When I was helping with the Nights Edge UT mod, from what I observed, the animated first person models where the choke point in production flow, some never made it into the game in time, as each one required so much work to get modeled rigged and animated right.
Then as the UT99 was eclipsed by new more advanced game engines, with the coder/Team leader having completed implementation of the main features he had wanted, the final version was released still missing most of the planned additions to the arsenal.
Nights Edge was a very cool mod, that did not get as much attention it deserved, despite having weapon models that IMHO looked even prettier/flashy then the ones in Inf and other UT mods.
If you do away with the First person model, then weapons that would have been a headache to implement (like chain feed machine guns) become no problem, and matching up and getting an iron sight view to look right/accurate on a 3D mesh is no longer a factor.
Several commercial games have used similar 2D textures for their weapon sight pictures, and have done well and even won awards and accolades.
for example
Using an accurate silhouette of the real weapons Sights, is as far as simulating the feel of looking down the real sights goes, just as equal in effect as having the same outline formed by a high poly 3D mesh.
The ease and speed of implementing accurate portrayal of the sights with a silhouette texture, balances out against whatever eye candy advantage a 3D mesh would otherwise hold.
Imagine having say, fifty guns with selectable attachments for each and using a realistic sight picture made from a texture.
Or having five guns without attachments, that formed the same sight picture by using a mesh.