silver@csu-cs.UUCP (09/12/83)
Just for fun, I gathered the following statistics on the news system at csu-cs. The background job ran for almost three (!) days. The algorithm went something like this: Find every ordinary file under /usr/spool/news. For each file, use awk to note the last line matching "^Posted:" (news format date line) or "!.* ..:..:.. 19..$" (notes format), and print whatever it got (maybe a blank line). Notes articles usually contain a line of the latter form (the true posting time) after the previous form (the gateway time). Here are various analyses of the statistics. I'm not pretending that this is a scientifically valid survey. For starters, it's no more accurate than the clocks on the posters's machines. Please, no flames on that account! (However, I still have the raw data if someone thinks there's another way to look at it.) Alan Silverstein, Hewlett-Packard Fort Collins Systems Division, Col ucbvax!hplabs!hpfcla!ajs, 303-226-3800 x3053, N 40 31'31" W 105 00'43" OVERALL STATISTICS: Count Description ===== ================================================= 5475 total lines of output (one line per file) - 114 blank lines (no date in file) - 31 lines containing "From" (not expected format) - 2 lines with wierd posting dates ===== 5328 total lines after tossing junk 4961 from news (93%) 367 from notes ( 7%) MONTH AND YEAR POSTED (reflects ad hoc expiration history): Count Mon Year Perct Rank ===== ======== ===== ==== 1 May 1982 0.0% 16 26 Jul 1982 0.5% 15 64 Aug 1982 1.2% 14 81 Sep 1982 1.5% 13 136 Oct 1982 2.6% 12 196 Nov 1982 3.7% 11 393 Dec 1982 7.4% 7 240 Jan 1983 4.5% 9 225 Feb 1983 4.2% 10 259 Mar 1983 4.9% 8 427 Apr 1983 8.0% 5 437 May 1983 8.2% 4 464 Jun 1983 8.7% 3 422 Jul 1983 7.9% 6 1274 Aug 1983 23.9% 1 683 Sep 1983 12.8% 2 ===== 5328 DAY OF MONTH POSTED: Count Day Perct Rank ===== === ===== ==== 339 1 6.4% 2 370 2 6.9% 1 208 3 3.9% 4 167 4 3.1% 14 161 5 3.0% 17 140 6 2.6% 22 149 7 2.8% 20 169 8 3.2% 13 116 9 2.2% 30 134 10 2.5% 27 139 11 2.6% 24 138 12 2.6% 26 138 13 2.6% 25 101 14 1.9% 31 150 15 2.8% 19 141 16 2.6% 21 173 17 3.2% 11 166 18 3.1% 16 169 19 3.2% 12 176 20 3.3% 8 166 21 3.1% 15 194 22 3.6% 5 177 23 3.3% 7 132 24 2.5% 29 175 25 3.3% 9 150 26 2.8% 18 132 27 2.5% 28 139 28 2.6% 23 173 29 3.2% 10 187 30 3.5% 6 259 31 4.9% 3 ===== 5328 DAY OF WEEK POSTED (easily available from news articles only): Count Day Perct Rank ===== === ===== ==== 724 Mon 14.6% 4 711 Tue 14.3% 5 873 Wed 17.6% 3 922 Thu 18.6% 1 911 Fri 18.4% 2 405 Sat 8.2% 7 415 Sun 8.4% 6 ===== 4961 HOUR OF DAY POSTED: Count Hr Perct Rank ===== == ===== ==== 177 00 3.3% 17 93 01 1.7% 21 107 02 2.0% 19 70 03 1.3% 22 65 04 1.2% 23 54 05 1.0% 24 94 06 1.8% 20 141 07 2.6% 18 226 08 4.2% 13 308 09 5.8% 8 355 10 6.7% 2 363 11 6.8% 1 322 12 6.0% 5 320 13 6.0% 6 336 14 6.3% 3 335 15 6.3% 4 307 16 5.8% 9 313 17 5.9% 7 218 18 4.1% 15 227 19 4.3% 12 211 20 4.0% 16 219 21 4.1% 14 238 22 4.5% 10 229 23 4.3% 11 ===== 5328