Rootsmagic Automerge test
There has been a lot of discussion about Rootsmagic version 4 being slow but in this case I am
very happy to report that it completely whipped the previous version in terms of speed.
Whilst using RM3 (for reasons I won't go into here) I was picking up some intermittent data
corruption problems. The file worked perfectly well, rebuilding of indexes was always successful but then it would
start failing mid stream on gedcom export. this was definitely a file issue as I tried it on different machines and
it seemed to be something to do with family linkage.
Anyway I decided to run the RM Automerges in an attempt to fix the problem, created a new RM3
file and also an RM4 file, opened both applications side by side and imported the suspect gedcom. I then ran the
Rootsmagic Automerges side by side, all 4 boxes checked and took a quick note of the time.
When I returned after 3 hours Rootsmagic V4 had already completed the task and
RM3 was reporting that it was still running it's First Pass. Well Rootsmagic 3 did not complete the operation until
around the 22 Hours mark, a whopping 19 hours later, so hats off to RM4.
Another interesting difference is the file properties on both Automerged files which are shown
below and I need to look into these in more detail to try to understand them. It does look like RM4 has merged more
individuals than RM3 and that would be good news, also with RM4 being a complete re-write you would not expect the
merge logic to be identical on both versions of the software.

|