DM Smith wrote:
Someone has ported the GNU diff algorithm to Java (see
http://www.bmsi.com/java/Diff.java)
It is written as the diff of two arrays.
I compared the code against analyze.c in GNU's diffutils and it looks
like a straightforward port.
I think it would be trivial to port it to C++.
Appropriately, it is licensed with the GPL license.
I think I might be more inclined to use a BSD licensed version of diff
as a basis and to re-write the wdiff functionality so that we own it.
Even if we go the GNU diff route, we might want to rewrite wdiff anyway
since it does fairly little and could be handled better with code
specialized to our needs. (E.g. should we necessarily present "heaven"
vs. "heavens" as a change of a whole word? Or should we just present the
difference of a single character, like "heaven{+s}". You could also
use this for orthographic variation, e.g. heaven vs. heauen (1611 KJV
spelling) > hea{v|u}en. And we could always make this flexible and
user-selectable.)
But the main reason I'd prefer a BSD licensed version of diff would be
that we actually have the potential of having all third-party code be
BSD/MIT/Apache/LGPL. The only exception, currently, is GNU regexp, I
believe. Once we get a replacement for that (probably by using either
ICU's Regex functionality or a BSD version of regexp), we'll have the
ability to permit groups to do non-GPL licensing from us (e.g. if ABS
wanted to do a branded version of BibleCS). There are no explicit plans
for anything like this (and, no, certainly no one has suggested making
Sword closed-source), but it would be nice to maintain the possibility
of having flexible licensing if we have the need.
--Chris
_______________________________________________
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
