home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!howland.reston.ans.net!wupost!psuvax1!psuvm!auvm!UCS.INDIANA.EDU!MORGANJ
- Message-ID: <CDROMLAN%93010514094839@IDBSU.BITNET>
- Newsgroups: bit.listserv.cdromlan
- Date: Tue, 5 Jan 1993 16:05:49 EST
- Sender: "CDROMLAN@IDBSU - Use of CDROM Products in Lan Environments"
- <CDROMLAN@IDBSU.BITNET>
- From: James Jay Morgan <MORGANJ@UCS.INDIANA.EDU>
- Subject: Plusnet2 statistics with Bindview
- Lines: 62
-
- This is cross-posted to CDROMLAN, CDPLUS-L and Medlib-l.
-
- In December I reported on CDROMLAN@IDBSU that because of the
- limitations of the CDPStats statistics module in Plusnet2
- systems, I was using Bindview to generate user statistics. Since
- this has been mentioned on CDPLUS-L and Medlib-L I thought I
- would post my latest findings to all three lists.
-
- From late July through December CDPSTATS shows a total of 6,309
- uses of our system. Bindview for the same period shows 16,463.
- Because of the way we administer our system, both figures
- underestimate usage, but the Bindview figure is more accurate.
-
- Using CDPStats we can generate detailed information about
- searches, and statistics from survey questions. However the
- module does not record data unless the user closes the session
- normally. Judging from the figures above, 38% of our users close
- the session normally. Most leave in mid-session and the session
- either times out or is continued by another user.
-
- We have open terminals with no passwords or charges within the
- library and the Learning Resouce Center, so the users have no
- incentive to close the session. If your system does give the
- users an incentive to close the session properly, the CDPStats
- statistics will be much more accurate. So far I've heard of 3
- incentives, 1) charging for use, 2) giving users passwords good
- only for a fixed amount of time, and 3) withholding printed
- results until the user closes the session. Each of these
- incentives would increase the number of properly closed sessions.
-
-
- Using Bindview we can total the number of logins by month, group
- and by user or station. This is still an undercount, because if
- another user continues an open session, that user will not be
- counted. We have set the timeout at 3 minutes, but during busy
- periods we have people waiting and terminals seldom timeout. In
- addition the CDPLUS timeout doesn't work in some situations, such
- as when the user leaves the station in browse mode and in some of
- the other informational screens. Some of the logins represent
- use by system administrators rather than users, but these can be
- easily subtracted from the total.
-
- I plan to do some observation in January to try to estimate the
- undercount during busy and during quiet periods. Right now I
- would guess that the undercount during a busy hour could be as
- much as 150%, since fewer than 40% of our users close their
- session. Of course such busy periods are limited to weekday
- afternoons, and the rest of the day the number of uncounted users
- would be lower. It would probably related the the percentage of
- time users leave the terminals in a no timeout mode. Overall the
- actual number of users is probably 20-30% higher than our
- Bindview login count.
-
- I would be happy to hear from anyone else who has worked on a
- formula for estimating use from the number of logins, or other
- methods of measuring use.
-
-
- Jim Morgan
- RUTH LILLY MEDICAL LIBRARY
- INDIANA UNIVERSITY SCHOOL OF MEDICINE
- MORGANJ@INDYVAX.BITNET MORGANJ@INDYVAX.IUPUI.EDU
-