At 09:21 AM 6/29/99 GMT, Werner LEMBERG wrote:
>
> R3: join causing -> left-join causing
>
> R4: join causing -> right-join causing (first instance)
> join causing -> left-join causing (second instance)
>
>Are you sure that this is the right fix?
After some further reading it does look like the defintions of
right-join causing and left-join causing are contradictory to how
they are used. I patterned my proposed fix on Rule R2, not the least
since you had found it unobjectionable, but it suffers from the same
problem you are mentioning here.
Therefore we must use these definitions:
right-join causing | superset of dual-joining, left-joining or join-causing
left-join causing | superset of dual-joining, right-joining or join-causing
in the understanding that right-join causing means causing a join on the
right side of the character that is the subject of the rule.
A./
PS: before you report it again, let me point out that the word liking as
used in table 6-8 is the same as the word joining elsewhere. This is being
fixed.
> A right join-causing
>character on the right *inhibits* joining to the left! I'd rather
>expect a `left join-causing character on the right' etc. Otherwise,
>this rule says that the medial form should be used if the left and
>right characters are either TATWEEL or don't join to the character at
>all...
>
>
> Werner
>
>
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:47 EDT