[Talk-GB] TfL Cycle Infrastructure Database (CID) conflation

[David Woolley]

On 06/01/2023 10:30, Steven Hirschorn wrote:
> I realigned roads and buildings with the Cadastral layer a few days 
> earlier and wonder if the CID import is working on cached data?

The first reported case involves duplication of recently added node, but 
newly added ones.  The normal double update checks should have failed 
that update, because the road version number would have changed. To me 
that means the data must have been refreshed immediately before the bad 
update.

Whilst they could have been misusing a refresh, it looks more like they 
didn't check for existing nearby features and added based on the 
coordinates in their source data without applying local map offsets.

I'm concerned that either the user interface failed to show the effect, 
or the user was totally unaware of the real world significance of what 
they were doing.

Another possibility is that the user did get a version error, but cured 
the symptom, by refreshing, rather than checking for the cause. That 
again suggests a lack of real world understanding.