Reading Datasets Across the Web
===============================
As of 26 Mar 2001, the interactive AFNI program has the ability to read
dataset files across the Web, using the HTTP or FTP protocols.  There
are two ways to use this, assuming you know a Web site from which you can
get AFNI datasets.

The first way is to specify individual datasets; for example

  afni -dset http://some.web.site/~fred/elvis/anat+orig.HEAD

This will fetch the single dataset, and start AFNI.

The second way is if the Web site has a list of datasets stored in a file
named AFNILIST.  If you specify this as the target for a Web dataset, AFNI
will read this file, and retrieve each dataset specified in it (one
dataset per line); for example

  afni -dset http://some.web.site/~fred/elvis/AFNILIST

where the AFNILIST file contains the lines

  anat+tlrc.HEAD
  func+tlrc.HEAD
  reference.1D

Note that the AFNILIST file can contain names of 1D timeseries files.
One way for the Web site creator to create an AFNILIST file would be to
put all the dataset files (.HEAD, .BRIK.gz, .1D) into the Web directory,
then do "ls *.HEAD *.1D > AFNILIST" in the Web directory.

The "Define Datamode" control panel has a new button "Read Web" that
will let you load datasets (or AFNILISTs) after you have started the
program.  These datasets will be loaded into the current session.
However, you cannot write out datasets read in this way.  Also, these
datasets are locked into memory, so if too many are present, your
computer system may get into trouble (i.e., don't download 10 60 MB
datasets at once).

ftp:// access is done via anonymous FTP; http:// access uses port 80.