The custom of “voting” for OSM tags on the wiki really is the tail wagging the dog here.
+1, no issue with 10 people introducing a tag catering for a niche interest, but a bunch of wiki voters deprecating a tag with 100.000 global uses is not a desirable IMHO, especially when there aren’t any issues to solve, it’s a mere renaming and the reasons are only cosmetic.
I’m happy that people like @ChillyDL are creating new tagging standards and deprecating unclear tags like site_type in the process. We can never expect the ENTIRE OSM community to join in the decision-making process, but if a few mappers with a practical mindset can get a proposal approved and inform the broader community and prominent users of big changes, then that’s a good way to evolve tagging schemes.
No. I could use stronger language here, but I suspect I’d get (rightly) censured by the moderators.
No-one doubts the good-faith efforts been put in by all sides here, and no-one doubts that there is a problem, but that net result of the “proposals” and the tag changes that followed that** have been entirely negative.
“Chesterton’s fence” should be the guide here. Changes to tagging recommendations need to take on board the different parallel ways that objects get classified***, how existing objects are classified****, and how accurate those classifications here*****. None of this happened in this process.
** The “rejection” of the original “crannog” proposal because a few people didn’t like “site_type”, the “acceptance” of archaeological_site as a key, the initial undiscussed tag changes that followed that and got reverted in the German community, and the subsequent tag changes that picked up non-archaeological items etc.
*** As I write this, those archaeological sites local to me in OSM are mostly categorised by “archaeological_site” OR “fortification_type”. The crannog proposal was based on a “settlement type” structure - plausible, but didn’t fit well with existing OSM data.
**** See here for comments about processing issues. These are of course also settlements.
***** This really needs to be done object by object by people familiar with the objects and the classification.
I don’t see why. The point of archaeological_site=* is not to change any tag values, but only the key. The meanings of the tags would not be changed, and archaeological_site is well-documented on the Wiki by now.
Your idea would divide the data between the old key and the new one, forcing people to support both without any idea about what the prominent tagging will look like in the future. I consider this undesireable. Documentation and knowledge about archaeological_site is already adequate to a level at which mappers and users can easily adapt to it once it is implemented in OSM.
You’re free to re-read the proposal if you don’t understand it.
(commenting on just this for now, since there’s clearly a more detailed explanation needed for the rest of it)
What exactly are you trying to say (and which of the many proposals are you actually referring to here)?
Proposal:Key:archaeological site - OpenStreetMap Wiki has been approved, and now you’re invalidating it because some criteria you mention are apparently not met. I say the proposal is good enough and we have a misunderstanding over the benefits of the tag changes.
I suggest that you draft new proposal guidelines and get those approved so we don’t have to debate the validity of approved proposals anymore.
What exactly do you think has been approved?
In addition, what about these questions about any objects encountered during any change:
How does this impact the other ways used in OSM to classify archaeological sites, which are broadly:
(in addition historic:civilization is used, but that’s orthogonal to any other classification)
Where OSM is unclear, how far are you prepared to go to categorize something properly? For example, it looks like most alleged crannogs in OSM now actually are tagged as something other than “name=Crannog” (which is a definite improvement over the last couple of months) we still have over a dozen “note=might be a crannog”.
FWIW, my understanding:
Yes.
That appears to be one of the intents of the proposal however as you point out, strictly speaking, no.
No. Presumably they wouldn’t be tagged with historic=archaeological_site so don’t fall under the proposal.
No. Presumably they wouldn’t have been tagged with historic=archaeological_site so don’t fall under the proposal.
Sure - if the editor is confident that this change would be correct and improve the current tagging. But neither of these were part of the proposal and nor are they strictly related to it (well, OK, they’re related but not in scope).
“It is proposed to change the main key for archaeological sites from site_type=* to archaeological_site=*.”
This means historic=archaeological_site + archaeological_site=* is now the approved way to tag archaeological sites, and it means that site_type=* on archaeological sites is now deprecated in favour of archaeological_site=*.
That seems to be out of scope for the approved proposal.
However, there are a few fortification_type tags which can be already re-tagged with the approved defensive_works=* tagging scheme.
The relevant guideline for this is Any tags you like.
In a successful voting on the OSM wiki, it had been decided on 2022-12-03T23:00:00Z to replace site_type= with archaeological_site=. 100,000+ POIs were affected.
Unfortunately, this decision was followed by the mechanical re-tagging of around 75,000 of these POIs without prior discussion, and which was then aborted.
To mitigate the effects of this re-tagging, after the discussion above, at the beginning of the year I tagged all archaeological POIs tagged site_type= simultaneously with archaeological_site= and vice versa, so both tags would have the same value. This was meant for a transition period to end in July 2023 for data users to adapt. Here is the documentation.
The next step would now be to mechanically remove the site_type= in all archaeological POIs simultaneously tagged archaeological_site= with the same value, as discussed above.
Are there any known instances where this will lead to trouble?
Step 2 is now completed, the removal of site_type= from all archaeological POIs that were simultaneously tagged archaeological_site=.
Not quite - have a look here and query by “site_type”. Taking tumulus as an example, both of the remaining ones are still dual-tagged.
Thanks for the links. Those are special cases where POIs using “archaeological tags” are not tagged with *historic=archaeological_site but e.g. as historic=yes or amenity=place_of_worship. For the sake of data quality, I did not include those in the mechanical edit yesterday but manually adapted the tagging on a one-by-one basis today.
Has there been any development on the tagging of Māori pā?
It does seem as though they should be better aligned with the historic tagging scheme (as historic fortified settlements) but obviously we need input from NZ mappers.
The NZ community were in favour of aligning the tagging scheme of pās, so I did today.
For info, archaeological site tagging was one of the things I mentioned in this diary entry (scroll down to “Lots more historical icons”). I’d like to do a flowchart that explains just how complex it actually is , but time hasn’t permitted yet…
New values in e.g. archaeological_site crop up all the time (see e.g. here). There’s a big crossover between “settlement” and “fortification”, which adds to the confusion.
At first glance it looks like one of those keys that can mean basically anything, so not especially useful. In particular, some usage seems to be small components of sites, and some seem to be the whole feature
At some point I’ll go through the combinations and see if I can uncover anything that I’d want to render on raster maps. There’s more of a case for adding to Garmin maps since much of what is there is “searchable” rather than “cluttering up the diaplay”.
In Britain many of them are (the illustration of a motte on the wiki is rather atypical) especially if built before the English Civil War. This is one quite close to me which is a scheduled monument (usually implies archaeological). It’s not been mapped as it is only really visible on Lidar.
I think most places mapped as motte & bailey will be like this. I don’t think it’s a useful tag on, say Clifford’s Tower (although this is the motte) or the Round Tower at Windsor (again the motte is under the current building). fortification_type=motte is not used much in comparison with the likely number of actual sites (a little under a thousand), but both in the UK, and globally seems to be mainly used with historic=archaeological_site.
Nottingham Castle was a medieval defensive castle, most of which was destroyed at the end of the Civil War (which Charles I started a few metres away) and what remains are the lower parts of outer curtain walls and a gatehouse restored in Victorian times… Later a stately home or small palace was built on the site of the inner bailey (incorrectly tagged as castle_type=defensive); this in turn was largely destroyed by fire during a riot in 1831 just leaving the external walls. The current building was largely designed as a museum by TC Hine. In fact most large sites of this type are palimpsests, with multiple layers which could be tagged in different ways. Britain is a little unusual in that changes tended to be abrupt (Dissolution of the Monasteries, Spoiling of castles after the Civil War).
