Re: [Tzdist] [tz] Timezone in Brazil

[Lester Caine <[e-mail]>]

On 12/10/15 00:40, Paul Eggert wrote:
> [e-mail] wrote:
>> But you are always gambling that the client machine has actually
>> compiled WITH backzone? If it is missing then data prior to 1970 is
>> suspect.
> 
> If backzone is used to build the tzdist server's data, then the server
> contains more data, some good and some questionable.  Some of the good
> data is before 1970, and some after.  Some of the questionable data is
> before 1970, and some after.  In this sense, it's just the same as when
> backzone is not used.  So there's no good reason to require
> implementations to reject time stamps before 1970 merely because
> backzone is being used (or is not being used).
> 
> I get the sense that you're trying to shoehorn backzone into a slot that
> it's not designed for.  It appears that you want some indication of how
> reliable data are.  There is no such indication anywhere in the formal
> part of tzdata, nor in tzdist, and this is true regardless of whether
> backzone is used.

But all YOU are saying there is exactly what *I* am saying ... you can't
use tzdb for reliable conversions prior to 1970! ... because the data is
unreliable. There IS a good reason to reject if it is important that the
correct data is used for normalizing. In my own case I need some way to
ensure that backzone IS in use by clients since this provide a lot more
correct set of rules than a generic result, but as you say ... there is
simply no way of knowing just what the client has PRIOR TO 1970! It is a
gamble when it should clearly identify the source ... so that even
faulty results can be reliably recreated.

I would prefer that the backzone data was ALWAYS included and then one
can guarantee that the client has a complete set of rules, many of which
are more accurate prior to 1970 than if that additional data is simply
ignored. What would be safer is if the extra UK aliases simply did not
exist if they are not returning the data that they should be!

-- 
Lester Caine - G8HFL
-----------------------------
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk