Table of Contents
Since 15APR98, AIPS has used the EGCS version of the GNU compilers
with great success, mainly as it gives a considerable performance
boost over the older gcc/f2c package. We now require use of the EGCS
compilers for compiling AIPS with
The final report on 31DEC02 is in the AIPSLetter dated 31 December 2002 (PS, 2906 Kbytes); also available gzipped, 1140 Kbytes). See also 31DEC02 AIPSLetter dated 30 June 2002 (PS, 753 Kbytes) for information on changes before about that date. It is also available gzipped, 320 Kbytes).
The final report on 31DEC01 is in the AIPSLetter dated 31 December 2001 (PS, 742 Kbytes); also available gzipped, 316 Kbytes). See 31DEC01 AIPSLetter dated 30 June 2001 (PS, 3900 Kbytes) for information on changes before about that date. It is also available gzipped, 966 Kbytes).31DEC00 AIPSLetter dated 1 April 2000 (PS, 156 Kbytes) for information on changes before about 1 April 2000. It is also available gzipped, 65 Kbytes).
Courtesy of Ernie Allen: see the various AIPSletters referenced above for detailed reports.
As of December 30, 2001 556 copies of 31DEC01 had been distributed, all by ftp. Of these, 448 appeared to be separate "sites" (including multiple computers at the same university however). 48 of these sites took full copies more than once during the year. Statistics for 2002 are unavailable. Our MNJ monitor suggests that 48 sites/computers ran MNJs on 31DEC02 during its development.
A tool to monitor shipments of the tar ball was finally installed in Spring 2003. We shipped the 31DEC02 tar ball to 88 "sites" (unique IP addresses) in the last 30 weeks of this calendar year. A plot of sites that have had at least one cvs access and of sites that have copied at least one of the tar balls is generated each week. The cumulative totals are quite significant. The cvs access implies that the site installed the 31DEC03 version of AIPS (MAKE.MNJ is not run while installing the frozen version) and/or ran the MNJ for 31DEC03. The tar-ball downloads count only downloads that completed successfully. In each of the totals, an individual IP address is counted only once. In general, an IP address is used by only one user, but more than one IP address will be used by those users that use dial-up connections. Thus, these totals are a modest over-estimate. Note, the totals also include multiple computers from a single institution such as NRAO.
Modified on $Date: 2012/01/03 16:01:48 $
[Eric W. Greisen]