On 12/19/2014 08:49 AM, Martin Burnicki wrote:
On the other hand, the intention of the file should be to use it, or the information it provides, so why should there a problem extracting data from it using a script and feeding the result to TZ DB and/or tzdist?
Yes, we can do that in the United States. But we cannot reproduce the input file from the IERS; all we can do is reproduce the output data. Which, effectively, is what we're doing now, using the NIST file as an intermediary. From our point of view it would be better if the IERS's file were explicitly placed in the public domain or under a free-software license like the GPL or the BSD license, so that we could distribute the file as part of the tz distribution.
Amusingly enough, the NIST file's expiration date disagrees with the IERS's. As a practical matter the NIST's date is more conservative and is a better choice for applications like tz and tzdist and this is another argument for using the NIST file.
Agreed. If the expiration date is only e.g. 2 days before the next potential leap second then alerting may be too late, except if the application alerts let's say 2 weeks (or whatever) before the current date reaches the expiration date.
Perhaps someday the IERS will address these two concerns.
Shall I contact the IERS folks with regard to this?
Sure, please feel free; you can cite this thread.