It will be interesting to see the next version of robots.txt that begins to appear in public repos to prevent this type of data scraping.
If somebody left a code with some license 10 years ago, then that's how he published it and it should be respected. If there are some additional files, changes or checkboxes required on the author side (who may be dead btw), then it is a theater not a solution.
I could totally go with let's abandon licenses (but would prefer starting with patents), what's online is ours. But if we want to keep them then the license itself is enough of a statement. It's not like we have a hard time writing software which interprets them... it's more like "but we would have to do a big refactoring".