In regards to tris in an organic mesh - 

Tris have a bad tendency to create artifacts in renders. They can also cause problems with morphs and joint bends depending on their placement. On the oposite side of the spectrum they can also be helpful, at times, with joint bends and morphs, but there are better ways of going about it than using tris.

This doesn't mean it will always happen. You might render a tri'd up figure 1,000 times and never notice it. Depending on your lighting and various material settings, a tri in the wrong place can cause headaches in various situations and is considered unprofessional, "lazy" modeling, unless you're modeling low-res game characters which use very different lighting setups and very basic stick-figure like rigging. 

So its not so much that tris are "bad", it's just that they serve their purpose in certain, specific situations - but are mostly reserved for less organic shapes.

Working with hard surfaces, doesn't really matter how many tris you use, since concrete wood and steel aren't intended to be very flexible. Their shapes always stay the same, unlike organics. 

Cloth/clothing is a different scenario. If it's conforming cloth, then it should usually be treated as an organic. If it's dynamic cloth, then tris can actually be beneficial in that regard as they can allow for more fluid draping, wrinkles, etc. But that is determined by the cloth dynamics engine being used and how it handles quads vs tris. 

The "lazy" label is due to it generally being much more challenging to create a human model without tris, but that requires a good understanding of topology and edgeflow and how different types of geometry work with each other. 

Modeling a human in all quads also helps when working with SubDs. Quads divide into 4s much more neatly than tris do, thus less artifacting. At high res levels too many tris can create "pinching" in the mesh, which can't really be smoothed out much no matter what you do. They just have to be post-worked. You're going to get some pinching anyway because all diamonds reduce to tris. So all diamonds will create pinching, but they're necessary and unavoidable as they direct edgeflow. The goal is to reduce as much as possible, and that's done by reworking the mesh over and over until you get it to an acceptable level. 

As for animated joint centers being built in, they would allow for more realistic joint bends - pretty much why they exist - in both the base figure and any FBMs you might be trying to use. 

Humans aren't hard-bodied robots. Joints do more than bend left, right, up, down, and twist. Multiple systems are at work there, and a joint can move a lot of different ways depending on the position you're aiming for. 

Weightmapping is not the end-all be-all of rigging, especially poser's weightmapping. It's just one tool in the shed. Whats the point in having a lawn mower if there's no feul for it? 

Poser's weightmapping does help, but it still can't create the most believable joint bends by itself, and seems to work best on a per-figure/character basis. Relying on weightmapping alone - or primarily - is actually more limiting than it was before, because now a lot of character morphs, if they are too much different than the base figure - require new weightmapping in order for the morph to really look and feel right when the figure is posed. Animated centers go a long way in alieviating that. Of course they have to be adjusted depending on the morphs that are applied, but that is one of the drawbacks - from a content creation stand-point - of having more advanced features that allow for more realistic figures. Unfortunately it also makes creating add-on content more of a job if you want it to look and function its best, but PP2014 has gone a long way in making content creation less of a hassle anyway.

There are some other points I was going to make, about SubD's and some other things, but too tired right now. Will address it later.

 

~Shane