[quote user="caddcop"]
I've even tried adding breaklines to force the triangles to form and retain edits, but found that other triangles can be effected by the added breaklines and will often change in a manner that requires even more breaklines to be needed.
[/quote]
Caddcop, just out of curiosity...have you experienced the same issues using the soft breaklines? Honestly, I haven't seen the behavior your speaking of myself...not that it's not happening, I don't doubt it, but just wondering if the newer feature type may alleviate the problem. The only issue I've seen with the soft breakline feature type is that it's not a supported feature in the older DTM format, so an export will keep the triangulation as shown in the terrain model until you re-triangulate the DTM, then they "disappear".
I too like the idea of an audit trail of modifications to the terrain model to prevent the issues Roadrunner is experiencing. It's one thing to disjoin the surface from the survey data, make changes to it, then re-join it because you're trying to add survey data to it...I could sort of expect the behavior that's being experienced (that's a whole different discussion), but another workflow that shouldn't have any issues (at least in my mind):
- Turn off (from Survey Settings) Create Terrain Model for All Field Books
- After importing survey data, you need to manually right-click field book to Create Terrain Model...
- Make edits as necessary
- Bring in add-on survey into new field book and create its own terrain model.
- Use the Create Complex Terrain Model command to join the surface data as required.
The problem that I'm noticing (which I'm sure you have experienced as well) is that once you try to do an Append, the flipped triangles go back to the way they were. This just shouldn't happen. Like you mentioned...some sort of audit trail would be able to prevent this from happening.
If you do a Merge instead, this isn't the case...the triangles stay flipped, but there are other unwanted results I'm sure depending on the dataset. Without the audit trail however, I'm sure some other command would flip them back at some point even if the merge was an acceptable method for the dataset.