How should we handle questions about features not yet in Blender?

Question

Currently, there are several questions on the main site that ask about features not yet in Blender, either planned, broken or not documented. How should we handle these type of questions, most of them are valid but have no set answers yet. I don't think answering these questions do the site any good at the moment. Answers like:

Probably the first release with Volumetrics will be 2.69 or 2.70.

or

This is in the roadmap and hasn't been addressed yet.

Answers like these do not pass on any useful information and detracts from what this site should be about, we are trying to create an archive of information here, not a list of feature requests or todos. It is also worth noting that most of these are cycles questions since it is still under development..

• Rendering smoke in Cycles?

• https://blender.stackexchange.com/questions/1296/is-there-going-to-be-the-option-to-mesh-your-sph-sim

• How to use the volume socket in Cycles?

• possibly Cycles Subsurface shader: GPU compute and CPU compute give totally different results

• Is it possible to limit the "Cycles Shadow from Transparency" per Object?

How should we handle these type of questions?

Answer 1

Unless the feature is available in some form...

• already written a branch or in blenders trunk.
• available as a script / addon.

... I think we should consider it out of scope.

Someone can ask if a feature exists to perform a certain task, but unless its in a blender release, or already implemented to be included in the next release - I think the answer has to be NO.

Otherwise we get too much into speculation and discussing ideas of features don't yet exist.

We could answer: "No but this isn't hard to add", giving some hints on how to add - a script, C files... etc. but this should be kept very limited. Once we start trying to help them program something themselves the discussion should be taken off stack exchange, or they can format their questions into something that fits into a scripting question in the cases where a python script can solve the problem.

Answer 2

In general terms
B.SE shouldn't be a seen as a destination for feature requests. It's OK to ask if a feature is implemented or where to find appropriate documentation and road-maps for planned or existing features. Better these questions get asked and answered on a public forum than repeatedly in private or IRC.

Asking about missing features / functionality
Sometimes functionality is only perceived to be missing, it might exist as an addon or relatively obscure key combo. (like control-A for skin modifier)

Questions about planned features
What else is there to do but reference the roadmap, preferably link it. Perhaps close it as off topic or beyond the scope?