On Mon, Oct 20, 2008 at 6:16 PM, Frank Mittelbach
<[log in to unmask]> wrote:
> Joseph Wright writes:
>  > Will Robertson wrote:
>  > > I noticed that recently too. I think it would be better to remove that
>  > > variant in the expl3 syntax, since it doesn't seem like it adds much of
>  > > importance but it adds a bit of obfuscation.
>  >
>  > As you probably guessed from the question, that would be my preferred
>  > solution too.
>
> I agree with both of you.let's mark it as obsolete for now and wait for Morten
> to speak up who introduced it (I guess)

It is used in a few cases such as

\def_long:Npn \tlp_if_in:NnTF #1#2{
  \def_long:Npn\tmp:w ##1 #2 ##2\q_stop{
    \quark_if_no_value:nFT{##2}
  }
  \exp_after:NN \tmp:w #1 #2 \q_no_value \q_stop
}

If there was no FT variant, this function would either have to read
the TF arguments twice or the quark test would have to be done
manually.

The FT variant was never meant to be heavily used - just a shuffle
variant so to speak.
-- 
Morten