lets start!

[justinelliotmeyers]

can we start to build this!!?
Thanks
Justin

[Cj-Malone]

👍

[Cj-Malone]

@oboswarva is pushing individual councils in the right direction.

https://www.owenboswarva.com/blog/post-addr11.htm

[oboswarva]

I've made a landing page with links to the Council Tax address lists that have been released as open data so far: https://www.datadaptive.com/addr/

I have more requests in the pipeline and I'll update the page as additional datasets become available.

(This effort is necessarily piecemeal though, and not a viable approach to delivery of an open national address file – that still needs political intervention.)

[Cj-Malone]

I have added the councils data to some Open Data address points I've been computing.

https://osm-uk.github.io/exports/posts/2025-10-19-os-open-data-apt/

This is mainly to show people how much can be computed, (coords, street, "city", postcode) and is not meant as an alternative to proprietary datasets. However I believe soon Open Data will rival these datasets making them a lot less interesting in the majority of use cases.

[regulate-tech]

Thanks for kicking this thread up and for the resources. Feels like a good time to restart work on the political side with @peterkwells. I'm also interested in the potential of the broad range of delivery services who now visit UK properties as potential address verifiers. Any insights into how they are collecting and using address data would be welcome.

[Cj-Malone]

I've signed an NDA with a dynamic parcel distributor (a long time ago), so I'm unsure exactly what I can say. However I'm sure we are all aware how delivery services work, the drivers are required to use an app, that app is on a phone with GPS, that phone is used while moving, and attempts to deliver a package are documented in an app with some metadata available.

Some delivery services don't use PAF/AddressBase so there addresses aren't encumbered with licensing requirements.

[oboswarva]

I've signed an NDA with a dynamic parcel distributor (a long time ago), so I'm unsure exactly what I can say. However I'm sure we are all aware how delivery services work, the drivers are required to use an app, that app is on a phone with GPS, that phone is used while moving, and attempts to deliver a package are documented in an app with some metadata available.

Some delivery services don't use PAF/AddressBase so there addresses aren't encumbered with licensing requirements.

That sounds like a parallel supply chain for collection and verification of address data, which is great, but presumably the delivery service will acquire IP rights in the collated data in much the same way as Royal Mail Group does when posties contribute updates to PAF. So there will still be licensing requirements for re-use unless the delivery service generously decides to dedicate the data to the public domain?

[Cj-Malone]

If I were a delivery company, my goals would be delivering packages quickly and efficiently.

These companies don't have proprietary road networks, they use (and some contribute to) OSM. Not OS, these are global companies + OSM is updated minutely. Addresses are the same, it's not a competitive advantage, it's the bare minimum.

The data will be noisy, fuzzy gps, typeos, inconsistent formatting and incomplete. Trying to sell it to companies already using OS is a waste of time. Giving this data to OSM UK to be cleaned and used in OSM, and then made available to there drivers improves the flow.

[Cj-Malone]

If OS is losing customers by the time PSGA is renegotiated, OS will be begging the government to buy a global license and release it as Open Data.

[oboswarva]

If I were a delivery company, my goals would be delivering packages quickly and efficiently.

These companies don't have proprietary road networks, they use (and some contribute to) OSM. Not OS, these are global companies + OSM is updated minutely. Addresses are the same, it's not a competitive advantage, it's the bare minimum.

The data will be noisy, fuzzy gps, typeos, inconsistent formatting and incomplete. Trying to sell it to companies already using OS is a waste of time. Giving this data to OSM UK to be cleaned and used in OSM, and then made available to there drivers improves the flow.

But IP rights still attach automatically so there must still be licensing issues to work through. Funnelling address data through OSM as the first/only point of release is fine for the purposes of OSM but presents a barrier for other open re-use cases, because OSM's standard open licence (ODbL) contains a share-alike condition that will present difficulties for many commercial re-users of bulk address data.

[oboswarva]

I can see a future scenario where Government might want to empower an aggregator to collate and standardise address data from delivery companies (mandated using the "smart data" powers in the DUA Act) if/when the Universal Service Obligation breaks down, but without a political steer towards open they would likely just use that to create commercial address products on the existing model.

[Cj-Malone]

But IP rights still attach automatically so there must still be licensing issues to work through. Funnelling address data through OSM as the first/only point of release is fine for the purposes of OSM but presents a barrier for other open re-use cases, because OSM's standard open licence (ODbL) contains a share-alike condition that will present difficulties for many commercial re-users of bulk address data.

I'm not suggesting OSM addresses should be the only goal. I believe having great OSM addresses is the starting point for OGL AddressBase. There is already tooling for editing OSM, there is already tools for geocoding with OSM, these companies are already using ODbL, and OSMF has already got guidelines on not hitting share alike in regards to geocoding.

I do think the government should release OGL addresses, I think it should have done it 2 decades ago. I don't think political discussions alone are going to do it, or it would have already happened. They need a push, they need an technical alternative that people can use. I think that should be OSM, not a new thing.

[oboswarva]

Geocoding isn't the issue. Address data is a mature market in the UK – there's a whole body of existing practice including case law related to the licensing of address data that focuses on rights that arise when addresses are cleansed and standardised. Using OSM for that is less bad than using AB from an open perspective but OSM addresses are not a credible starting point for an "OGL AddressBase" because ODbL isn't forward-compatible with OGL. And, licensing issues aside, OSM activity of this nature won't scale to produce a technical alternative (i.e. a open national address file, regularly updated) that's viable for anything but hobbyist use cases.

[regulate-tech]

A key plank of the historic argument for PAF was that Royal Mail delivery staff were the only people going to most properties frequently. We now have multiple entities touching properties daily with their own commercial incentives to verify addresses and record additional useful data about them. If our interest is in the most accurate and up-to-date information then finding a way pool the information from these visits seems worth exploring. This would mean addressing IP and commercial questions as well as developing appropriate quality control etc processes. Non-trivial but not impossible. Commercial entities might anyway build a "PAF 2.0" to monetize their address knowledge following the same logic as Royal Mail and worth considering how this could be steered in an open direction.

[Cj-Malone]

OSM addresses are not a credible starting point for an "OGL AddressBase"

I'm not suggesting they are. I'm suggesting OS will eventually release the existing AddressBase as OGL. Once the government pays them to do so.