[PSR-6] Renamed "pool" to "cache" for better usability, made naming consistent

[webmozart]

Hi there!

Thank you @tedivm for explaining the cache approaches on your blog. I like the proposal.

As an outsider of the cache working group, I would like to bring in some unbiased feedback:

1. The concept of a cache "pool" is great for discussion of various approaches. The end users of the PSR, however, don't need to know what a "pool" is. For them, it is IMO sufficient and easier if the "pool" is simply called "cache".
2. The item has getKey(), but get(). This is a bit weird - IMO the code becomes easier to read if get() and set() are called getValue() and setValue().
3. The pool has getItem() and getItems(), but deleteItems() only. I added deleteItem() for the sake of completeness.

I have some more suggestions which require more discussion however. I'll submit them in a separate PR.

Thanks for your hard work! :)

Cheers,
Bernhard

[milesj]

I agree with all these changes. After recently attempting to build a cache layer based on this PSR, I came to the same conclusions.

[Crell]

The naming here was discussed at obscene length in the past. The idea of keeping the "pool" in the name was to make it clear that we were talking about a specific pool, bin, etc, not about the cache driver itself. That was a point of confusion before.

(I personally preferred "bin" to "pool", but the consensus on list was to go with pool.)

[dragoonis]

I am happy with the term Pool. If we thought about changing it to Cache I would have no objections.