On Mon, 2007-10-22 at 16:38 -0400, Ian Eslick wrote:
> Well here are the open decisions:
>
> - Characters as an independent type or a numeric type?
Can we usefully think of Unicode characters as a numeric type?
The real problem is that each culture uses a different collation
order, so it is hard t
Well here are the open decisions:
- Characters as an independent type or a numeric type?
- Case sensitive vs. insensitive string/symbol/pathname sorting?
(insensitive was the policy implemented by the original developers)
- Settle on a standard type ordering and find some way to do a range
qu
I think this is the best policy that we can generate.
I personally don't like it, as I've mentioned before. I don't think
we should have an arbitrary order. If you don't have a specified order,
then you don't really have a btree in meaningful sense. You can't do a
range query over something th
Another detail to iron down is the implications of change-class and
redefining a class via defclass.
change-class is pretty easy as it is an explicit call by the user to
change a given instance. I added a warning mechanism that signals if
you are going to delete data from a store by droppi
Lisp, by the way, only allows comparisons between characters in terms
of their code. So sorting them as one of numerous numeric values, or
independently by code, is adding to the lisp spec, not violating it.
Ian
On Oct 22, 2007, at 1:18 PM, Ian Eslick wrote:
There have been a number of dis
There have been a number of discussions about sorting issues and
guarantees among the various data store implementations. This is
difficult due to different backend serialized representations. This
is my understanding of the current guarantees. Let's commit to this
as part of locking dow
