Wisdom for audio mixing 5.1 to 2 channels

[TCmullet]

I found this interesting post on another forum (2011) when Googling "convert 5.1 to stereo" (not in quotes):

http://www.debianuserforums.org/viewtopic.php?f=9&t=1264

It deals with the subject of converting 5.1 audio to 2, and from reading it (tho much is over my head), I gather that to do this properly is not a trivial matter and numerous freewares have taken shortcuts that do not produce the best result.  I see in the AviDemux audio filter section the options for remixing to various target formats:

Mono
Stereo
Stereo + surround
Stereo + center
Stereo + center + surround
Stereo front + stereo rear
5 channels
5.1
Dolby Pro Logic
Dolby Pro Logic II

I see in the AviDemux docs:

The mixer will change the number of channels that you hear. The most common usage is to go from 5.1 sound to stereo. You will probably want Dolby Pro Logic II downmixing. Beware that to be efficient, Dolby Pro Logic audio must not be encoded with a low bitrate (~(nbsp)<(nbsp)160(nbsp)kb/s).

This is a start, but could someone please:
1.  Explain why one would use each of the various options?  On the surface, one would think "stereo is what I want".  Then one sees "stereo + center" and thinks "Oh, the center will get left out if I don't pick this option?".  I get the impression that picking "stereo" takes only the left-front and right-front and DISCARDS the other 4 channels.  This is not what one would normally want.
2.  Report whether or not the concerns of the poster in that 2011 forum (the link I gave above) are addressed by the options in AviDemux, and how so.

I used to believe that all but the left and right were INCLUDED in the left and right, and separated by a decoder.  (This is memories from the 70s and 80s.)  Intuitively, if the channels are indeed discrete (there is no center, surround or bass signals matrixed into the L & R), then I would think that "Stereo + center + surrounds + bass" would be what puts ALL of it into a 2-ch. mix.  But this is where (I think) the 2011 guy is warning us that it can easily be implemented quite incorrectly and (according to him) IS done incorrectly often.

BTW, after a brief fling with Dolby-Surround in the 80s, I've leaned to getting everything into well-done 2-channel with a high-end audio system that "acts" 3-dimensionally.  But the world has changed.  Plus I no longer have the ability to HAVE more than simple stereo speakers for the moment.  (Economic hardship.)  Yet don't want to needlessly lose any audio content from the recordings.  I didn't know that my new Hauppauge Colossus capture card was going to bring Dolby 5.1 audio to me.  But now that it does, I have to deal with it.  (I *will* be keeping the 5.1 in some contexts, but have to remove it for others.)

[mean]

In dolby prologic mode you have center & rear "hidden" in the stereo channels
You can play it stereo, but if you have a prologic enabled system, it will recreate center and rear

[TCmullet]

That's great!

1.  Are you saying that BOTH Prologic formats do that (hide the center and rear into the front L & R)?

2.  Am I correct to infer from your brief doc that Prologic 2 is better than Prologic 1?

3.  That's great that center and both rears are encoded (hidden) into the stereo.  But what about the bass channel?  We certainly don't want to LOSE it.

I hope you'll answer all 3 Qs.