On 2016-10-20 15:27, Pierpaolo Bernardi wrote:
about Italian time, I see: # From Paul Eggert (2006-03-22): # For Italian DST we have three sources: Shanks & Pottenger, Whitman, and # F. Pollastri # Day-light Saving Time in Italy (2006-02-03) # http://toi.iriti.cnr.it/uk/ienitlt.html # ('FP' below), taken from an Italian National Electrotechnical Institute # publication. When the three sources disagree, guess who's right, as follows: I don't know about the other sources, but: - the url given to cnr.it is dead. - The authoritative source for time in Italy is the national metrological institute, which has a summary page of historical DST data at http://www.inrim.it/res/tf/ora_legale_i.shtml - Some of the data in TZ does not agree with this page. I think that there must be VERY strong evidence that the INRIM is wrong before preferring other sources?
Hi Pierpaolo, That page does not cite any of its sources: - some of them could be those used in the tzdb or the tzdb itself; - tzdb prefers official government sources as references, better if viewable online; - perhaps you could contact INRIM to find out if they can provide references to information of which the list was not previously aware? Also perhaps you (or they) could say what the phrases: "dalle ore X del" and "alle ore X estive del" mean in this context? Google Translate says they mean something like: "from hour X" and "from summer time hour X", implying local (standard or daylight) time is used in both cases. I have attached comparisons of the data from the table on the web page, converted by scripts to tzdb rules, flagged with i in the Type column relabelled Src, and conversions of those date/times showing 24:00 to the next date 00:00, flagged with =, both with the day of the week appended to the day in the month, for easier checking. Those were merged with tzdb Italy rules, flagged with z. Comparison was done visually and entries where there are discrepancies have a ? appended to the end of the tzdb line; sometimes others where I thought data may be questionable. Some assumptions were made in the comparison: - web page times are local standard time for daylight start entries and local daylight time for daylight end entries; - daylight start entries with the same time value with or without s (standard time suffix) are the same; - daylight end entries with local time an hour later than an entry in standard time are the same: entries with the same time value with and without s (standard time suffix) differ; - there may be more/fewer discrepancies if these assumptions are wrong. Some tabs between leading columns were converted to spaces to shrink line length for attachment. If you contact INRIM, please feel free to forward the comparison. Scripts used and intermediate files are available by email if anyone cares. -- Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada