Some formats are just bad and don't have enough tools, despite their historic popularity. Excel export is one of them- there are not a lot of libraries that re

Some formats are just bad and don't have enough tools, despite their historic popularity. Excel export is one of them- there are not a lot of libraries that reliably read it in all cases, and it isn't even exported properly in many cases (so good luck fixing that). Also changing formats like Apache Hadoop did allows things like sub-linear time (on the master node) file splitting- which is not possible if you are hunting arbitrarily far for matching quotes. So superficially it may look like the opposite of opposing broken formats is "recentism", but there are often good reasons to start over with better trade-offs and better implementations.