diff --git a/draft-ssmith-cesr.md b/draft-ssmith-cesr.md
index e7898f7..13b9d9e 100644
--- a/draft-ssmith-cesr.md
+++ b/draft-ssmith-cesr.md
@@ -402,7 +402,7 @@ Special CESR count codes support re-synchronization at each boundary between int
 
 Given the popularity of three specific serializations, namely, JSON, CBOR, and MGPK, more fine grained serialization boundary detection for interleaving CESR may be highly beneficial both from a performation and robustness perspective. One way to provide this is by selecting the count code start bits such that there is always a unique (mutually distinct) set of start bits at each interleaved boundary between CESR, JSON, CBOR, and MGPK. 
 
-Furthermore, it may also be highly beneficial to support in-stride switching between interleaved CESR text domain streams and CESR binary domain streams. In other words the start bits for count (framing) codes in both the *T* domain (Base64) and the *B* domain should be unique. This would provide the analogous equivalent of a UTF Byte Order Mark (BOM) [BOM]. A BOM enables a parser of UTF encoded documents to determine if the UTF codes are big endian or little endian [[19]]. In the CESR case this feature would enable a stream parser to know if a count code along with its associated counted or framed group of primitives are expressed in the *T* or *B* domain. Toghether these impose the constraint that the boundary start bits for interleaved text CESR, binary CESR, JSON, CBOR, and MGPK be mutually distinct.
+Furthermore, it may also be highly beneficial to support in-stride switching between interleaved CESR text domain streams and CESR binary domain streams. In other words the start bits for count (framing) codes in both the *T* domain (Base64) and the *B* domain should be unique. This would provide the analogous equivalent of a UTF Byte Order Mark (BOM) [BOM]. A BOM enables a parser of UTF encoded documents to determine if the UTF codes are big endian or little endian. In the CESR case this feature would enable a stream parser to know if a count code along with its associated counted or framed group of primitives are expressed in the *T* or *B* domain. Toghether these impose the constraint that the boundary start bits for interleaved text CESR, binary CESR, JSON, CBOR, and MGPK be mutually distinct.
 
 
 Amongst the codes for map objects in the JSON, CBOR, and MGPK only the first three bits are fixed and not dependent on mapping size. In JSON a serialized mapping object always starts with `{`. This is encoded as `0x7b`. the first three bits are `0b011`. In CBOR the first three bits of the major type of the serialized mapping object are `0b101`. In MGPK (MsgPack) there are three different mapping object codes. The *FixMap* code starts with `0b100`. Both the *Map16* code and *Map32* code start with `0b110`.