Improve README

Author: Dirbaio

README.md

@@ -13,11 +13,19 @@ a length byte. This requires a lookahead of up to 254 bytes when doing streaming
 
 rCOBS outputs the length byte at the *end* of the chunk instead of at the start, completely
 eliminating the lookahead requirement for encoding. The tradeoff is decoding now has to be done starting
-from the *end* of the message backwards, so no lookahead is possible. The message has to be 
-read in its entirety before it can be decoded.
+from the *end* of the message backwards, so streaming decode is not possible at all. The message has to be
+read in its entirety into a buffer before it can be decoded.
 
 This makes rCOBS ideal for situations where data is *encoded* in constrained, embedded systems and *decoded* in 
-more capable systems, where the full buffering requirement is not a problem. 
+more capable systems, where the full buffering requirement is not a problem, such as:
+
+- Logging
+- Crash reporting
+- Telemetry/monitoring
+- Sensor data feeds
+
+This crate implements a streaming `no-std`-friendly encoder, and a non-streaming `std`-only decoder (enabled with `std` Cargo feature).
+`no-std` decoding is intentionally not provided: if you need to decode on embedded devices, standard COBS is a better fit for you.
 
 ## Examples
 
@@ -31,12 +39,16 @@ rCOBS-encoded:  11 22 33 44 05
 Message:        11 22 00  33
 COBS-encoded:   03 11 22  02 33
 rCOBS-encoded:  11 22 03  33 02
-
+Cr
 Message:        11 00  00  00  42 42 42
 COBS-encoded:   02 11  01  01  04 42 42 42
 rCOBS-encoded:  11 02  01  01  42 42 42 04
 ```
 
+## Credits
+
+Idea of reversing COBS to allow encoding with no lookahead by @jamesmunns. Implementation by @dirbaio.
+
 ## License
 
 This work is licensed under either of