[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
Direction and name of d.cursor
- To: zzdev@xxxxxxxxxx
- Subject: Direction and name of d.cursor
- From: Tuukka Hastrup <Tuukka.Hastrup@xxxxxx>
- Date: Tue, 11 Jul 2000 10:17:32 +0300 (EEST)
I guess one might think I'm sort of reluctant to work ie. to code.
Tuomas assigned me the job of reversing the directions of d.cursor. The
reason was to make the first cell of the rank - the headcell -
"special" - the data cell which 0...n cursors are pointing to. Are you
relly sure this is how we want it? In my opinion, the name of the
dimension represents the relation type to the positive direction, and
negative direction is thus the opposite relation (eg. +d.next gives the
next cell, -d.next gives the previous one).
Consistently, going poswards would give the cursors which point to this
cell, and going negwards would reveal which cells this cell is
"cursing", pointing to. This is the way it is at the moment. Maybe we
could accept that both headcells - the negative end and the positive
end - can be important.
Other possibility is changing the name of the cursor dimension to
switch the relation. Make it "d.curses" or "d.cursee" or "d.pointing" -
I'd prefer a noun, like d.cursee, as this should also be consistent.
Then I'll change the code.
If you read the 10 kilobytes I sent you the other day, you found some
diverse examples of one-to-many relations. Often, _none_ of the cells in
a rank is "special", but is found otherwise. That lessens the
importance of the headcell. Take this example:
Here, you get the "special" cell going up the rank and then negwards on
d.contains. (Notice that here again, going poswards on d.contents gives
you the contents.) The headcell (chapter 1) has no special meaning here.
Thanks if you bothered to read this through. If you say that Tuomas's
suggestion is the right way to do it, then I'll do it.