I too would like to get rid of the ids, but I took a good look at the algorithm and I do not think it can easily be adapted.
the algorithm seems to work on sequences, which might work with a uniform tree structure, but our tree structure is just about as nasty as it gets a tree type (branch as instance) can be anywhere in the tree and even be present within the branch itself. Our problem is a lot more similar to the math graph theory. I remember APL (whoever knows that today) had some functions to compare and work on that, I will search my old library (meaning physical paper). rgds jan i. On 19 June 2015 at 18:28, Peter Kelly <pmke...@apache.org> wrote: > > On 19 Jun 2015, at 11:20 pm, Peter Kelly <pmke...@apache.org> wrote: > > > I experimented with this a while back using the Myers diff algorithm > > > [1] http://www.xmailserver.org/diff2.pdf > > This is a fascinating algorithm BTW, and one of my favourites of all time, > because of the way in which it re-casts the problem (diffing two files) > into a totally different form (finding the shortest path through a graph), > and explains the solution in terms of this latter form. The paper is pretty > heavy reading, but there’s a good implementation here: > > > http://www.codeproject.com/Articles/42279/Investigating-Myers-diff-algorithm-Part-of > > There’s an implementation of this in Editor/src/util.js (function diff). > It’s not actually used anywhere any I can’t remember why I added it there, > but useful nonetheless as as starting point if you’re looking for an > implementation. > > — > Dr Peter M. Kelly > pmke...@apache.org > > PGP key: http://www.kellypmk.net/pgp-key <http://www.kellypmk.net/pgp-key> > (fingerprint 5435 6718 59F0 DD1F BFA0 5E46 2523 BAA1 44AE 2966) > >
