Re: "changed" field should be deleted

[Dale Amon as Operator <]

On Sat, 6 Jun 1998, David Kessens wrote:
> But people that are querying your data are coming from different time
> zones, whether you like it or not. My database 'person:' object is
> stored in the RIPE database from the times when I was still working at
> the RIPE NCC. Now, I have a 9 hour time difference. The working area
> of the RIPE NCC is quite big and it does go over many time zone
> boundaries. Replies regarding the 'changed:' attribute discussion came
> from people at the US east coast, west coast, new zealand and europe ...
> 

I guess I was questioning whether there is a reason for someone else
to have this information. I guess there are reasons
why someone trackingdown a routing problem might be
interested in when something changed. But if that is the case, then
why is the time signature generated at source? I should send a
changed: line with an id-code of some sort, which could be company
internal; the database should generate a UTC time sig at the
instant of the update. Otherwise the date sig is pretty pointless.

I can personally remember working out a large database change in a file
over a couple days before sending it. So that dates were when I
generated the data, not when I updated the database. Well, I may
have fixed them, but you get the point.