New issue 1109: [experimental] default way to write images in VR?
Note differences in ways to write images in VR: (1) you specify an fname in the yt.volume_render() function originally (2) you get the resulting im object from yt.volume_render() and then write it using im.write_png(fname), (3) you take the resulting sc object from yt.volume_render() and then sc.render(fname). i guess it is good to have these three methods for getting the images to file, but it might be worthwhile to explicitly state that there are three methods for writing your VR to file, so people don't get confused on which is the main method (as i did and just had to figure it out by looking through different examples, source, etc).
New issue 1107: [experimental] consistency in camera motion
The various means of moving the camera can be confusing. As an example, `rotate()` doesn't rotate the direction of the camera so much as it rotates the position of the camera relative to its focus.
`yaw`, and `pitch` do the same thing, even though those are normally words associated with changing the orientation of something and not it's position. How to reconcile? I'm not opposed to leaving as is, but we need to be explicit about this in narrative docs and docstrings so as to not confuse users.
New issue 1105: [experimental] make subpage in docs for TransferFunctionHelper
in the VR docs, TransferFunctionHelper has a very useful long notebook demonstrating how to use it. unfortunately, it's right in the middle of the narrative docs, and it really breaks up progression of the description of using VR. I think it should be split out into a different page in the narrative docs and linked there as well from the cookbook/notebooks page.
New issue 1103: [experimental] render() with off_axis_projection yields no image
the way to render an off_axis_projection requires yt.write_image() instead of sc.render() like the VRs even though it is using the same infrastructure. inconsistent. why? sc.render() *runs*, but even with a clip_ratio, it just gives a black screen because all of the values are NaNs. bug? problem? if i'm doing it wrong, perhaps others will, and it could be documented how to use it correctly.