dmsmith555 at yahoo.com
Mon Jan 30 09:37:29 MST 2006
Sergio Queiroz wrote:
> Le lundi 30 janvier 2006 à 09:54 -0500, DM Smith a écrit :
>> I found a possible enhancement for our BitSet representation at
>> The typical use case is that our set is sparse and often contains
>> clusters of verses.
>> The only problem I see is that there is no method to determine the
>> cardinality of the set, as size() is not implemented.
> Isn't this their .length() method?
Now I see cardinality().
length() provides the range of integers that is currently represented in
the set. So if 0-982 can be represented, then 982 is returned. (i.e. it
is the highest index + 1)
size() is implemented, but deprecated.
I looked at the license and we cannot use it. It uses the Java Research
License and it appears to also use LGPL, but the jar does not contain
the LGPL license nor refers to it. The website mentions that license. We
are fine with LGPL but the JRL only allows the code to be used for
research and personal use.
That said, I think I am going to experiment to see what kind of gain it
provides. If we can reduce the runtime RAM requirements, that would be good.
The folks at Lucene are looking into it, because in their license
context, they allow for commercial use. They might create a clean-room
implementation of it. We could use that.
More information about the jsword-devel