Hi David, As we said year nr worked. Next try was: vol.,no. 52, [...]year, month, week [...] weekly: 2nd try: vol., no.52,[...]year,month,week[...] weekly: No prediction at all. 3rd try: no., [...] year, week.: Numbers do ok (example nr. 1,2,3), year adds up at every issue, week subtracts 3 at every issue. After 9 issues changes into: season. 4rd try: vol., no. 52[...], week, year [...] weekly: Result: Issuances start at the bottom of the page, nr. do ok, vol. is ok, weeks add up, year same as above. 5th try: vol., year, week; Result: vol. adds up, year also, week subtracts 3. Nr. 10 has become season: summer. What is going wrong?
Regards, Marjolein From: open-ils-general-boun...@list.georgialibraries.org [mailto:open-ils-general-boun...@list.georgialibraries.org] On Behalf Of David Fiander Sent: donderdag 17 februari 2011 14:29 To: Evergreen Discussion Group Subject: Re: [OPEN-ILS-GENERAL] Weekly serials in 2.0 Can you give us some examples of the patterns that you attempted, and details of how they didn't work, like the volume, issue, date information for the issue before and after the real change, and what the predictions said? - David On Thu, Feb 17, 2011 at 06:22, Marjolein Kremer <m...@iisg.nl<mailto:m...@iisg.nl>> wrote: Hi, We have been trying to predict a weekly serial. We tried several options for the enumeration. We always used the option: restarts at unit completion. We did not use a calendar change. We used the option Can compress or expand. We never got a working pattern (either it changed years during the same volume, or it did not even generate the prediction). The only option that worked was Year, nr. Has someone succeeded in making a working "weekly" pattern? Marjolein Kremer IISG Books and Serials Acquisitions
