NIST Web Metrics
Directory Structure
|
VUG
Home Page
NIST Web Metrics
Directory Structure
|
 |
NIST Web Metrics Directory Structure |
 |
The software relies on the setting of certain symbolic names to locate key directories. This may be done with a command such as:
setenv WEBLOG_DATA /home/joe/WebMetrics/WebLogsThere are four names to be set; note that these should be set to absolute pathnames, not relative, so that they are valid no matter from where within the directory structure they are invoked.
Here is the entire process that a usability engineer (UE) could undertake and the parts of the directory structure that are involved in each step. For the purposes of this example, assume the website is named "site-2b-tested", the user session is named "sally", and the tasks are named "browse-store" and "buy-sweater". Note that most of this is done automatically; the purpose here is to show how the various pieces of the Web Metrics system fit together.
Comments [in brackets] explain the purpose of the various sub-directories. Names that appear between $Dollar-signs$ are variables, the others are literal.
STRUCTURE FOR USERS' DIRECTORIES
--------------------------------
As of 2002-April-03: [Best viewed with fixed-width font]
.../ [normal parent of cgi-bin]
cgi-bin/ [user's general-purpose directory for server executables
setenv = CGI_BIN]
webmetrics.env [the single well-known place that knows about
the location of all the other directories]
WebSAT executable(s)
WebCAT executable(s)
WebVIP executable(s)
[other Web Metrics tools are local executables]
.../ [User can put WebMetrics itself pretty much anywhere]
WebMetrics/ [everything else related to Web Metrics
setenv = WEBMET]
WebTools/ [files related to the software itself;
setenv = WEBMET_TOOLS]
resources/ [generic routines and images used by several tools]
images/ [used by local copies of the documentation]
ogl/ [OpenGL utilities]
geom/ [geometry utilities]
c-util/ [low-level C utilities]
gluttk/ [GLUT/Tk material under here, perhaps. Could also
live outside of WebMetrics]
weblink/ [tools to analyze link structure of a website]
linklint [follows links within a website]
convert-ll.perl [analyze, convert linklint output]
WebSAT/ [have at least source and doc for each tool]
source/
doc/
WebCAT/
source/
classes/
doc/
Experiments/
WebVIP/
source/ ??
doc/ ??
start.html [The page from which the WebVIP cgi-bin
scripts are invoked]
FLUD/ [the FLUD script]
FLUDViz/ [FLUDViz executables and source]
VisVIP/ [VisVIP executables and source]
libvip/ [subroutines needed for compilation]
tcl/ [Enhanced Tcl interpreter (Unix only)]
TreeDec/ [TreeDec scripts]
WebLogs/ [FLUD-related experimental results;
setenv = WEBLOG_DATA]
$Website-1$/ [everything related to an *instrumented* website]
...
$Website-n$/
site-2b-tested/ [for example]
userbegin.html [control page for the experiment]
options.html [customized html page for this instrumentation]
website/ [the actual instrumented website]
setup/ [information on how WebVIP instrumented the site]
copysite.txt [URL of the original site that was copied,
if w3mir is used (does not exist otherwise)]
filelist.txt [list of the files to be instrumented]
options.txt [list of interactions to be captured]
pathform.txt [contains the experiment name]
script.txt [contents of the javascript insertion]
summary.html [html composite of the above files]
webstruct/ [files for site structure]
fileF.txt [Linklint output]
url2nn.dat, pages.dat, links.dat [convert-ll output]
node_loc.dat [Used by VisVIP to remember a website layout]
up_color.dat [Used by VisVIP to assign colors to userpaths]
sessions/ [results of test sessions]
$Session-1$/ [all the files resulting from a single
... session go in the same sub-directory]
$Session-n$/
sally/ [Using "sally" as an example of a
session name]
sally.ulog [WebVIP output]
sally.html, sally.parse [parser output]
browse-store-t1.up [navigation paths from the sally
buy-sweater-t2.up session; first two tasks are
browse-store and buy-sweater.]