[CORE] Doc: Add developer guide for adding new backend / component for Gluten

Beginning with https://github.com/apache/incubator-gluten/issues/7031, after months of iteration on the new `gluten-core` module, we are basically have an ready-to-use Spark accelerator SPI extracted out from the old Gluten code base.

Now (after https://github.com/apache/incubator-gluten/pull/9900) `gluten-core` is a thin Spark SPI layer that is not directly relevant to specific accelerator technologies and their programming languages, e.g., C++ / Substrait / Velox / ClickHouse, etc.

We'll add some developer-oriented documents showing how to add a new backend in Gluten, E.g., a backend (component) for accelerating a specific lake-format processing, or a backend with brand-new native accessing layer or programming language, and how to make them cooperate smoothly with the existing Velox / ClickHouse backend in Gluten.

Start with the gluten-core module, the changes in pull request 9900, and the earlier discussion in issue 7031. The guide is done when it explains how to add a new backend or component, including a native access layer or programming language, and how it cooperates with the existing Velox and ClickHouse backends.

Written by the indexing model from the issue text.