|
Thread Rules
1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution.
2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20)
3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible.
4. Use [code] tags to format code blocks. |
Can anybody help me regarding on RTSP streaming?
We have an android mobile app as client of our audio stream and use VLC as the server streaming the audio. We are using a wifi connection to connect them and when we were testing, our app disconnects when:
a) VLC is paused for more than 5 seconds
b) when we rapidly change the seek time (e.g. from 1:30 to 2:50 to 2:10...)
here's what we've tried to solve it so far
when the client disconnects, we try to let it reconnect immediately and then timeout when it attempts to reconnect for the 3rd time. unfortunately we feel this is too crude.
is there any other way to solve this?
we were using the university's wifi when we were testing this, could network congestion be a factor?
we also tried VLC to VLC streaming and tried to replicate the errors but we can't.
|
On March 20 2013 03:31 BisuDagger wrote:I'm a c++ programmer. Been using script for almost 2 years though. Now I have a huge job opportunity at my work but I need to learn linux. Any advice how the hell I learn linux, the only thing I know is I need to use VIM and Eclipse. But I need the best starters guide available! 
Here's a basic guide to learning Linux:
1. Download a live Ubuntu image.
2. Burn to CD, insert and boot from it.
3. Play around in the live session until you feel comfortable with the interface, then install it on a partition of your HD.
4. Boot it up and slowly work towards being able to do what you normally do.
5. Get stuck on a problem, search google or go on freenode IRC and ask people.
6. Do 5 over and over and over and over.
Linux, at least in my experience, is all about your patience when it comes to solving problems. Ubuntu and many other distros have come a long way and a lot works fine out of the box, but as soon as you start to go out of the bounds of "I just want to surf", you eventually run into issues which boggle the mind and when you look for answers, you'll be forced to learn how to configure things which is generally the big hurdle with linux. Don't dive in hoping to learn exactly how X window system works etc, it will take you ages, just solve problems as you find them and you'll get used to it.
If you want to take it to the next level, you can try building a linux distro from scratch, that's what I did with Gentoo once. You basically format your computer, boot up a very basic linux system, build gentoo from source, then work up to a fully functional distro one package at a time. Gives a lot of knowledge about how Linux works, but is quite a timeconsuming project.
|
Do any of you have any experience with regression testing?
What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here?
edit: f*** nvm
http://www.testingeducation.org/k04/RegressionExamples.htm
General functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.)
Also with regards to the comments earlier about how operating systems (and other) courses, give you the grey matter you need. I would argue that it's still significantly more beneficial to just teach students real world applications than it is to try to indirectly teach them the critical thinking skills they need. Because learning how to develop a website may be more beneficial than learning how to develop an operating system in the long run for the majority of programmers.
This is a dangerous idea though, I agree, as it leads to an idea known as competency based learning where everything is measured. So if a course's impact on a student's competency to do a job in the real world is not measureable, then the course is deemed worthless. I don't know if my definition of the idea is right, but competency based learning is being pushed forward and might be the new face of education.
My fear is that competency based learning may make students learn things that are too specific to graduation.
http://gettingsmart.com/cms/blog/2013/01/is-competency-based-learning-the-future-of-education/
Jargon-obsessed as we are, researching CBL online will probably find you phrases like ‘bundling’ and ‘concept learning model’. In an everyday idiom, however, it’s easy to sum up: CBL is the future.
In the new age of CBL, students earn their qualifications by proving their ‘competency’, proceeding at their own pace, not by spending a set number of hours in the lecture theatre. Writing about carving burrs into veneers, the theory goes, may only show that you can write about carving burrs into veneers. It doesn’t mean you can actually do it, and besides you can write something today and forget it tomorrow. It is an indirect measurement of your knowledge and abilities. The act of carpentry, or whatever your chosen vocation, requires a deeper level of understanding than a simple exam paper can prove. If you can perform a task in front of an assessor; however, then you can earn certification which proves you can actually carve/sail a ship/dance the tarantella.
|
Question: Is there a good book, online site or method to learning to use windows visual express 2012? I'm lost in the IDE (been trying to mess around with the new kinect sdk)
|
On March 21 2013 06:02 obesechicken13 wrote:Do any of you have any experience with regression testing? What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here? edit: f*** nvm http://www.testingeducation.org/k04/RegressionExamples.htmShow nested quote +General functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.)
Many companies have the following process: for every bug fix made to the product, a test is introduced into a test suite to ensure that the bug is never reintroduced. This test suite, made from prior bug fixes, is called a regression test suite.
|
We have so many flavors of testing it's hard to know which qualifies for which category.
I think of regression testing as "it used to work, does it still work after this change?" We have unit tests for almost everything that are supposed to be run after you make a change to a particular item, so even though it's not derived from a past bug per se, I'd still think of that as falling under regression testing.
|
On March 21 2013 07:56 Craton wrote:
We have so many flavors of testing it's hard to know which qualifies for which category.
I think of regression testing as "it used to work, does it still work after this change?" We have unit tests for almost everything that are supposed to be run after you make a change to a particular item, so even though it's not derived from a past bug per se, I'd still think of that as falling under regression testing.
We're still more in the feature design stage so we don't have that much automation yet. I'm not sure if we have unit tests that run regularly.
|
We are too. It's part of Scrum.
|
What makes C++ so famous for game development? You don't have to describe in details.
Thanks.
|
On March 21 2013 09:04 darkness wrote:What makes C++ so famous for game development? You don't have to describe in details. Thanks.
Optimization.
|
Czech Republic4646 Posts
On March 21 2013 06:02 obesechicken13 wrote:Do any of you have any experience with regression testing? What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here? edit: f*** nvm http://www.testingeducation.org/k04/RegressionExamples.htmShow nested quote +General functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.) Also with regards to the comments earlier about how operating systems (and other) courses, give you the grey matter you need. I would argue that it's still significantly more beneficial to just teach students real world applications than it is to try to indirectly teach them the critical thinking skills they need. Because learning how to develop a website may be more beneficial than learning how to develop an operating system in the long run for the majority of programmers. This is a dangerous idea though, I agree, as it leads to an idea known as competency based learning where everything is measured. So if a course's impact on a student's competency to do a job in the real world is not measureable, then the course is deemed worthless. I don't know if my definition of the idea is right, but competency based learning is being pushed forward and might be the new face of education. My fear is that competency based learning may make students learn things that are too specific to graduation. http://gettingsmart.com/cms/blog/2013/01/is-competency-based-learning-the-future-of-education/Show nested quote +Jargon-obsessed as we are, researching CBL online will probably find you phrases like ‘bundling’ and ‘concept learning model’. In an everyday idiom, however, it’s easy to sum up: CBL is the future.
In the new age of CBL, students earn their qualifications by proving their ‘competency’, proceeding at their own pace, not by spending a set number of hours in the lecture theatre. Writing about carving burrs into veneers, the theory goes, may only show that you can write about carving burrs into veneers. It doesn’t mean you can actually do it, and besides you can write something today and forget it tomorrow. It is an indirect measurement of your knowledge and abilities. The act of carpentry, or whatever your chosen vocation, requires a deeper level of understanding than a simple exam paper can prove. If you can perform a task in front of an assessor; however, then you can earn certification which proves you can actually carve/sail a ship/dance the tarantella.
CBL has its place, but as general approach to education is useless. The general trend in the western education is nonsensical (but maybe unavoidable). Making bigger and bigger percentage of population get higher education means in general that level of said higher education goes down as people are getting only marginally "smarter" (if at all). So programs are dumbed down and CBL (or rather bastardized versions of it) is a nice way to make even higher percentage of population pass the higher education.
It is getting to a point that being a carpenter requires college degree. Even for being average programmer college/uni degree is not necessary. That is also the reason for the notion that some people hold that uni CS courses are useless. That is because point of CS uni education is to educate computer scientists, not programmers. To be average programmer it would be enough to have specialized "vocational" schools on a level of high school. No higher education is necessary. Approach to programming should be the same as to engineering. You have your "technicians" that can easily be educated on high school level (with some follow up training and regular updates) and then you have the actual engineers that do the high level work and that require uni education and who direct the technicians and others to do what needs to be done. But software development is still too young of an profession and this specialization will take time and some more standardized techniques need to emerge so it can work.
As for the specific example. Developing a big project (website) should be part of the uni education. But in no way can it replace for example courses on complexity, discrete math or logic. Or even the ones on compilers or operating systems. Again that is because the goal of uni education is to create the "upper class" of programmers. The fact that every programmer nowadays is expected to have CS degree is just side-effect of PC-driven education systems we have nowadays.
|
Czech Republic4646 Posts
On March 21 2013 06:17 Kambing wrote:Show nested quote +On March 21 2013 06:02 obesechicken13 wrote:Do any of you have any experience with regression testing? What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here? edit: f*** nvm http://www.testingeducation.org/k04/RegressionExamples.htmGeneral functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.) Many companies have the following process: for every bug fix made to the product, a test is introduced into a test suite to ensure that the bug is never reintroduced. This test suite, made from prior bug fixes, is called a regression test suite.
That seems very narrow view of regression testing. Regression testing as we used it was supposed to make sure that new changes to the system do not introduce bugs in parts of the system that were not supposed to be affected by those changes. Basically they exist to discover side-effect bugs of new changes and also errors in analysis of the new changes.
In the case of project I worked on it was a set of tests that covered all critical functions of the system. Using just tests based on previous bugs does not seem to address the issue.
|
On March 21 2013 22:03 mcc wrote:Show nested quote +On March 21 2013 06:17 Kambing wrote:On March 21 2013 06:02 obesechicken13 wrote:Do any of you have any experience with regression testing? What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here? edit: f*** nvm http://www.testingeducation.org/k04/RegressionExamples.htmGeneral functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.) Many companies have the following process: for every bug fix made to the product, a test is introduced into a test suite to ensure that the bug is never reintroduced. This test suite, made from prior bug fixes, is called a regression test suite. That seems very narrow view of regression testing. Regression testing as we used it was supposed to make sure that new changes to the system do not introduce bugs in parts of the system that were not supposed to be affected by those changes. Basically they exist to discover side-effect bugs of new changes and also errors in analysis of the new changes. In the case of project I worked on it was a set of tests that covered all critical functions of the system. Using just tests based on previous bugs does not seem to address the issue.
That is a fairly common way of building a regression test library, though. It has a certain compelling logic to it, too: It guarantees you'll only ever have to fix a particular problem once, and it guarantees you'll invest test-writing time in components in direct proportion to their propensity to develop bugs.
|
On March 21 2013 09:04 darkness wrote:What makes C++ so famous for game development? You don't have to describe in details. Thanks.
I've heard tons of reasons, some of them make sense, others do not. The most classic argument is that higher level languages like C# are too slow. This argument is quite weak, careful benchmarking shows that the difference is quite small nowadays, wellwritten C++ code and wellwritten C# code will often perform very very similarly.
The best reason I've heard has to do with the fact that DirectX etc is quite hard to work with effectively from higher level languages. Other than that, C++ offers a lot of flexibility which can be nice for doing larger scale optimizations.
|
On March 22 2013 01:14 Tobberoth wrote:Show nested quote +On March 21 2013 09:04 darkness wrote:What makes C++ so famous for game development? You don't have to describe in details. Thanks. I've heard tons of reasons, some of them make sense, others do not. The most classic argument is that higher level languages like C# are too slow. This argument is quite weak, careful benchmarking shows that the difference is quite small nowadays, wellwritten C++ code and wellwritten C# code will often perform very very similarly. The best reason I've heard has to do with the fact that DirectX etc is quite hard to work with effectively from higher level languages. Other than that, C++ offers a lot of flexibility which can be nice for doing larger scale optimizations.
Two words ruin C#: garbage collection.
|
On March 22 2013 01:26 CecilSunkure wrote:Show nested quote +On March 22 2013 01:14 Tobberoth wrote:On March 21 2013 09:04 darkness wrote:What makes C++ so famous for game development? You don't have to describe in details. Thanks. I've heard tons of reasons, some of them make sense, others do not. The most classic argument is that higher level languages like C# are too slow. This argument is quite weak, careful benchmarking shows that the difference is quite small nowadays, wellwritten C++ code and wellwritten C# code will often perform very very similarly. The best reason I've heard has to do with the fact that DirectX etc is quite hard to work with effectively from higher level languages. Other than that, C++ offers a lot of flexibility which can be nice for doing larger scale optimizations. Two words ruin C#: garbage collection.
garbage collection(GC) with respect to bad performance or?
In my opinion, GC is still borderline ok; it frees lazy programmers from the burden of doing memory management, as long as they don't get hit with related bugs.
The reason you'd use C# is reflection and widgets for game developer tools.
|
On March 22 2013 01:53 ddengster wrote:Show nested quote +On March 22 2013 01:26 CecilSunkure wrote:On March 22 2013 01:14 Tobberoth wrote:On March 21 2013 09:04 darkness wrote:What makes C++ so famous for game development? You don't have to describe in details. Thanks. I've heard tons of reasons, some of them make sense, others do not. The most classic argument is that higher level languages like C# are too slow. This argument is quite weak, careful benchmarking shows that the difference is quite small nowadays, wellwritten C++ code and wellwritten C# code will often perform very very similarly. The best reason I've heard has to do with the fact that DirectX etc is quite hard to work with effectively from higher level languages. Other than that, C++ offers a lot of flexibility which can be nice for doing larger scale optimizations. Two words ruin C#: garbage collection. garbage collection(GC) with respect to bad performance or? In my opinion, GC is still borderline ok; it frees lazy programmers from the burden of doing memory management, as long as they don't get hit with related bugs. The reason you'd use C# is reflection and widgets for game developer tools.
More specifically, stop-the-world GC (can) ruin the soft real-time requirements of a video game (in general, any computationally intensive interactive simulation). Note that computers are powerful enough and runtime technology advanced enough this is only true of games that are pushing the boundaries of what current technology can afford. Beyond this point, you are making an engineering trade-off between the convenience of GC, the performance hit and behavior of GC on your program, and manpower optimizing your program around these constraints.
And yes, C# (and other higher-level languages) serve very well in the space of creating supporting tools for games development such as level editors, assets managers, etc.
|
Two more words ruin C#: dynamic binding
|
On March 22 2013 01:00 AmericanUmlaut wrote:Show nested quote +On March 21 2013 22:03 mcc wrote:On March 21 2013 06:17 Kambing wrote:On March 21 2013 06:02 obesechicken13 wrote:Do any of you have any experience with regression testing? What should it include? Wikipedia hints that regression testing may just be a small set of tests for things that have broken in the past. But what I've always done is I've run through every single feature in the test suite. What is the best practice here? edit: f*** nvm http://www.testingeducation.org/k04/RegressionExamples.htmGeneral functional regression: We retest the product broadly, including areas that worked before, to see whether more recent changes have destabilized working code. (This is the typical scope of automated regression testing.) Many companies have the following process: for every bug fix made to the product, a test is introduced into a test suite to ensure that the bug is never reintroduced. This test suite, made from prior bug fixes, is called a regression test suite. That seems very narrow view of regression testing. Regression testing as we used it was supposed to make sure that new changes to the system do not introduce bugs in parts of the system that were not supposed to be affected by those changes. Basically they exist to discover side-effect bugs of new changes and also errors in analysis of the new changes. In the case of project I worked on it was a set of tests that covered all critical functions of the system. Using just tests based on previous bugs does not seem to address the issue. That is a fairly common way of building a regression test library, though. It has a certain compelling logic to it, too: It guarantees you'll only ever have to fix a particular problem once, and it guarantees you'll invest test-writing time in components in direct proportion to their propensity to develop bugs.
Yes. It depends on what your definition of regression is.. Different shops will have different interpretations of what regressions are, which is fine. Where I have worked before, regressions are specifically re-introduced bugs (in the context of software that is being maintained rather than freshly developed). The sorts of tests you described, mcc, we would simply call functionality test suites.
The other benefit of what I described as a regression test suite is that the developer who writes it is typically forced to check in said test with his or her bug fix. At the time, they have the most insight in the bug, so they are one most capable of writing a comprehensive test for said issue.
|
On March 22 2013 02:44 aurum510 wrote:
Two more words ruin C#: dynamic binding
Eh, dynamic dispatch? Or type dynamic?
|
|
|
|
|
|
EPT
