Add implementation guidelines to the documentation

[asfimport]

I'd like to add a paragraph to the documentation providing implementation guidelines:
An execution engine (or framework, or UDF executor, or storage engine, etc) can use only a subset of the arrow Arrow spec or extend it given the following constraints:

Implementing a subset:

1. If it is only producing (and not consuming) arrow vectors.

• any subset of the vector spec and the corresponding metadata can be implemented

2. If it is consuming and producing vectors

• there is a minimal subset of vectors to be supported (To Be Defined)
• production of a subset of vectors and their corresponding metadata is always fine
• consumption of vectors should at least convert the unsupported input vectors to the supported subset (for example Timestamp.millis to timestamp.micros or int32 to int64)

An execution engine implementor can also extend their memory representation with their own vectors internally as long as they are never exposed. Before sending data to another system expecting Arrow data these custom vectors should be converted to a type that exist in the Arrow spec.
An exemple of this is operating on compressed data.

Reporter: Julien Le Dem / @julienledem
Assignee: Julien Le Dem / @julienledem

PRs and other links:

• PR

_{Note: This issue was originally created as ARROW-252. Please see the migration documentation for further details.}

[asfimport]

Wes McKinney / @wesm:
+1 I agree with these sentiments

[asfimport]

Jacques Nadeau / @jacques-n:
+1 as well

[asfimport]

Julien Le Dem / @julienledem:
PR: #120

[asfimport]

Julien Le Dem / @julienledem:
Issue resolved by pull request 120
#120