You're perfectly valid in questioning whether a user interface task should be throttled to counteract the ability to be about to dangerously do too many things too quickly. This is common with electrical systems where you need to de-bounce a button such that a switch state can only be changed if the change in state lasts longer than a few microseconds and that you can't repeat a button press for a different value of other few microseconds.
In analog systems, this filtering is called hysteresis. You're filtering out the input noise so that intermittent transient changes in input value are filtered out if they do not exist long enough to change the hysteresis system to a different state. Much like many other electrical engineering principles, this can be applied to digital systems very easily through timers and event states and there doesn't exist an actual analog component if the system is a UI.
A serial over USB or some other USB device class interface to a host computer running in the microcontroller firmware. You're also not limited to just a single microcontroller and can chain them together through various bus interfaces and interrupt event signalling to get better concurrency from the system on a hardware level.
Aside, there's nothing at all wrong with using larger systems to prototype out an idea before committing to the additional engineering time necessary to trim something down for a lower cost before going into mass production. You'll never now if you want to scale something out or keep it as a one off oddity. The part you already have on your workbench is -ALWAYS- better than the one you have to go out to buy if you're hashing out wild ideas.
If the video or audio streams are transmitted in XMPP messages (which they most likely are), then they are. XMPP runs over a TLS connection with X.509 certificates that Google has picked up from their own private certificate authority that is a part of the default trust store of most major browsers and TLS encryption library implementations for a while. What they do with it is inside the black box, and the rest is a question of your trust for them as a corporate entity.
Re:Still no support for TLS 1.1 / 1.2
on
Firefox 21 Arrives
·
· Score: 1
Firefox gets its SSL/TLS support from the underlying NSS library (another Mozilla product), and TLS 1.1 support was added there in NSS 3.14 released on Dec 18, 2012, and NSS-3.14 is the most current version on my system. While the browse dev team needs to pick up the pace on their side of implementing support for it, it is now present in the crypto library.
You confuse the sanctity of education with its utility. You'll probably never find a PhD touching a line of code to do with that feature set. They'll be too busy churning out whitepapers like crazy about it. It'll wind up being the guys with certificates and lesser degrees actually making the clever hacks..
Doesn't really matter what you or I think. It only matters what the Vatican and Italy think, and Italy only matters because the Vatican is functionally an Italian state even if they are allowed to claim sovereignty. Plus, if anyone wants to wage war with the Vatican, they have to go through Italy first.
But you can only get that efficiency the first time you write that line of perl. The poor schmuck that comes after you then has to spend ten times as long maintaining it by trying to figure out your obfuscated perl contest submission.
KDE technically isn't C++. It's written in Qt, which is a set of bastardized extensions to C++ (see meta object compiler) that produce generated C++ code. It's unfortunate that Nokia (who bought Trolltech way back when) invested so much time and effort into something the is effectively obsoleted by Boost.
We're talking about computer science majors, not physics majors.
Yes, I agree that you should hire computer scientists to do R&D, not routine system administration. However, you should always develop software with the greatest understanding of your users and the systems that your software will integrate with. In a majority of cases, your users are going to be either system adminstrators or package maintainers just as much as end users. Someone has to deploy your software to end users, and that's usually the sysadmins in the end user's IT department.
If your software is held together with duct tape and bubble gum, barely builds, doesn't follow common system conventions, or tries to reinvent half of unix, your programmers don't know enough about basic system administration to be useful. If you're lucky, there isn't a competing project that satisfies the same niche and the world at large just has to live with your incompetence. If you're not lucky, there's another company or group that's writing software to fill the same niche, and they did hire the computer scientist that knows about system administration, so they have a robust build system, clean RPMs or DEBs or MSIs or whatever else package format du jour is in vogue. Their software is going to be what your potential customers' IT staff recommends over yours.
My point is that someone that considers themselves a great scientist is only useful if they understand enough about what's out there to build on top of it. No one needs to have the wheel reinvented ten times over by ten different 'geniuses' that all claim that the standard implementation of the wheel is insufficient, broken, flawed, or not useful since they didn't cut the stone by hand themselves. To claim that you can improve the state of the art without understanding the state of the art is nothing more than pure hubris.
Then they get a sad reminder in 3-6 months what happens when you fire the people that were meeting your deadlines. I've seen entire departments and sometimes companies fold after they tick their engineers off enough to go find new jobs after their staff gets below the threshold of enough manpower to complete the tasks that sales and marketing are busy promising to the customers.
I'm not totally sure, but I sure as hell wish that class was a required course for anyone with a CS degree. I fail to understand how so many people can claim to be geniuses at writing software and be so brain dead when it comes to understanding how production grade software is actually put together and used in the real world (ie, not the duct tape and bubble gum projects they put together for their senior thesis).
The Earth is not dark matter. Even discounting the wide array of man made radiation that our species puts out, the Earth also puts out a considerable amount of detectable radiation. It's just an insignificant amount in comparison to the sun.
I wish universities taught more of the fundamentals of IT and network administration. You have no idea how many boy-wonder-junior-vbscript-hackers I've have to deal with that don't even understand the basics of IP networking, but they all claim to be able to code entire systems to replace windows or unix or the internet from scratch in a matter of hours.
A university Computer Science degree should be an addendum to basic IT skills, not a replacement for them. As such, that means that it should be the burden of the universities to teach their students the basics in addition to the higher level topics.
Joint Architecture for Unmanned Systems (JAUS)
It originates in the DARPA research ecosystem, so it's got a lot of alphabet soup going on. I'm not sure if this is what the commercial sector will eventually converge upon as a standard, but what they do converge on will resemble that to some degree atleast.
That's not how peering works. You don't host boxes inside someone's network. You host BGP routers in the same carrier hotel as the other company and traverse those links to get into their network instead of paying for metered access through a leased line of some sort.
http://en.wikipedia.org/wiki/Wheel_(Unix_term)
Because it's been that way since TENEX (circa late 1960s), and a majority of sysadmins are crusty curmudgeons that hate change. It's just the laziest (and least time consuming) way to give users admin privileges without sharing the root password is add them to the wheel (or admin) group instead of re-writing/etc/sudoers everytime they want to give someone access to that one special command that requires UID 0 privs to run right.
Go in to the linux box you supposedly own and type 'man 5 sudoers'.
You can specify which programs a user can run as root with fine grained controls that only allow specific users or specific groups to run only specific programs. It is not an all or nothing tool even if a majority of lazy sysadmins use it that way. If it didn't have that level control, there would be no point in using it over/bin/su.
Grover Norquist's quote states that he wants it to be small enough to drown in a bathtub. Not sure where fitting in a bedroom came in, but we all played telephone in grammar school.
Slashdot Mirror
You're perfectly valid in questioning whether a user interface task should be throttled to counteract the ability to be about to dangerously do too many things too quickly. This is common with electrical systems where you need to de-bounce a button such that a switch state can only be changed if the change in state lasts longer than a few microseconds and that you can't repeat a button press for a different value of other few microseconds. In analog systems, this filtering is called hysteresis. You're filtering out the input noise so that intermittent transient changes in input value are filtered out if they do not exist long enough to change the hysteresis system to a different state. Much like many other electrical engineering principles, this can be applied to digital systems very easily through timers and event states and there doesn't exist an actual analog component if the system is a UI.
A serial over USB or some other USB device class interface to a host computer running in the microcontroller firmware. You're also not limited to just a single microcontroller and can chain them together through various bus interfaces and interrupt event signalling to get better concurrency from the system on a hardware level. Aside, there's nothing at all wrong with using larger systems to prototype out an idea before committing to the additional engineering time necessary to trim something down for a lower cost before going into mass production. You'll never now if you want to scale something out or keep it as a one off oddity. The part you already have on your workbench is -ALWAYS- better than the one you have to go out to buy if you're hashing out wild ideas.
If the video or audio streams are transmitted in XMPP messages (which they most likely are), then they are. XMPP runs over a TLS connection with X.509 certificates that Google has picked up from their own private certificate authority that is a part of the default trust store of most major browsers and TLS encryption library implementations for a while. What they do with it is inside the black box, and the rest is a question of your trust for them as a corporate entity.
Firefox gets its SSL/TLS support from the underlying NSS library (another Mozilla product), and TLS 1.1 support was added there in NSS 3.14 released on Dec 18, 2012, and NSS-3.14 is the most current version on my system. While the browse dev team needs to pick up the pace on their side of implementing support for it, it is now present in the crypto library.
ctags and cscope are quality development tools and integrate well with emacs and vim.
You confuse the sanctity of education with its utility. You'll probably never find a PhD touching a line of code to do with that feature set. They'll be too busy churning out whitepapers like crazy about it. It'll wind up being the guys with certificates and lesser degrees actually making the clever hacks..
DoD skunkworks concept whitepaper != functional implementation.
Anyone that has worked in any proximity to government contracting knows how much the marketing materials embellish.
I'm fine with paying someone three times the salary of an individual from the group of 140 that they keep in line.
My favorite way of getting a free C/C++ compiler is to wipe the box, install linux, and then install gcc.
That's what software is for.
Doesn't really matter what you or I think. It only matters what the Vatican and Italy think, and Italy only matters because the Vatican is functionally an Italian state even if they are allowed to claim sovereignty. Plus, if anyone wants to wage war with the Vatican, they have to go through Italy first.
But you can only get that efficiency the first time you write that line of perl. The poor schmuck that comes after you then has to spend ten times as long maintaining it by trying to figure out your obfuscated perl contest submission.
KDE technically isn't C++. It's written in Qt, which is a set of bastardized extensions to C++ (see meta object compiler) that produce generated C++ code. It's unfortunate that Nokia (who bought Trolltech way back when) invested so much time and effort into something the is effectively obsoleted by Boost.
We're talking about computer science majors, not physics majors. Yes, I agree that you should hire computer scientists to do R&D, not routine system administration. However, you should always develop software with the greatest understanding of your users and the systems that your software will integrate with. In a majority of cases, your users are going to be either system adminstrators or package maintainers just as much as end users. Someone has to deploy your software to end users, and that's usually the sysadmins in the end user's IT department.
If your software is held together with duct tape and bubble gum, barely builds, doesn't follow common system conventions, or tries to reinvent half of unix, your programmers don't know enough about basic system administration to be useful. If you're lucky, there isn't a competing project that satisfies the same niche and the world at large just has to live with your incompetence. If you're not lucky, there's another company or group that's writing software to fill the same niche, and they did hire the computer scientist that knows about system administration, so they have a robust build system, clean RPMs or DEBs or MSIs or whatever else package format du jour is in vogue. Their software is going to be what your potential customers' IT staff recommends over yours.
My point is that someone that considers themselves a great scientist is only useful if they understand enough about what's out there to build on top of it. No one needs to have the wheel reinvented ten times over by ten different 'geniuses' that all claim that the standard implementation of the wheel is insufficient, broken, flawed, or not useful since they didn't cut the stone by hand themselves. To claim that you can improve the state of the art without understanding the state of the art is nothing more than pure hubris.
Then they get a sad reminder in 3-6 months what happens when you fire the people that were meeting your deadlines. I've seen entire departments and sometimes companies fold after they tick their engineers off enough to go find new jobs after their staff gets below the threshold of enough manpower to complete the tasks that sales and marketing are busy promising to the customers.
I'm not totally sure, but I sure as hell wish that class was a required course for anyone with a CS degree. I fail to understand how so many people can claim to be geniuses at writing software and be so brain dead when it comes to understanding how production grade software is actually put together and used in the real world (ie, not the duct tape and bubble gum projects they put together for their senior thesis).
The Earth is not dark matter. Even discounting the wide array of man made radiation that our species puts out, the Earth also puts out a considerable amount of detectable radiation. It's just an insignificant amount in comparison to the sun.
I wish universities taught more of the fundamentals of IT and network administration. You have no idea how many boy-wonder-junior-vbscript-hackers I've have to deal with that don't even understand the basics of IP networking, but they all claim to be able to code entire systems to replace windows or unix or the internet from scratch in a matter of hours. A university Computer Science degree should be an addendum to basic IT skills, not a replacement for them. As such, that means that it should be the burden of the universities to teach their students the basics in addition to the higher level topics.
I think they learned their lesson with the whole GAIM/Pidgin fiasco.
Joint Architecture for Unmanned Systems (JAUS) It originates in the DARPA research ecosystem, so it's got a lot of alphabet soup going on. I'm not sure if this is what the commercial sector will eventually converge upon as a standard, but what they do converge on will resemble that to some degree atleast.
That's not how peering works. You don't host boxes inside someone's network. You host BGP routers in the same carrier hotel as the other company and traverse those links to get into their network instead of paying for metered access through a leased line of some sort.
but, but, but, THAT'S TOO HARD when passing the buck is so much easier!
http://en.wikipedia.org/wiki/Wheel_(Unix_term) /etc/sudoers everytime they want to give someone access to that one special command that requires UID 0 privs to run right.
Because it's been that way since TENEX (circa late 1960s), and a majority of sysadmins are crusty curmudgeons that hate change. It's just the laziest (and least time consuming) way to give users admin privileges without sharing the root password is add them to the wheel (or admin) group instead of re-writing
Go in to the linux box you supposedly own and type 'man 5 sudoers'. You can specify which programs a user can run as root with fine grained controls that only allow specific users or specific groups to run only specific programs. It is not an all or nothing tool even if a majority of lazy sysadmins use it that way. If it didn't have that level control, there would be no point in using it over /bin/su.
Grover Norquist's quote states that he wants it to be small enough to drown in a bathtub. Not sure where fitting in a bedroom came in, but we all played telephone in grammar school.