[alok-g]

This is very helpful. I'll look at the suggestions.

>> They are unstable because they require huge amount of resources ...

This actually is not enough of a reason in theory for the programs to crash. The programs should stop the processing when they run out of resources and allow a graceful exit.

The root cause for crashes then, I am guessing, must be lousy code. Either the entire video editing industry has gotten used to it, so that they are not demanding better solutions and thereby the product companies are not prioritizing fixing these crashes, OR, there are some common libraries that they all using which crash, and these product companies feel helpless. Hopefully someone knows the actual answer! :-)

[blattimwind]

> The root cause for crashes then, I am guessing, must be lousy code.

Huh? Almost every application just panics when it runs out of memory. Not least because it's actually difficult to do stuff without using memory within typical application frameworks.

[alok-g]

At least memory-intensive applications should be tested under such conditions.

When I was at college more than a decade back, I tested many applications under such conditions and my own application too. Most Microsoft's applications survived even with less than 100 bytes of RAM remaining and a disabled virtual memory! This included Microsoft Office, Visual Studio, et al. My own application survived too with a fix or two. Windows Explorer, funnily, crashed. Many non-Microsoft applications crashed.

More recently (2012), I tested two commercial Java profiler applications, JProfiler and Yourkit. They both survived without issues under such conditions.

[baldfat]

I would seriously blame codecs. The programming from video editors I am sure is 100% about compromises. You can't get a smooth preview playback of 4K unless you have sunk thousands of dollars into your computer. I am guessing that they are always playing the compromise of you can't really get this but we will give you that all the time. The issue comes when they are playing around with how to control latency and i/o bottle necks.