Documentation: Add new sections such as 'requirements'

Documentation/book.md

@@ -12,6 +12,29 @@ sansfont: Inter
 monofont: "IBM Plex Mono"
 ...
 
+# How to use this document
+We encourage you to read this document in its entirety, as it provides a comprehensive
+overview of the VMP project. Feel free to pause and experiment with the code and
+examples provided.
+
+The goal is to provide a comprehensive guide for both deployment and development.
+First by giving you a birds-eye view of the architecture, and then by diving into
+the technical details.
+
+If you are only interested in deployment and administration, you can skip the
+in depth technical chapters which may require knowledge of the GStreamer multimedia
+framework.
+
+# Requirements
+This document assumes you have a basic understanding of multimedia processing, and
+have experience with administering a Linux system.
+
+If you are interested in development, you should first familiarize yourself with
+Objective-C, and the GStreamer multimedia framework.
+Various GStreamer demos and tutorials are available on the official website, and
+specific demos which were eventually integrated into the VMP project are available
+in `Demos/`.
+
 # Chapter 1. Introduction
 
 The importance of livestreaming university lectures and other related events has
@@ -34,7 +57,7 @@ equipped with video and audio capturing capabilities, as well as the power to
 blend and encode multiple feeds in real time. The resulting stream is then published
 via RTSP, or as a raw network stream.
 
-In theory, compositing—the combination of multiple video and/or audio feeds—can
+In theory, compositing — the combination of multiple video and/or audio feeds — can
 also be transferred to a data center. However, this necessitates
 hardware-accelerated video processing, typically performed using GPUs. Since
 edge devices also need to encode the streams to minimize bandwidth, which in
@@ -50,6 +73,34 @@ requirements at a fraction of the cost of comparable commercial equipment.
 
 ## Chapter 2. Architecture 
 
+The VMP project is by design hardware-agnostic. Unique hardware capabilities, can be
+described by a profile which are usually located in `/usr/share/vmpserverd/profiles`.
+
+On startup, the *VMP daemon (vmpserverd)* reads the profiles and probes the
+enviroment to determine the best profile to use.
+
+But what does the VMP daemon do? It is the core of the VMP project, processing
+the video and audio feeds, and exposing them via RTSP. Additionally, there is
+also an option for recording the processed streams to disk. But that is just
+the gist of it.
+
+The VMP daemon is built on top of the GStreamer multimedia framework, which
+provides a pipeline-based model for constructing multimedia workflows.
+
+In practise, you do not have to be aware of the underlying pipelines as we
+provide a set of predefined *channels* and *mountpoints* for most use cases.
+
+Think of *channels* as a way to describe the processing of an input stream, such
+as a capture card. A channel will always expose an interface in which other
+modules can grab the resulting stream. An example would be a presentation channel
+that grabs the presentation feed from a capture card via V4L2 (let's say
+`/dev/video0`), and scales it to 1920x1080p.
+
+A *mountpoint* then takes one or more channels, processes and encodes them, and
+packages them into a network stream. This is then exposed via RTSP at a path
+described by the mountpoint. A mountpoint in turn mounts a network-stream to
+a given RTSP path.
+
 ### Chapter 2.1 Terminology
 
 In this section, we define key terms to ensure consistency throughout this document:
@@ -101,6 +152,8 @@ filtering operations, making it hardware and vendor-agnostic.
 
 ### Chapter 2.3 Channels
 
+### Chapter 2.4 Mountpoints
+
 ## Chapter 3. Deployment
 
 ## Chapter 4. Development