HD gear. It says something when a system that could, theoretically, be up and playing in 5 seconds from power-up is almost a factor of magnitude away from what the hardware allows it to do. It really takes the cake when such a system is about as "fast" as an electromechanical variant. Yeah, VCRs are nowhere near the quality of even DVDs, but still.
I'm analyzing plenty of other code, thank you, slashdot's is very, very low on my list. It might have even fallen off the list :)
Maybe they just don't fucking read books. It may be that simple :(
The stories are nothing but conversation starters. They are usually so badly edited so as to be useless for anything more than that anyway. It takes a truly amazing amount of presbyopia on the part of Dice to presume anything else. Nobody cares much about your lousy stories, slashdot.
I'm sorry to disappoint you, but who the fuck has the time to analyze the source code of a news site they simply read? Sure, if you're in web design of one sort or another you'd want to learn from the others by reverse-engineering, but that's not what any significant fraction of users is doing. I'm pretty sure slashdot could randomly insert links to bank accounts, with passwords, into the html comments, and it'd be missed for a long time.
You're mischaracterizing the issue. Narrow columns work just fine - get any paperback book and see how many characters there are per line. Compare that to the current layout on a modern wide screen. Long lines are unreadable. That doesn't mean that there should be just one column of content, though. Someone with brains can surely come up with a way of laying out a discussion in multiple columns. Sure it may not be a trivial algorithm, but that's, supposedly, what the professionals are for. Why can't slashdot innovate? Why can they only copy all the bad things others do?
Why won't you hire actual, you know, people who do usability professionally, and have them tell the designer folks what should be the constraints for design? Unfortunately, most designers demonstrably suck in the usability department. They simply don't have the training nor understanding necessary to do it right, in spite of the correct examples glaring at them from stuff that has been around for very, very long before they were born...
Widening it is actually the wrong thing to do. There are two orthogonal issues: 1. Use of the screen real estate. 2. Keeping the line lengths reasonable. The solution has been found a couple hundred years ago, just look in a newspaper. Slashdot really needs a multi-column layout.
The bigger problem is that the widescreens make a lot of web content unreadable. Long lines are horribly hard to read. If an entire paragraph takes two lines of a couple hundred characters each, filling a wide screen, then that's a problem.
Slashdot needs to switch to two-column mode on wide screens. There's a reason why books are formatted for pages that are longer in vertical direction than horizontal.
The commercial interests, big commercial interests are negatively affected by this spying. It's going to hit some bottom lines big time pretty soon. If we're to believe in the strong arm of lobbying, domestic spying should end any day now, right? Riiight :)
As flayzernax said: swap them for some prostitutes. Keeps the number incarcerated locally constant :)
Let me know when there's a way to pull a few selected files out of a git repo, change them, and then push them back. Until then, shut up, mmkay?
Where the heck did you get the idea that I think git stores diffs? All I'm saying is that Git's merge support is useless for binary files, because it doesn't know anything about them. It'd be great if I could have git show visual differences in STEP files, but it doesn't do that, so merging is at the file-level, not content-level, so it becomes no better than what subversion has.
Now pray tell how much will I pay and how long will I wait when I want to check out 20 megabytes worth of drawings at customer's site, running through a 3G wireless modem - a common scenario for me. The repo is a couple gigs large in the current revision, and has the overall size of about 100G.
git is not the end-all, be-all. Where you don't need massive distribution or you work with binary files (say, a CAD repository), git and mercurial just get in the way. They still doesn't have sparse/partial clones where you could just download selected subdirectories. That's the make-or-break feature of SVN in some of my projects. Our CAD repository is over a gigabyte. When I work on a project, I only need our parts library and the project folder. git and mercurial are great for software, but software is not the only thing version control systems are used for.
Good luck trying to get a chargeback on your card if you use it for a donation. When the word "donation" is heard by the rep, they'll end the conversation right there. Donations with credit cards are like Western Union money transfers. Once you pay, the money is gone.
Note that a kickstarter contribution is not a donation, it's a payment for service, and you can certainly ask your credit card company for a chargeback if your backed project doesn't deliver.
And what's so different between gpg in javascript vs. gpg in machine code?