How to best handle values longer than 255 characters?

Recording the inscription of a memorial (Node: 8999569936 | OpenStreetMap) has led to me hitting the 255 character limit…

A similar issue was discussed in, where splitting the multiple speciality-values across several more specialised keys appears to be the suggested solution. Since inscription inherently is a single value, I don’t think that solution applies to this case…

Key:inscription - OpenStreetMap Wiki suggests putting the data in another database, such as Wikidata.

For now, to not throw away data, I’ve put the excess part of the inscription in note=. That is not really a good solution in my opinion, though.

Does anyone have suggestions for a better way to record the inscription (prefereably within OSM)?


Well, what I think some people have done is split it into multiple keys, like inscription:1, inscription:2, etc., though I don’t know as there’s a consistent standard as there exist some inscription_2 and inscription2 and such as well.

It’s not clear to me what data consumers exist that make good use of the inscription information, or if any of them try to work out ways that it might be split across tags.


The common way across many different keys is the version with colons: inscription:1. Underscores are used more rarely, and mostly in old, imported tags like ‘name_1’.

I agree that we shouldn’t use OSM to store huge amounts of text. Wikipedia, wikidata and alike are much more suited to store context about objects in OSM. Nevertheless there are quite many cases where a text is just a bit too long, or e.g. destination signs with many entries. In these cases I think it’s a good solution that can be supported rather easily in any tools using the data.


Thank you for your input @PeterCooperJr and @mueschel!

It seems like “inscription:2” is paired with “inscription:1” in only 56% of cases, reflecting that there are multiple variants of this method:

I guess the variants starting with inscription are somewhat compatible with consumers unaware multiple parts of the value…

On the other hand, the version starting with inscription:1 prevents unaware consumers from processing incomplete data by accident.


I’d really love to know if there’s anything out there really using inscription (particularly for memorials), other than systems that just allow one to query a particular node to show all tags. I’ve mapped it on several items, but I sometimes feel like I’m just doing it to feel like I did something and that nobody else will ever see it.

OsmAnd displays inscription for memorials under the Details menu (but does not understand the | used for line breaks). Searching for bits of an inscription does not seem to work.

1 Like

I had a look on the popularity of different solutions:

Specifically for historic=memorial, wikidata is tagged 19.600 times, no use of inscription:1 nor inscription:2

We also have a large amount of ‘description1’, mainly from an import in 2010.

Across all tags I think it’s safe to say that the version with a colon (:1) is the most common - if we ignore imports like ‘name1’ from Tiger data.

Not quite:
Almost 50% of the inscription:1 tags are used on historic=memorial.
The table with combinations in Taginfo cuts of at a count of 1000. Here’s the overpass query for these: