This is a summary of this page in Finnish.
The National Land Survey of Finland (NLS) will open their topographic database for everybody on May 1 2012. At this time it seems certain, that the license will allow importing any parts of that date to OSM. ("Attribution only" type of license). Other databases will also be opened, including aerial imagery and a laser scanning elevation model. (Full list)
What to do
The consensus is, that any imports should be agreed on beforehand, and that nobody should do imports on their own. Please ask for other users' opinions before you even consider uploading data from the NLS database! Discussions can take place on the OSM Forum Users: Finland section, in #osm-fi IRC channel. If you have ideas of what to import, please read first the Import/Guidelines and mention your plan on the firstname.lastname@example.org mailing list.
As a start, people present in the Fi:OSM Finland Meeting 2012 on 2012-02-06 agreed that some data is good to import, at least in areas without any previous mapping. In rural areas this will bring some evenness to OSM data density; in major cities, there probably isn't much that could be imported.
The NLS Topographic database has information on roughly 500 different classes (listed in Finnish this file). Some feature classes are inappropriate for OSM, such as some symbol points, but many classes can be mapped to distinct and already used tags. Splurting data into OSM without any consideration of the existing data is bound to annoy mappers. If anything is imported, it ought to be sensibly merged with the existing data.
Where OSM data already exists, it might have more attributes than what's available in the Topographic database. Their database does not include cadastral boundaries. All "empty" areas are consider wood, i.e. where no other area element exists (farm, swamp, residential, lake etc.).
- Consider what is usable
- Consider how it can be integrated with what is already in OSM
- Consider how their updates are processed
- (Consider how they can spot errors in their data based on OSM edits)
One (but not the only) way to divide the features is into three groups:
- Can be imported easily
- For example municipality borders: known to be exact, better than what's in osm now, and definitive
- Protected areas, (military) danger zones
- Can be imported manually/semiautomatically
- Some objects could be imported automatically as-is, others need to be matched to existing features
- When the features already exist in OSM, they (often) contain more attributes than the NLS data but the geometry is possibly less accurate. Matching can be automated, but needs to be verified one by one before upload.
- Includes: many waterways, lakes, and several point features: place names, addressing, and for example masts
- Can't be imported without merging everything manually.
- Lot's of manual editing is needed
- Some features might be imported, but probably only in areas where they have not yet been mapped at all
- Imported objects needs to be selected/redrawn one by one, manually
- For example: missing roads, power lines, landuse
The second way to divide the features is by their maintainability:
- Features that the average OSM user can not update/verify (for some definition of "can")
- For example administrative borders
- Features that the average OSM user could maintain, but where we lack the user base to care for the whole country's worth of data
- On the other hand, what if the updates are available, and the tools to feed them into OSM improve?
- Features that the Finnish OSM users can reasonably maintain.
- For example roads, new buildings (at least with the up-to-date aerial imagery)
Here, importing features from at least the first group is discouraged by the general import guidelines, unless the features serve a significant public interest (like the municipality borders).