It seems that I can't get \definecomplexorsimpleempty to work right. The command I define works fine unless the simple version is called, in which case the first character in the body of the command is checked \doifnextcharelse[... which thus fails but in some manner eats it (even though it is output). The problem with the \doifnextcharelse is that the character it checks against is \catcode'd to \@@active, but it won't work when the character has been checked. Is there any way to fix this? I understand if my problem isn't that clear, so I've attached the module. The idea is to be able to write \startbnfgrammar <exp>: <num>. <num>: "0";"1". \stopbnfgrammar and this is then typeset as a table, but currently you're forced to write \startbnfgrammar[] <exp>: <num>. <num>: "0";"1". \stopbnfgrammar or the first < in <exp> gets eaten by \doifnextcharelse and the \def<{... won't be called causing problems down the line when the > is encountered, which expands into a \egroup which stops tex as there is no matching \bgroup. Now, is there any way to get \doifnextcharelse to properly deal with \catcode'd characters (such as my <) that it encounters when looking for other characters (such as the [ for optional arguments to commands)? I'd really like to resolve this, so any comments or tips are appreciated. Second, is there any good reference to programming TeX? I mean, one can figure out most stuff from reading CONTEXT's source for example, but it would be great to find out the exact implications of builtins such as \relax and so on. nikolai -- ::: name: Nikolai Weibull :: aliases: pcp / lone-star / aka ::: ::: born: Chicago, IL USA :: loc atm: Gothenburg, Sweden ::: ::: page: www.pcppopper.org :: fun atm: gf,lps,ruby,lisp,war3 ::: main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}