Update README.md

Add info about the version to use, current status, etc

README.md

@@ -4,23 +4,32 @@ Implements "OCF Compresion" as used by Cerner PowerChart.
 
 The compression is basically just the same LZW compression as used in TIFF images.
 
-Background
-===========
-
-I wanted to produce a simple decoder that was somewhat "C# like" rather than just another
-port of the C implementations. As such, it trades speed for "style".
+Please use the `OCfLzw2` version of this library. The previous version in `OcfLzw` folder is 
+no longer under development as it was too slow.
+
+An alternate implementation by Bruce Jackson may also be of interest:
+ * https://gist.github.com/pgodwin/7d66729444173146ad698d154f2b9b6c
+ * https://blog.brucejackson.info/2013/03/deconstructing-lzw-decompression.html/
+
+## Current Status
+The code has been extensively tested against more than 10 million "wild" LZW blobs without issue. 
+It handles compressed streams with a chunk-size of up to 13 bits. 
+
+## History
+I wanted to produce a simple LZW decoder which could specifically handle "BLOB" data from Cerner PowerChart.
+I also wanted something which was somewhat "C# like" rather than just another port of the C implementations. 
+As such, it trades speed for "style".
 
 This port is based on two Java Implementation of TIFF LZW:
 
  1. PDFBox was the original source (https://pdfbox.apache.org/.).
  2. TwelveMonkeys TIFF LZWDecoder (https://github.com/haraldk/TwelveMonkeys)
 
 I've made use of PDFBox's NBitStream for reading the variable bit-widths from the stream.
-However it's LZW decoder seemed to be very heap-intensive (and really didn't need the hash of a dictionary),
+However it's LZW decoder seemed to be very heap-intensive (and I really didn't need the hash overhead of a dictionary),
 with decompressing 1 million values taking over 2 min on an i5-4590s.
 
-As I want to run this across millions of rows, the performance just wasn't there.
-Instead, the main decoding loop is from the TwelveMonkeys Image IO extensions,
+Instead, the main decoding loop is based on the TwelveMonkeys Image IO extensions,
 specfically - https://github.com/haraldk/TwelveMonkeys/blob/master/imageio/imageio-tiff/src/main/java/com/twelvemonkeys/imageio/plugins/tiff/LZWDecoder.java
 
 Rather than a dictionary and node-tree, it just uses an array. This reduces the number of objects created, and has a much
@@ -30,14 +39,12 @@ As a result, running 1 million values on the same i5-4590s now completes within
 
 I think there's room for further optimisation, but I want to try and ensure the code remains readable.
 
-Current Status
-===============
-The code is passing it's basic tests, but I need many more samples to ensure it works correctly across my use-cases.
 
-Todo:
- - Package and publish as nuget packages
- - Add SQL CLR functions (likely as a separate library)
- - Further performance improvements
+### Todo:
+
+ * [ ] Package and publish as nuget packages
+ * [ ] Add SQL CLR functions (likely as a separate library)
+ * [ ] Further performance improvements
 
 ## Licenses
 License: http://www.apache.org/licenses/LICENSE-2.0