[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


>> The flexibility of JOI comes at the expense of always requiring the
>> elidable to terminate the previous construct.
>{broda joi brode} requires no terminator.

The reason we don't need a terminator in:

  le broda .a le brode

is because .a is only legal for sumti, not selbri -- we have ja for selbri.
The comparable:

  le broda ku joi le brode

needs "ku" because joi has wider application than .a, so without the {ku} we
wouldn't know whether joi was connecting {le broda} to another sumti or
{broda} with another selbri.

If JA were allowed to fill both roles, you'd have to say:

  le broda ku ja le brode

You couldn't have

  le broda ja le brode

because scanning that requires more than an LALR(1) parser.

Hmmm but now that I go through that train of logic I've changed my mind --
having to put ku before ja in that case isn't so bad, really, since .a is
still there if you prefer it.  If it can be done so that no existing use of
JA requires a terminator, only new uses, then maybe I favor the change.

>> I would have proposed long ago to make multiple non-logical
>> selma'o with different cmavo assigned to each - the exact oopposite of your
>> proposal. %^)  (WE never seriously considered that one either.)
>Why would you like to make the grammar more complicated? I think that
>there are already too many selmaho for very specific things that should
>not be there. Having to learn exceptions and special cases is difficult.
>The more selmaho there are the more difficult the language becomes.

One measure of difficulty is the number of selma'o; another might be the
number of situations where terminators are needed.  Selma'o are for parsers;
for learners, the choice is between explaining "use .e for sumti, je for
selbri" or "put terminator before je when connecting sumti".  Neither seems
particularly easier to me.
 Chris Bogart        \  /  http://www.quetzal.com
 Boulder, CO          \/   cbogart@quetzal.com