Yes, no problem

That’ll be wall_loops. This is 13 years old, presumably pre-dating that on map features. There may be others around Cambridge. We know them as v-grips locally.

FYI I am going to add back 'porters' to entrance here as without it the porters lodge disappears from the University Map (https://map.cam.ac.uk)

Building restored and attributed to new operator in changeset osm.org/changeset/79337521
@Hilton Hotels - just because you gave up the business was no reason to delete the building. You should just have removed the operator and contact details tags.

Hello, The University map team have received a bug report from a Fellow of Trinity College that says "The Sides" is not the correct name of this passage and they say "I think it would be more correct if it read 'SCREENS
PASSAGE'." Do you have any comment about this? If "The Sides" is a colloquial name, we can keep that as alt_name.

Though the formal changeover of the route is at Outpatients, Whippet (whose timetable it is in the first place!) has asked that we start it at Francis Crick Avenue, as that's the best stop to head into town from the west side of the hospital, but it still shows as "towards Addenbrookes". Also, they told me the LMB stop doesn't serve the U service, so dropped from the relation.

Well if it makes you feel better, feel free to change it. Personally I think that's unnecessarily pedantic. In any case, the road will be gone within a year (all being well).

Yes, perhaps not the best idea. I've changed it to "haul road", lower case. It's a temporary road for construction traffic, and this tends to be what it is referred to as, though it isn't an official name. I think it's better to have something here than nothing, which would be the alternative.

Regrettably in making this change, you managed to move a node at the junction of St John's Street, Trinity Street and All Saints Passage to Hobson Street. I've changed it back.

This just seems like a hack, and it'll all end up going round the same circle again in future.

But also, it effectively means University of Cambridge as an entity effectively disappears off the map completely (except for the University's own map) because it would be represented as a non-standard tag, including in the application you seem to be envisaging. Operator simply doesn't encapsulate "university-ness".

It was "trigpoint" who said tagging for the renderer.

Now, that *would be* tagging for the renderer! I really don't think that is the answer. (Nor is "operator" on its own, that's not enough information).

Cambridge University is not a campus University. There are about 10 areas you might call campuses in the conventional sense, maybe, and also 30 or so University colleges, which are independent institutions which are also part of the University but not operated by it. There are numerous smaller "campuses", in the limit being individual properties, or areas within other properties (of which these hospital ones are just a selection). For example, this: osm.org/way/48910156 is different only in scale, not status. If I try to take your logic to the limit, there would be no amenity=university in for University of Cambridge, because there is no single amenity. I chose to take it to the other limitwhen this was all first mapped in 2011.

I don't think by making a change here you're going to get the answer to your "use case" you want anyway, because Cambridge University (and indeed others, it's not unique) just doesn't fit your model of an institution in one or a few large campuses. It is way more complicated than that. It seems like you want to re-engineer the data so it provides the answer you want to get, but I'm not sure the question makes sense in the context of this institution. Are you counting Campuses? If so, by what judgement is Douglas House less a Campus than, say, Homerton College, or the "Silver Street/Mill Lane site"? And what about co-operative ventures? Or if you are counting University's you won't get the answer 1 for us, short of completely re-organising it all as a single relation, because of the multiple campuses and collegiate system. You'd either get zero or a dozen or so, or 50, or 1500 or so depending what you look at. You also can't say that a building isn't a university because some universities (not Cambridge, obviously) *are* single buildings.

A relation which contains all the University premises might be an answer, but it would probably also need some new tags, which identify campuses or property areas, as opposed to buildings, and similar. It's really hard to maintain relations though, and it really doesn't add anything to the information, it's just a different way of grouping. building=university has been suggested, but it doesn't solve the problem of things which aren't buildings like the hospital ones, and it means the building can't be tagged as anything else (e.g. a residential block which is university premises). amenity=university_premises or some such might work, but then requires everything else that understands universities to change.

But theses all require someone to make an investment of significant amounts of staff time and real money to cope with non-upward-compatible change.

I did think about the way the University is tagged long and hard before I did it, and published the method. The interpretation of amenity=university in the way you want to make it largely post-dates that.

(Whatever else, just locally changing some blobs in the hospital area inconsistently with the rest is surely the worst possible answer.)

Not just those. This area is not a university campus, but some parts of it are University. Not is it a hospital completely. And some of it is private commercial. There is already a boundary for the campus, but it's a diverse place.

Operator is completely insufficient to represent the data here or elsewhere in Cambridge. University of Cambridge "operates" many things that aren't "university". building=university might be sufficient, though some of these locations aren't actually separate buildings. I think amenity=university expresses the intent here. There are tens of thousands of such instances across Cambridge that have been used in this way for five years or more. It would be a mammoth job to change these, simply because you disagree with the way the tag has been used.

Not at all: tagging for the renderer is where you tag something to make it appear the way you want on some particular renderer. These really are parts of the University, and the rest (almost all) the hospital isn't; grouping them together is just wrong, it makes them indistinguishable. In any case changing the tagging on a whim under our feet isn't helpful.

These are part of the University of Cambridge, not the hospital. By removing these tags, you wreck the University of Cambridge's online map, so we propose to put them back in. If you don't like this, perhaps we can have a discussion about it, but in the meantime, we need our map back!

Thanks, I think I've got them all now, updated in changeset osm.org/changeset/39732659

(My comment should, of course have said -3 not 3)

Yes, I think it is still true. mapnik doesn't use layer, so for the mapnik based renderers (which are most), it won't make any difference, but for any renderer that uses layers as intended, an area at one layer will by definition obliterate anything at lower layers.

Just to say, I've reinstated the building tags on the individual buildings and changed the multipolygon you groups these into into the hospital. These are separate, though connected, buildings, apart from anything else with a different number of levels. Importantly, this change broke the University of Cambridge online map, one of the main map customers for this area.

No it wasn't, good catch. I've fixed it.