Thanks, Sally.

Today is the anniversary of AFNI's release to the general public -- that is, to non-MCW users. I was originally planning Feb 1, 1995, but something happened to make that date slip (I don't recall what).

For historical reference, AFNI was conceived of in late June 1994, and the first version (what I now call version "0.5") was semi-usable by sometime in September 1994. In December 1994 I rewrote the I/O for AFNI, discarding the original dataset format in favor of the .HEAD/.BRIK format. Shortly after that I wrote 3dmerge and 3dttest. (And also the infamous memo stating that I was quitting all work on AFNI for the rest of time.)

The first version of AFNI was developed on an HP-UX workstation, then ported to an SGI-IRIX system in Steve Rao's lab at MCW. (It crashed on the first try on the SGI, because of lack of memory. Steve ordered some via FedEx so we could try it again the next day.) The first users at MCW, besides myself, were Julie Bobholz and Julie Frost. I don't know who the first users outside of MCW were.

For that matter, I don't really know who uses AFNI outside of MCW and NIH even now, except by noting queries here on the message board. I can guess that about 500 computers seem to be running AFNI, based on version check logs in the last month. About 100 at NIH, 50 at UCSD, 35 at Princeton, 30 at MCW, 25 each at U Chicago and Harvard-MGH, 15 at U Wisconsin, and 10 each at Emory, Stanford, Brown, Northwestern, and San Diego State. Plus smaller numbers at diverse other sites. Outside the USA, the greatest number of users are in Canada and Italy (about 15 each). We have given classes in using AFNI at the NIH, Princeton, U Iowa, U Pisa, and Woods Hole.

bob cox