As more and more quirks are found, all is not lost, but there's now a greater chance of a language fork to occur (just like the forke between C and C++ that are no longer compatible. or the second fork with C# (C++ is in my opinion a slowly dying language which has failed all its objectives, while _javascript_ has succeeded a lot and solved all its practical problems by retaining an excellent upward compatibility for its evolutions and still resisting to changes needed for security issues)
Lua is still young, and may leed to Lub for a second corrected major version fixing its major inconsistencies and removing some spurious legacy that were not really useful:
- e.g. long-bracket string literals,
- inconsistancies created by curification, to be solved by either making the semicolon required to separate _expression_ statements or by adding syntaxic distinction between newlines and other whitespaces
and adding a few other things like
- the integration of metatables and function environment in the syntax,
- a specification for other types of literal constructors than just basic strings,
- integration of annotations in the syntax,
- integration of cyclic table constructors (using either tags, or local variables which could also be added inside expressions),
- integration of ternary conditional expressions or n-ary switch expressions without having to use complex anonymous functions syntax,
- suppression of the [] notation of keys in table constructors,
- better semantics for variable scopes and references to values
- better specification for the lifecycle of values themselves
and in the proposed default implementation, fixing:
- the broken table indexing,
- the costly representation of tables using too much memor
- the weak string hashes .
