Open Krakonos opened 9 years ago
Oh, and a match that should be suppressed: don't complete to keys that are already set on the current item (values however can be completed thus).
Status changed from New to In ProgressI'll see what can be done.
OTOH, you realize that you can pretty much customize your tagging experience the way you want with the Tag Templates, do you?
Replying to [comment:2 koying]:
OTOH, you realize that you can pretty much customize your tagging experience the way you want with the Tag Templates, do you?
Yes, but I guess old habits just die hard :)
Assigned to deleted (koying -)Target version set to 0.18-devStatus was set to "In Progress" over 3 years ago. Is this issue still being worked on? Intelligent autocompletion is something I've been looking forward to since I started mapping 2 years ago.
Maybe this issue should be split into a feature request (AI) and a bug report (duplication of keys).
If the AI thing is too complex, I suggest using:for autocompletion of keys: take the most abundant (or: recently entered) fitting key not aleady assigned to the current object.for autocompletion of values: take the most abundant (or: recently entered) fitting value for the respective key.
(Ok, this is about polishing a rather well-working feature, so has low priority. Some of the suggestions would likely need further tweaking based on user response.)
I imagine the tag completion in the tag editor could be improved if a match was selected from several "buckets" of different priority. If a match is found in more than one bucket, the one with the highest priority would win. (A shortcut would cycle through the matches.)
Priorities could be (from highest to lowest): strings from loaded GPX waypoints; tags that I typed in in this session (prefer the match I typed in most recently over other ones in this session); tags that were downloaded from OSM.
For ways, prefer tags that are on other ways; for nodes, prefer tags that are on other nodes. This would be beneficial to my editing habits even if nothing else was implemented.
For values, complete values that have the same key first, but also match values from other keys (with very low priority though).
Another possibility would be a spatial awareness, i.e. prefer tags that are used close by.
Or... Save all tags that I ever type in myself (cumulated over all editing sessions), and select from this pool. This alone could also be a big improvement (especially if we filter out the name key, as that tends to be unique).
What I'm missing most right now is that Merkaartor seems to only do completion on downloaded tags, but not on my own edits in this session. As a first step, could these be dynamically added to the list of possible matches as I edit along?