One user has made 34 edits (from ch 176861218 5Jan 1pm, to ch 176956303 7Jan 6:30pm) with widespread significant changes across much of the Barbican estate, including but not limited to
- the deletion of correctly mapped buildings (incl entire west wing of school), building parts incl 3d info, named pedestrian areas, footways, walls, lake and garden features;
- the deletion then redrawing anew of some features incl buildings, footways, steps, walls etc, and arbitrary and not necessarily accurate switching between area and node - in both directions - for some amenities, and including loss of mapped tag info including heights, building levels, step incline directions etc
- damage to geometry including pulling nodes aside (poss intended temporary but never fixed) and moving excellently aligned nodes to non-vertical, non-aligned Bing superficial positions, including some nodes and not others where precise 3d mapping had been done
- the replacement of building with (sometimes roughly defined) man_made=bridge areas where the topmost surface is walkable, regardless of what is underneath
- causing building-inside-building issues where was previously carefully handled with building:part, address/entrance nodes etc
- causing multiple fix changesets to be required for duplicate ways, coincident nodes, self intersection, and for addresses without eg addr:city (but also missing many other details ie just putting housenumber and nothing else)
- breaking a foot route (including removing ways and junctions that exist that make up part of that route)
- has left gaps with no structure at all mapped (building or their âbridgeâ areas) eg N of arts centre, messed up NW corner of Guildhall Sch. of Mus.&Drama, the now âunexplainedâ covered or tunnel status of roads such as W end of Beech Street and S end of Bridgewater Street, etc
- changeset descriptions do not describe the content, all are just âBarbican Estateâ, with typically just Bing imagery reported used and no other notes on sources etc provided, and so giving no clue of the changeset intent and if eg any deletions were accidental
- used inapplicable man_made=chimney on 8 hot air vents (top of elec. substation - presumably transformer etc cooling vents) giving spectacularly wrong 3d rendering at f4map:
F4map Demo - Interactive 3D map
where you can also see âz-fightingâ where 3 buildings are now double-mapped and lack of 3d detail on school (and lack of school W wing !), and height on W block etc. Can also see how the W tower has had its shape messed up (the 3 tall tower are similar shapes, but rotated, to each other).
Positive changes are (the only ones found, subject to further review):
- updating details for 1 restaurant
- dividing building block terraces into separate houses for 3 blocks (but causing building-in-building issues in order to get block name to render)
For those not familiar with the Barbican, it is hugely multi-level and complex place. The view from above and what is present at the uppermost level does not tell all about buildings, paths and roadways underneath. For reference (and amusement / striking fear into any mapper) you can review the sectioned drawings visible at
and the graphic at the top of the page at
There is one legitimate question to be resolved: for an elevated surface which is traversable above (eg walkways) in those cases where there is little below, where to draw the line between man_made=bridge vs building[=roof] :
Clearly where there is an enclosed space (with walls) such as the Silk Street entrance (see 3rd picture at The Barbican: London's Brutalist masterpiece - ICON Magazine )
or the Barbican Hall (the concert hall !!), then there is a building and deleting these and replacing with man_made=bridge is destructive enough to be vandalism.
Over the main road Beech Street, see https://www.geograph.org.uk/photo/2927901 there is a narrow strip with (mostly) only a single layer of foot-accessible top surface which in isolation could be a bridge (although the main footway is mainly parallel to, not crossing, the roadway under), but in the context of the Barbican is part of the overall structure of the whole.
But there are other large areas now mapped as âbridgeâ where there is at very least a full height wall enclosing a space under the top surface, if not actual âindoorâ space(s) at various levels, where the previous podium building was appropriate and, I think, acceptable and preferable rather than marking the structure as a man_made=bridge (with, as current, âentranceâ nodes now not connected to any building).
The user needs education on good changeset comments, preservation of object history, respect for other mappers data (incl features such as 3d, walking routes etc that they might not be interested in at the moment), handing of and making allowance for non-vertical imagery, image alignment, etc.
I would really like to revert ALL of these changes (and carefully reapply the very few positive changes) to restore the deleted buildings, damaged alignments and 3d information, as well as history and connectivity of perfectly well mapped objects, restore the lake edge features etc etc.
Reversion is currently just possible, with careful handling of the relevant overlapping fix changesets along the way:
3x BCNorwich geom error corrections caused by this user
3x rskedgell addr:city missing - noting changesets have larger scope so selective reversion or just manual resolution is needed of the incomplete addr: objects added by this user
4x SomeoneElse2 named foot route repair (1 as consequence of changes here, 3 others just version changes with changes in other locations not touched by these edits)
For that named foot route through the Barbican, there exists (and was mapped) an elevated (âHighwalkâ) route across Gilbert Bridge and continuing N (just passing the theatre fly tower) to a junction where left leads to Frobisher Crescent (upper public walkway) and right connects to Speed Highwalk. This junction and
multiple connections are currently deleted/broken - see changeset comment on 176951607
Can others take a look at this sequence of changesets, noting the level of deletion (and sometimes but not always redrawing) where I believe there was very little wrong (subject to the difficulty of mapping such a complex 3d place) with anything previously mapped. Note that building alignment had been done
(across most/all of the area) using Cadastral-aligned Mapbox imagery which (prior to the Julyâ25 update) was superbly vertical in London; this quality of alignment cannot be directly reproduced now without reference to the previous/reverted node positions.
I will attempt to post (separately below) the list of changeset numbers, showing the [new;modified;deleted] object counts and some quick notes of the content for reference, also the ~10 other user changeset numbers that will trip anyone up if they want to step back through a series of reverts as I did (I have successfully stepped back to the starting point and locally saved an .osm for reference)
If there are other positive changes that should be kept, it would be useful if those could be noted below.
And please, can I suggest no âhelpfulâ fixing of any issues in response to reviewing these changes until it is decided what to do, noting that this will complicate and could possibly prevent reverting of any/all if that is decided as the best option.
Cebderby (Clive)
|