At 11:41 AM +0200 on 6/20/99, M. Uli Kusterer wrote:
>>put (item (item (item 5 delimited by "." of c) to (item 6 delimited by return
>>of d) delimited by space to 2 of a) delimited by comma of b) into
>>what_the_hell_is_that
>Your syntax wouldn't clean that up anyway:
>
>put (item (item (item 5 delim="." of c) to (item 6 delim=return of d)
>delim=space to 2 of a) delim=comma of b) into what_the_hell_is_that
>
> Is just as bad.
Well, you left out a few parentheses... but, true. Putting the "delim="
part after the word "item" (not after the number) would help... but I think
that expression is ugly no matter what.
>But "delimiter is" doesn't even resolve to a real English
>sentence, so why should we prefer that over the others if it even
>introduces more ambiguities that the others?
How does delimiter is not resolve to a real sentence?
"delimiter is tab."
is a perfectly complete sentence minus one article. But 'the' is normally
option al in HT. So we could allow the full form:
put item (the delimiter is tab) 3 of...
>
> The only way we can truly work around this problem is if we allow
>scripter-defined chunk types. But that would make it much more verbose
>again:
Not only would it make it verbose, it'd make xTalk mutable... Imagine this:
put item1 1 of item2 5 of item4 7 of data into...
Even worse if a new "item-word" could be a multiple delimiter...
