Encoders are required by the spec to emit uppercase only for this reason, since L is plenty distinct from 1 (whereas I and 1 can still get mixed up, and therefo

Encoders are required by the spec to emit uppercase only for this reason, since L is plenty distinct from 1 (whereas I and 1 can still get mixed up, and therefore the former is an alias for the latter). A bit unfair to uppercase L that it has to be so neglected just because of its lowercase counterpart :) The original Tcl implementation treated uppercase L and lowercase l separately (i.e. uppercase was standalone and lowercase aliased to 1 like I/i do), but in the process of writing up that documentation site I quickly learned that it ended up just confusing me more than it helped, so I just kept lowercase and uppercase L together and distinct and called it a day, figuring that discouraging lowercase output (while still allowing lowercase input) was a good enough workaround for the lI1 issue.