It's never a good idea to start optimizing so soon for the very simple fact that you don't really know how many other effects there will be by the time the game is programmed (you think you know, but you don't, really).

Always keep a brute-force renderer (VB per each object) up-to-date and when the feature-set of the engine is finally finished, only then go about putting them to a single VB/IB, or grouping render calls by texture/shader/whatever.

There could be million reasons why you don't see the second object:

- incorrect matrices

- incorrect position of the object

- incorrect lighting

- the object is hidden by some other object

- object is outside view frustum

- object is rendered to a different RT

- .....

You know, it may seem like it is a good idea to introduce an optimization or two, thinking you may actually save the time later.

But the first time I actually wrote the optimization render methods, it took me less than an afternoon to write one method that renders everything from a single static VB/IB and another method that renders eveything from a dynamic VB recreated based on current view frustum.

The point is, you won't save the time, since every time you'll make some adjustments to a scene, you have to deal with the early optimization too, thus prolonging the debugging, which is going to happen perhaps even hundred times before you actually get to doing the optimizations.

And, most important of all, if this VB/IB access is not actually a bottleneck, you just lost yourself a lot of time for no particular gain...

Of course, there's no better way to remember it than banging your head on the wall afterwards

The number one Headache is simply remembering the API function calls

It is actually dangerous to just remember the parameters. You always want to check the other behaviour of the fn that is mentioned in the docs.

The time you save by remembering the parameters is nothing compared to the time you loose when stuff stops working and it's only after you finally consult the dx docs [with the error you were given] you figure out the problem.

If you really work all the time just at the rendering layer, then you will, eventually, remember the parameters and everything else. But the renderer is usually, just about 10-15% of the total effort. So, once you are done with the low-level rendering stuff, you spent majority time at gameplay/ai/toolset anyway.

There was a time, I remembered majority of dx functions with their parameters. But fast forward 8-12 months, which I spent just with the gameplay, AI, tools and I forgot most of them. Which is a good thing.