"Discussion" quality of answers, and soft-ness of questions

Question

Proposal: Computer Graphics

There are a lot of "soft questions" here, and usually the tools on SE sites don't like questions which are open ended, or 90% discussion of different ways and the pros/cons of each.

How will this be dealt with? I have had before (on code-review) cases where I've given a pseudo-code algorithm and explained how it works only to have someone to go "but the OP is using Java". While it wont directly carry across I can see answers like:

"Adding to your current system would be difficult, to achieve X you need a structure that better allows you to do Y and Z, like for example..." which is entirely valid!

The myriad of existing game engines clearly shows us there is no "one correct way" so there is foundation to this concern.

(this concern stems from stuff I can see coming up like "how to implement shadows" or something. Also since programmable-pipelines have come along how engines handle stuff like shader attributes is HUGELY different, and they do have to specialise for indented uses quite early on)

Answer 1

As most codebases are very unique because somebody is writing his own framework, first renderer or using Engine XYZ it would suffice to give a short description of why we think the questioner has a certain problem.

And then hand him the best practices to solve or avoid the problem in form of pseudocode, a paper or an "Explain me like im Five" text.

The best way i've come across such situations was communicating with graphics e.g. drawing on provided screenshots and then pointing out the computational problem and they would have to adapt it and come back with their result.

So the basic SE-Question-Iteration.

I just imagined the question of 'Why is my texture blurred?' and would assume the asker gives a screenshot and information about his 'softwarestack' so we could isolate whether its an engine setting problem or some algorithmic problem or whatnot.