Idris Samawi Hamid ادريس سماوي حامد mailto:Idris.Hamid@colostate.edu 17. Mai 2017 um 16:04 On Wed, 17 May 2017 07:25:41 -0600, Hans Hagen wrote:

...

...

...

cd:instances

While we wait for support cd:instances: How about this heuristic to distinguish abstract commands from pre-supported instances:

======= If slanted and containing and containing *no* capital letters, treat as an instance. =======

(slanted is not part of the coding, instances is)

Sure, but we have to distinguish abstract commands from pre-supported instances of abstract commands, both of which are typeset as slanted.

So we need a simple name for the genus that encompasses both abstract commands and pre-supported instances of abstract commands.

Perhaps 'generic'. All commands typeset as slanted in setup-en etc. are generic commands. So to rephrase

======= If a command is generic and contains *no* capital letters, treat as a concrete instance. =======

So whatever hook marks off a command to be treated as slanted is what marks it as a generic command.

Below this is the entry for \section in my own version of the setups files. All commands with instances have the attribute variant="instance" and the predefined instances are listed under cd:instances. Commands which are slanted in the manual are produced by a \defineXXX command which is documented with the generated="yes" attribute. For your list you can ignore the values under cd:sequence and use only the listed instances. cd:sequence cd:arguments cd:instances Wolfgang