I have to disagree with you in two independent regards.
BL> IMO this is a feature not bug because list use space as one of delimiters.
Only in the 1-dimensional case. That is, list will divide
'one two and three'
into four items to columnize (along the spaces) but otherwise respect
the items as present in the input data. That is,
'one', 'two' ,: 'and three'
will be seen as three items to be columnized, irregardless of the
spaces in the spaces in the shape 3 9 array. (I would have thought
that my proposed example would have made clear that "list" already
does the right thing for character data, i.e. respect items; it just
fails to do so when listing symbols. Going via ": (what "list"
currently does) is the wrong approach for symbols; going via 4 s:
(as I suggested) would be right).
BL> Thin space could be used here just as that used as thousand separator in
BL> french locale,
J Dictionary, Section I, paragraph 1, sentence 1:
The alphabet is standard ASCII
This is the restriction for J source code and programming environments
that any J programmer should be aware of if concerned about the portability
of his/her scripts. Likewise, it is important that standrad utilities for
developers don't depend on anything beyond "standard ASCII".
I very much appreciate that J supports the handling unicode data, but much
more so that J can be used on *any* platform that just supports ASCII.
We are far away from having everything capable to do Unicode.
Martin
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
