On mapping strategies:

Poser scenes can get quite heavy. One prime consideration you have to make, regarding maps, is their file size. Poser will load up as much as you can give it in terms of memory weight. But, it will start randomly dumping stuff once it gets overloaded. Geometry "can" have a big footprint, especially if you're using several characters in a scene with 50k+ polys a piece. But, adding on several heavyweight texture maps, each unique for each character, can put you over a decent memory budge very, very, quickly. Especially with 4kx4k eyeball maps... (Just kidding, kinda.)

The base figure cr2 for V4 is about 20 meg, with nothing but a simple map and deformers. Some of these bits are just instructions for Poser, so they effectively do more than their weight suggests. But, adding on morphs can dump a great deal into a figure in a scene. For instance, I have a dev figure that I play with with lots of custom morphs. It doesn't even have everything I could put into it and it's 200+ meg. That's pretty hefty, but if pile on a good hi-res texture set in jpg format, which is a compressed format, those textures are going to take up an extra 20+meg, alone, which is 10% of that very hefty figure's weight. If you're going for the lightest figures you can get, since there may be many in a scene, you're still looking at 20+ meg for each, unless you're doubling up by using the same texture sets for multiple figures (Which, I believe, Poser handles by instancing on the maps, not loading up multples of each set.) So, the textures you're applying could weigh more than the actual figures, even with compressed file types.

(Note: Unless you're using natively uncompressed images, you're not actually going to "gain" resolution and information by uncompressing a natively compressed image format... ie: Taking a jpg and saving it as a tiff is not going to "reveal" secret, hidden, image information which gives you some sort of visual advantage, since the original never had that information, anyway. AFAIK, Poser treats all image data as just plain "data", with certain exceptions for version and certain file-type specs. But, these can be handled just as easily with a jpg as a tiff/bmp/pds in regards to the Material Room. The only exception may be, as noted, with extreme closeups and with images who's original format is uncompressed.)

What do you need vs how much operating capital you have and what is the final resolution of the texture going to likely be? Remember that resolution doesn't mean squat unless you're going to actually be making use of it. A very lightweight, compressed, jpg texture set is just fine for the diffuse texture of figures that will be rendered at a distance. In that sort of render, the pixel resolution is small in consideration to total render size. But, you may wish to keep a higher resolution for transmaps, since lower resolution in those have the chance of being much more noticeable, especially if you have some procedurals coupled with them.

In Conclusion: To me, there's one Rule that would probably serve just about any situation: Unless you are rendering extreme close-ups, you may not notice any difference whatsoever across the spectrum of compressed-vs-uncompressed formats. And, do not forget that how these maps are filtered in the final renderer matters as well, so selecting None, Crisp, etc.. for your texture filter will matter a bit. (One does have to note that some versions of Poser are more capable than others in regards to map file-types, so I may not be familiar with some versions that make special use of image filetypes.)