-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Paul Koning escribió: ...
From: Olson, Arthur David (NIH/NCI) [E]
...
From: Carlos Raúl Perasso [crperasso@gmail.com] Sent: Wednesday, March 31, 2010 2:32 PM To: olsona@lecserver.nci.nih.gov Subject: Proposal to the Paraguayan Government
Hi Olson...
I would like to know, considering the tz datebase update procedure and the frequency at which mayor Linux distros, Java, Oracle and so on replicate this updates, how many days do you think there should be between a rule change and the next date it affects.
Carlos, I lived the experience of the Venezuelan Gov't timezone change from UTC-0400 to UTC-0430 in 2007, which gave us very few months to prepare a new release of tzdata and then push it into the most used Linux distro (Debian) in critical gov't infrastructure (such as the power utility where I used to work) I believe the Sun people responded on a timely fashion for Solaris, yet I don't know how it might be at this moment while they merge. Our Microsoft Windows team had a hard time, patch arrived very late for them to simulate the time change (I assume they don't use Olson's DB any way) and they had long downtimes. Even if the Olson database is pushed on a timely manner, some vendors will give you surprises. After this timezone change, a SCADA system failed to report a switch in a substation. Because they mixed various versions of PLCs, they found out some of them were up-to-date, some of them were not and some of them didn't need a timezone change on the PLC itself since they reported in UTC... Such a decision is not to be taken lightly by any government (take into account we're not used to yearly timezone changes such as DST in most of the LatAm countries) nor it should be taken just with political basis. After this 2007 change, both the Venezuelan gov't and political opposers recommended to change the timezone again in order to save energy. In a country where only the northernmost cities experience an under 1 hour difference in daytime, year-round, moving the timezone again without taking into account a DST measure (which would be unusual for our lat) would be disastrous, some argue that it'd even lead to energy waste. I wish you luck in your considerations for timezone changes in .py, - -- José Miguel Parrella Romero (bureado.com.ve) PGP: 0×88D4B7DF Debian Developer Caracas, VE/Quito, EC -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBCAAGBQJLtCPrAAoJEMAyQqmI1LffZGYP/0fSP0CZSVDcA/2EIstFtfPK Am9b3oEJHut1jtxG568v13jf7x9WJlOB70BoI3LVHZS6zogrtagIHTdsD6H+HBcg F1iW4cEOe4lp4ilKMd876yeFRqu4ZtvUVdZ9TvHCxUWF+KCBunkf9BsWAVIlnQz1 f76mVcWyl7sY4tqp8F03h5RBlH/sSjzvKxPWZZHc5318IpvrRdP879PojgDaKEHJ +Fqh/YXjTgLi7wUuybgjLCxqjp2lWSqF5DPhJQHOYUe3bjo3MD7kqb4My87g223P 3CpHK243LG8GwG0ZECSGs6DH5I9gNNkKhPp/09Z7ld5Ii5ZgQtcQtDraMDnMfMBH /fsszY4aUqXs+KIf88VXlI+rTwbV2eaDKzm9ZJMYAHIPvNfyWXNyENbe12nDLUHH 5fHbeLqfPIspHvP7476ncLzx1fUTNYl8JwXReNSRJtJTklBV21vbTLl9dgeoLxJY pZ8j9a4VUTBHUkJ6vjyju4HQT94O9GnHmh5S3HIwlLXua8p8XVq88E2hWrKpdDQt 7PhcEzVzTkr+9qlyyfX/V+EcVYtYSXMGE5yc8Zk0O2B1PPXBqTrKXn1xaHZHjND3 Te8uFoww1jy9z+b7tLEJGGkHvoQhgClW9VRAKqU3x6kZM5hxcc8kKneC3Ulcm3tF r+gRgc79LmzL65paGweH =XCBd -----END PGP SIGNATURE-----