existing-records.md

@@ -66,7 +66,7 @@ This would mitigate the problems caused by accessors being partial functions sin
 
 
 
-Note from Simon.  I hate that the above defn of 'f' has just one argument (val {x="foo")), 
+**Note from Simon**.  I hate that the above defn of 'f' has just one argument (val {x="foo")), 
 whereas it looks as if it has two.  (This is a problem with existing Haskell.)  It looks
 like 'f' has an argument 'val' and another arguement that is a free-standing record, 
 something we really want in the end anyhow.  Not sure how to fix this.  `val@{x="foo")`?
@@ -125,6 +125,15 @@ Note the change in the type of the stored field.
 At the moment, such a record update must be written using the data constructor, not the update syntax.
 
 
+
+**SLPJ:** That isn't true.  Haskell 98 already supports what you suggest here, and the code above compiles with GHC without flags.  
+
+
+
+However I would like to argue for *removing* this feature.  It causes a quite unreasonable amount of
+pain in the type checker, especially once we start thinking about GADTs and existentials.  And I think the benefit is modest at best.  I don't think programmers generally expect record update to be a type-changing operation.
+
+
 ## 'Open' statement