Yes, I read the note at the time. In fact, I referenced the note in the fixme. The purpose of the fixme was to request that the box was resurveyed, in the hope of a new collection plate with the correct ref being installed in the future.

Hi spiregrain,

It's true that the postbox in this changeset has gone, but it has been relocated further along Upton Lane. I have re-established it here:

node/12954700670

I agree. I’d prefer it if post boxes didn’t have names in OSM. This change was to reformat the name that a previous mapper had already set on the post box, to conform with the other mixed-case names that exist.

https://osm.mathmos.net/postboxes/osm-name.html

I understand your point and I agree to a certain extent. I suppose the real question is, what is the meaning of the network tag in a network relation? Your point of view is that, in this case, it should refer to the over-arching network. That is a reasonable point of view, although it raises the further question of what the network tag should be in a network relation that doesn’t have a higher level network, for example London Underground. I suppose it could be omitted in that case, for consistency. I don’t have any objection if you want to change it, but it might be worth doing nothing until the naming of the London Overground lines come into effect later this year. I suspect that, when that happens, there will be a de-emphasis on the London Overground name and the mapping community might propose that the whole London Overground relation hierarchy is tagged with network=National Rail.

Thanks for your question. There was a debate a few months ago amongst other mappers over whether the network tag for the London Overground lines should be London Overground or National Rail. A similar debate took place for the Elizabeth line. The outcome of the London Overground debate resulted in the route master and route relations network tags being set to London Overground (by other mappers), although it is interesting to note that the debate over the Elizabeth line went the other way. The rationale for my changeset is to simply set the network tag in the London Overground network relation to match the network tag in the child route master and route relations, for consistency. So, should the same rationale apply to the network tag in the network relation of the other rail lines in the UK? Yes, I think it should, but presumably they are already set to National Rail, as their child route master and route relations are presumably also set to National Rail.

At the time, it looked like a school site that was being redeveloped for residential use (although the imagery may have been a bit out of date). I think it is probably now completed for residential use, and so the area can be reclassified as residential, and the addr: tags can be moved from the area to the building(s).

Thank you for the comments. I have found that PO Box addresses are not often mapped in OSM, but when they are mapped, they invariably exist at the delivery point. It seems wrong and misleading to use addr: to map a PO Box Address at a location that isn’t the delivery point (in this case mapping a Gloucester delivery point in Cheltenham). Also, in this particular case, the postcode of GL1 2XG appears to be an invalid postcode, albeit the court advertises this postcode as part of its PO Box address. However, I take your point about map users possibly viewing this node and not appreciating that the court has advertised a PO Box address for its address to receive post. So, I have revised the note tag to orient it towards map users and to echo the PO Box address, albeit I still think the postcode is wrong.

Hi,
Are you sure about the postcode change from CF14 5DU to CF14 5YP for Marks & Spencer in this changeset? I can see that CF14 5DU is the postcode specified on the M&S web site but it is not in Code Point Open and it appears to be a postcode that was made obsolete in 2002. M&S is using CF14 5YP in its FHRS entry. So, I think the M&S web site might be wrong. For those reasons, I think the correct postcode is CF14 5DU. Could you please have a look at this again to see if the postcode in OSM should be changed back to CF14 5DU. Thanks.

Postcode BA2 4AB was made obsolete in 2020. The hotel has probably not changed its web site and FHRS entry yet. I have seen this type of thing happen quite often, where Royal Mail re-assigns a different postcode to an address but the organisation continues to publicise the old postcode, sometimes for many years.

I removed BA2 2QL because it's an obsolete postcode. I couldn't determine an alternative and that suggests to me that the entities that had that postcode are probably disused.

Welcome to OpenStreetMap. I have noticed that the two postcodes that you added to 10 Park Drive were already there (before your edits, addr:postcode=E14 9GD;E14 9JX;E14 9QY;E14 9ZW). I think the editor might have misled you into thinking that some postcodes were missing, because the display tends to only show a portion of a multi-valued postcode field. I have removed the duplicate postcodes. I hope this false start doesn’t put you off from making further edits to this and other OpenStreetMap elements in the future. Regards Mike

The source is code-point open. In this case, there were some partial postcodes already in place for SY4 5UP. I am working on trying to clear all of the partial postcodes in the UK. Here I added the missing elements for SY4 5UP and also added postcode SY4 5TW for completeness of Somerset Way. See https://osm.mathmos.net/addresses/pc-stats/SY/SY4/5/ (which should show these postcode centroids in green tomorrow, after the nightly update).

Regards

Mike

Thanks for the comment/question. The general answer is yes, I am relying on the CodePointOpen dataset. In the past, I’ve also relied on the address data from FHRS, although that is not as easy to utilise as it used to be, due to the way the daily reporting of mismatched data has changed. Occasionally I might get an address in my local area from social media, e.g. where a new fast food outlet announces their arrival but they haven’t yet registered with FHRS. Lastly, I occasionally obtain address data from an organisation’s web site, e,g. where I recently amended the postcode for School 360 at E15 2QS.

No objections, I agree with that proposal. I was probably a bit too heavy handed in changing the building tags on those objects.

Hi Ken, Recently I've been working to try to clear the street name warnings in the E postal area report (see https://osm.mathmos.net/addresses/street-warnings/E.html). That building was highlighted in the report as an address with a single postcode (E13 0JN) but with a missing street name. I added the addr:street tag based on the postcode. I didn't spot that the building had addresses associated with its entrances, because those addresses don't include postcodes, and so don't appear on the report. If a building has multiple postcodes then it is legitimate to include all of those postcodes. In this case, the addr:postcode should become "E13 0JN;E13 0JU". But that would create the dilemma that you describe, because the building would then have postcodes with different street addresses (Stratford Road and Maud Road). There are some buildings in the E postal area mapped with multiple postcodes (see https://osm.mathmos.net/addresses/pc-errors.html) but, as far as I know, none of them have postcodes covering different streets. In the case of this building, and any others like it, I think probably the best thing to do is to remove addr:postcode and addr:street from the building and add the addr:postcode (and addr:city) tag to each entrance. Regards Mike

Thanks for the feedback. You may have seen that I reinstated the three-dots placeholder in the partial postcodes. I thought I might then consider whether that is the best way of flagging a partial postcode that needs fixing (of which there are at least 10,000 according to the OSM GB Postcode Errors report) or whether a fixme tag or a note might be a better way of doing it. But since then, another mapper has updated the properties with the full postcodes, which is obviously the best final solution.

Fair enough. I'll switch them all to caterers for consistency.

Yes, they are so-called dark kitchens. They exist to service the fast-food delivery sector. There are a few clusters of dark kitchens in various parts of London. Typically, they are in temporary units of disused car parks or in railway arch premises. There is an argument for tagging these type of outlets as either fast food or caterer but I’m usually guided by how they have registered themselves with FHRS.

Hi, could you please take a look at the mapped position of Columbia Primary School, because I believe it is currently in the wrong place. I think the mapped site at the junction of Columbia Road and Pelter Street should be Columbia Market Nursery School, not Columbia Primary School. Columbia Primary School is, I believe, further east along Columbia Road, in the area of Ravenscroft Street / Ezra Street. Thanks.

The current postcode of RH16 8LL for the Cricket Pavilion in Balcombe is being flagged in the daily postcode errors report (https://osm.mathmos.net/addresses/pc-errors.html). The web site for the Balcombe Cricket Club specifies it as RH17 6PA. So, I suggest that is the correct postcode.