consensus_encoding: update package docs for consensus encoding clarity

diff --git a/consensus_encoding/src/lib.rs b/consensus_encoding/src/lib.rs
index 20b87a11..a2655f1b 100644
--- a/consensus_encoding/src/lib.rs
+++ b/consensus_encoding/src/lib.rs
@@ -2,30 +2,31 @@
 
 //! # Rust Bitcoin Consensus Encoding
 //!
-//! Traits and utilities for encoding and decoding Bitcoin data types in a consensus-consistent way,
-//! using a *sans-I/O* architecture.
+//! Traits and utilities for encoding and decoding Bitcoin data types using a *sans-I/O*
+//! architecture.
 //!
 //! Rather than reading from or writing to [`std::io::Read`]/[`std::io::Write`] traits directly, the
 //! codec types work with byte slices. This keeps codec logic I/O-agnostic, so the same
 //! implementation works in `no_std` environments, sync I/O, async I/O, and hash engines without
 //! duplicating logic or surfacing I/O errors in non-I/O contexts (e.g. when hashing an encoding).
+//! This crate only supports deterministic encoding and will never support types like floats whose
+//! encoding is non-deterministic or platform-dependent.
 //!
 //! *Consensus* encoding is the canonical byte representation of Bitcoin data types used across the
-//! peer-to-peer network and transaction serialization. This crate only supports deterministic
-//! encoding and will never support types like floats whose encoding is non-deterministic or
-//! platform-dependent.
+//! peer-to-peer network and transaction serialization. Bitcoin types which support consensus
+//! encoding implement the [`Encode`] and [`Decode`] traits.
 //!
 //! # Encoding
 //!
-//! Types implement [`Encode`] to produce an [`Encoder`], which yields encoded bytes in chunks
-//! via [`Encoder::current_chunk`] and [`Encoder::advance`]. The caller drives the process by
-//! pulling chunks until `advance` returns [`EncoderStatus::Finished`].
+//! Consensus encodable types implement [`Encode`] to produce an [`Encoder`], which yields encoded
+//! bytes in chunks via [`Encoder::current_chunk`] and [`Encoder::advance`]. The caller drives the
+//! process by pulling chunks until `advance` returns [`EncoderStatus::Finished`].
 //!
 //! # Decoding
 //!
-//! Types implement [`Decode`] to produce a [`Decoder`], which consumes bytes via
-//! [`Decoder::push_bytes`] until it signals completion by returning `Ok(DecoderStatus::Ready)`. The
-//! caller then calls [`Decoder::end`] to obtain the decoded value.
+//! Consensus encodable types implement [`Decode`] to produce a [`Decoder`], which consumes bytes
+//! via [`Decoder::push_bytes`] until it signals completion by returning `Ok(DecoderStatus::Ready)`.
+//! The caller then calls [`Decoder::end`] to obtain the decoded value.
 //!
 //! Unlike encoding, decoding is fallible. Both `push_bytes` and `end` return `Result`. I/O errors
 //! are handled by the caller, keeping the codec logic I/O-agnostic.
@@ -33,7 +34,8 @@
 //! # Drivers
 //!
 //! This crate provides free functions which drive codecs for common I/O interfaces. On the decoding
-//! side we provide:
+//! side we provide functions which take a consensus encodable type parameter `T: Decode` to select
+//! the output type's associated decoder.
 //!
 //! * [`decode_from_read`]: Decode from a stdlib buffered reader.
 //! * [`decode_from_read_unbuffered`]: Decode from a stdlib unbuffered reader (4k buffer on stack).
@@ -42,22 +44,24 @@
 //! * [`decode_from_slice_unbounded`]: Slice can contain additional data after decoding completes.
 //! * [`decode_from_hex`]: Decode from a hex string without heap allocations.
 //!
-//! Each function above takes a type parameter `T: Decode` to select the output type and its
-//! associated decoder. The following variants instead accept a [`Decoder`] type directly,
-//! instantiated with [`Default`], and can be used when the output type does not implement [`Decode`]:
+//! The following variants instead accept an agnostic [`Decoder`] type directly, instantiated with
+//! [`Default`], and can be used when the output type does not implement [`Decode`]:
 //!
 //! * [`decode_from_read_with_decoder`]: Counterpart to [`decode_from_read`].
 //! * [`decode_from_slice_with_decoder`]: Counterpart to [`decode_from_slice`].
 //! * [`decode_from_slice_unbounded_with_decoder`]: Counterpart to [`decode_from_slice_unbounded`].
 //! * [`decode_from_hex_with_decoder`]: Counterpart to [`decode_from_hex`].
 //!
-//! And on the encoding side we provide:
+//! And on the encoding side we provide similar functions for consensus encodable types.
 //!
 //! * [`encode_to_writer`]: Encode to a stdlib writer.
-//! * [`drain_to_writer`]: Drain an encoder to a stdlib writer.
 //! * [`encode_to_vec`]: Encode to the heap.
-//! * [`drain_to_vec`]: Drain an encoder to the heap.
 //! * [`encode_to_hex`]: Encode to a hex string.
+//!
+//! As well as variants for agnostic [`Encoder`] types.
+//!
+//! * [`drain_to_writer`]: Drain an encoder to a stdlib writer.
+//! * [`drain_to_vec`]: Drain an encoder to the heap.
 //! * [`drain_to_hex`]: Drain an encoder to a hex string.
 //!
 //! # Feature Flags