Joseph Wright

I suspect the reason we kept these internal is they need a rather precise form of the base name. We might want to broaden that out or re-work the interface...

> > > why plural and singular? > > > > > > `\quark_new_conditional:Nn` generates `\_quark_if_:n(TF)` or `\_quark_if_:N(TF)`, depending on the first argument. > > `\quark_new_recursion_conditionals:N` generates one of six...

OK, I've been looking over this and I see why we kept them internal - the interface is a bit odd. The current `\__kernel_quark_new_conditional:Nn` is reminiscent of `\prg_new_conditional:Nnn` or similar,...

Or ```latex \quark_recursion_new:N \q_mymodule_tail \quark_recursion_new:NN \q_mymodule_tail \q_mymodule_stop ``` i.e. including one or (perhaps) both quarks explicitly

My favoured would be ```latex \quark_new_conditional:Nn \q_mymodule_nil { T , F , TF } \quark_recursion_new:NN \q_mymodule_tail \q_mymodule_stop ``` and auto-generate, BTW (actually, I'd drop the `n`-type argument in the first...

The model used by `l3keys` means that one can't apply a property inside `\keys_set:nn`. Rather, a key called (say) `input-comparators.put_right` would be entirely independent of `input-comparators`. You could therefore define...

I think overall we should avoid this: it's best if the typeset code is the 'live' version.

> What is the current status of this? Is there any decision? At the moment we've made no change