Saturday, January 21, 2006 - Sat, 21 Jan 2006

Finally everything’s set up and I can compile Behold again and continue on with development. I hope I can remember where I am, since I left it right in the middle. But the main benefit of this pain should be no more system crashes that lose work, cause frustration and lose time.

For fun, I thought I’d try to load a large file with Behold on my new machine. I took the same 24 MB one I mentioned in my May 30, 2005 post. It loaded after about 8 minutes, but still red-lined on both the 1 GB of memory and the 1.5 GB swap file.

I do realize that this indicates that Behold is currently an inadequate tool for very large family files. Even a minute to load is simply too long. I haven’t yet addressed tuning to make Behold faster. There is a lot I can do there, but it will have to wait until all the alpha functionality is in place and I get to beta testing.

The program that impresses me the most speed-wise is GENViewer by MudCreek Software. It is blisteringly fast, loading GEDCOMs at about 20,000 names a second! I know it must only load indexes into memory and keep the data on disk, but even so, I’d love to find out how they get that sort of speed.

GENViewer has a light version that is freeware. I recommend you try it. MudCreek also has a GENMatcher program for those of you who were looking for a way to quickly compare two genealogy files for matches, or one genealogy file for duplicates. I must say these are two very nicely thought-out programs!

The first conceptual difference between GenViewer and every other program with Behold is other programs load data quickly but take time to generate reports every time you want one. Whereas Behold takes time to load initially but then its report is everpresent. The second main difference will be when editing is added to Behold since, unlike every other program, with Behold you will do your data entry and editing directly on the Everything Report.