DALI 1.1 comments

[Rob Seaman]

On May 11, 2016, at 5:05 AM, Paul Harrison <paul.harrison at manchester.ac.uk> wrote:

> On 2016-05 -10, at 21:10, Arnold Rots <arots at cfa.harvard.edu <mailto:arots at cfa.harvard.edu>> wrote:

> 
>> I don't have the text of ISO-8601 handy here, but my recollection
>> was that it is silent on what a time stamp without explicit time zone
>> information means.
> 
> I do not have a copy of the text of ISO-8601 either, but many authoritative sources make this interpretation clear e.g. https://www.cl.cam.ac.uk/~mgk25/iso-time.html#zone <https://www.cl.cam.ac.uk/~mgk25/iso-time.html#zone> , https://en.wikipedia.org/wiki/ISO_8601#Times <https://en.wikipedia.org/wiki/ISO_8601#Times> and that is how java https://docs.oracle.com/javase/8/docs/api/java/time/package-summary.html <https://docs.oracle.com/javase/8/docs/api/java/time/package-summary.html> interprets strings for instance and http://tools.ietf.org/html/rfc3339 <http://tools.ietf.org/html/rfc3339> recommends UTC but adds the Z to be explicit.


IVOA is a standards organization, and reports to Comm-B2 of the IAU, itself a member of the International Council for Science (ICSU). ISO is an independent entity. Which is all to say that IVOA should expect to invest in purchasing necessary external standards as needed. If IVOA recommends that astronomers should use an ISO format, members of the pertinent working groups should have copies. While it cannot be the policy of IVOA to sidestep paywalls, copies of many prominent documents have escaped into the wild, e.g.:

	http://www.uai.cl/images/sitio/biblioteca/citas/ISO_8601_2004en.pdf

or may be available through university libraries, etc. Failing that, surely the recurring use astronomers have made of ISO-8601 over the past two decades would justify simply paying the IP bounty of 138 CHF:

	http://www.iso.org/iso/catalogue_detail?csnumber=40874

IIRC, the price was significantly less during the run-up to Y2K (though my old copy has been superseded).

> There are a wide range of existing software systems (e.g. language libraries, databases) that use the strict interpretation of ISO-8601 which mean that the timestamp without will be interpreted incorrectly unless the system timezone is set to UTC, and while I can imagine that many astronomical servers around the world do have their timezones set to UTC, I doubt that all astronomers laptops are similarly set with their timezone to UTC.

I’m not sure that is precisely what is required by the witches’ brew of standards, but Steve’s observation remains:

"The standards bodies responsible for computer time representations
explicitly disregarded advice to be able to handle precise time,
so anyone trying to do that must diregard the standards."

In other words, attempting to adhere to standards such as POSIX (http://pubs.opengroup.org/onlinepubs/9699919799/) may require rejecting other, perhaps more fundamental, standards. There is no self-consistent choice, other than the surprisingly controversial step of promoting real world phenomena to be ultimate arbiter.

> This issue is why I made the suggestion earlier in this thread that only ISO 8601 be used only for the case of UTC timescale and be forbidden for other timescales - then the Z serves the double purpose of being a designator of UTC timezone and time scale whilst keeping compatibility with non-astronomical ISO 8601 usage.

One is at a loss how to convince users that “2016”, for instance, a compliant ISO-8601 specification, can only be deemed to have started at Greenwich midnight.

Rob
seaman at lpl.arizona.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dal/attachments/20160511/8bb29f4e/attachment.html>