I would really like to see explanation of the user hearable jitter.... it would have to be in tens of milliseconds range and how does that happen in reality?
It is a receiver issue here. Different receivers handle timing differently. The problem isn't when you get off by 1 bit, the problem is when you get the clocks mismatched so that it drops a few every second. Your receiver can get off and it can get to be audible. You need REALLY good speakers to do this though or other effects will dominate jitter by a lot. What is worse is that you also tend to need a low to mid end receiver (one that does not buffer and post process LPCM signals). That combination is doomed for failure regardless IMHO.
edit. Why doesn't that jitter occur with decoded picture that has much higher bandwidth requirements?
It does. It shows up in two forms. One is "snow" - which you will rarely see on any decent HD receiver. The other is missing pixels at the end of the line. Missing pixels will be colored either white or black. Remember though that the edges of the lines are fixed, so this would only appear at the edges of a screen. Most people's televisions have 1-5% overscan anyway so the affected areas would never be drawn.
Note that there is another "timing" problem with picture that is far more prevalent. That is when the television or receiver doesn't get the pull down right.
Even more to the point though, most recievers or televisions do some form of post processing - meaning they buffer incoming information and try to correct it for artifacts in the source. That tends to crush any artifacts caused by transmission.
edit2: How would jitter happen if client has nice buffer to fight against lag. all jitter would be null&void before actually occuring
It doesn't happen if a good buffer is available - hence the reason such buffers are useful in processing.
Just more in general, there are two main reasons people prefer bitstreaming:
1) Post Processing - this one is the big one. Post processing includes things like Pro Logic II processing. In other words, taking a 2.1 signal and turning into a 5.1 or a 5.1 and simulating a 7.1. Several receivers that people would consider mid to high end will not post process LPCM sound. For example, the Onkyo 605s will not. The Onkyo 705s will. Personally, I use a Denon that does post process LPCM sound. So in my case, there is no advantage to have the sound bit streamed. Note that most of the receivers that don't are older. However, when you spend $2000-$5000 on a receiver you expect it to last 5-10 years. I know people who bought high end receivers only a few years ago that will not post process LPCM. They look for bitstreamed solutions for obvious reasons.
2) Configuration/quality issues. Jitter falls into this. This is really a very setup specific category. Some people have very expensive pre/pros and amps. They prefer their processing to happen in their receiver to make sure they get the most out of their equipment. The more components you add into the mix, the more likely small differences are going to matter. Jitter is the most often cited example, but there are others. Out of the two, I think most people in this category know very well they are in this category and don't generally see the difference as positive or negative. They just choose what is best for their own equipment.
I think the whole topic is really overdone though. If you are really into AV you choose what is best for your setup. I have no idea why people got hung up on the details in the case of the PS3.
) Not sure if firmware 2.20 will decode & output all 7.1 channels, or start with 5.1 channel output. (Wait and see. 
) For what it counts movies natively encoded in 7.1 LPCM, already output perfectly fine (as mentioned in forums at some unmentioned bitrate). So if the firmware/decode works well then 7.1 seems reasonably possible.
