Historically one of the most heated flame-wars in the Scheme
community. We have to decide whether we want to preserve backwards
compatibility with standards up through the R5RS, or reaffirm the
switch to case-sensitivity in the R6RS. The charter gives precedence
to the former, so unless we can reach a rough consensus (> 90%) to the
contrary, the default should remain case-insensitive.
Whatever we choose we should realize that some implementations are likely
to use a different default in their preferred environments.
Note that the default case-sensitivity is orthogonal to the issue of
whether and how behavior can be toggled on a per-file or
per-expression basis. If there is a toggle, it may be specified by
WG1 for both groups, or only by WG2.
Below are some common arguments for each side. The number of arguments
on each side is irrelevant - people should weigh each argument according
to its importance.
- R[0-5]RS and IEEE Scheme compatible
- possibly easier for beginners not expecting case distinctions to be significant
- allows using different cases as stylistic differences in source
- prevents using separate identifiers which differ only in case, which is considered poor style
- less confusion when code is read aloud, e.g. variable names
- R6RS compatible
- the popular default in a number of modern implementations
- easier compatibility with external data (XML, some filesystems, FFIs)
- users from other languages usually expect case sensitivity
- using different cases for the same identifier can be confusing
- transliterations of math formulae may be easier to read with case distinctions
- more general - can write a (begin/ci ...) macro to implement the alternative
- allows using separate identifiers which differ only in case (e.g., capitalizing a class but not an instance)
- simpler to implement in the presence of complex character sets
- most programmers won't be able to recognize case-folded equivalents from languages they don't speak