diff --git a/constantine/proof_systems/constraint_systems/r1cs_circom_parser.nim b/constantine/proof_systems/constraint_systems/r1cs_circom_parser.nim
index a2254b93..ec6e6a05 100644
--- a/constantine/proof_systems/constraint_systems/r1cs_circom_parser.nim
+++ b/constantine/proof_systems/constraint_systems/r1cs_circom_parser.nim
@@ -121,6 +121,7 @@ type
   R1csCustomGatesList* = object
   R1csCustomGatesApp* = object
 
+  ## XXX: Make this a `R1CS[T]` which takes care of parsing the field elements
   R1CS* = object
     magic*: array[4, char]
     version*: uint32
diff --git a/constantine/proof_systems/constraint_systems/wtns_binary_parser.nim b/constantine/proof_systems/constraint_systems/wtns_binary_parser.nim
index 0a349b5e..1d3e6bd4 100644
--- a/constantine/proof_systems/constraint_systems/wtns_binary_parser.nim
+++ b/constantine/proof_systems/constraint_systems/wtns_binary_parser.nim
@@ -40,10 +40,7 @@ import
      - Content: Each witness element is a 256-bit (32 bytes) unsigned integer in Big Endian format
 ]#
 
-# We use `sortedByIt` to sort the different sections in the file by their
-# `WtnsSectionKind`
 from std / sequtils import filterIt
-from std / algorithm import sortedByIt
 from std / strutils import endsWith
 
 type
@@ -58,7 +55,7 @@ type
     num*: uint32 # number of witness elements
 
   Witness* = object
-    data*: seq[byte]
+    data*: seq[byte] ## Important: The values are *not* Montgomery encoded
 
   Section* = object
     size*: uint64 # NOTE: in the real file the section type is *FIRST* and then the size
@@ -82,6 +79,9 @@ type
                            # of each different section in the file and then parse them in increasing
                            # order of the section types
 
+## XXX: Add `Wtns[T]` type, which takes care of converting field elements and
+## does not contain `seq[Section]` anymore
+
 func header*(wtns: WtnsBin): WitnessHeader =
   result = wtns.sections.filterIt(it.sectionType == kHeader)[0].header
 
diff --git a/constantine/proof_systems/constraint_systems/zkey_binary_parser.nim b/constantine/proof_systems/constraint_systems/zkey_binary_parser.nim
index 169159d9..adcb6a77 100644
--- a/constantine/proof_systems/constraint_systems/zkey_binary_parser.nim
+++ b/constantine/proof_systems/constraint_systems/zkey_binary_parser.nim
@@ -108,6 +108,9 @@ type
                            # of each different section in the file and then parse them in increasing
                            # order of the section types
 
+## XXX: Add `Zkey[T]` type, which takes care of converting field elements and
+## does not contain `seq[Section]` anymore
+
 func header*(zkey: ZkeyBin): Header =
   result = zkey.sections.filterIt(it.sectionType == kHeader)[0].header