Ok, so I have been doing some UV tutorials, and have learned quite a useful technique that I thought I would share.
Before I can explain the concept you need to understand a little something about UV maps.
When you open your UV mapper in your modelling program, you're presented with a square graph with the U and V axis going from 0 to 1.
Now, most of the time, we crush all the UV data into that square to create our texture map. But what do we do if say, we want a character and a gun to share the same UV map (as in A. in image below)?
so, first, UV map the character using the entire UV space. Then once that is done, uv map the gun using the entire UV space.... now here's the clever bit.... you now shrink the character's width so it only fills the left half of the UV map. And shrink the gun so it only fils the right half. (as in B.)
Although your texture co-ordinates are now squished, you can create your texture, fullsize at say 1024 by 512. (as in C). because the UV map is calculated on values from 0 to 1 the texture will still be correct. in the UV map the character texture uses the left 50% of the map, and 100% of the height, and although the texture image is twice as wide, the character part of it is still 50% of the width and 100% of the height.
Thought people might find this useful as an alternative to using multitexturing.
Below is an example image (A,B and C are referenced in the text above).
