Fortress versus making Fotrran better?
807575Jan 24 2008 — edited Jan 29 2008Sun is working on Fortress, which is sort of a modern language aimed at projects that have traditionally used Fortran. I would rather just see Fortran be improved, if at all possible. The key to success is better management of the language standards. Maybe something like Fortress is necessary for this to happen.
The Fortress mention new successful languages that appear to be successful because they are open source projects. I think that the reason they are successful is not so much because open source is desirable, but because committee driven language development, like Fortran, is so unproductive.
The advantage of successful open source languages like Python comes from the language syntax being developed at the same time it is being implemented. Committees end up defining language features that are often difficult to implement, inefficient, and can have basic design flaws. I think that dropping the BITS feature from F2008 late in the game may be a good example of this type of problem.
However, open source projects can make some poor decisions that tailor to the personal preferences of a small group, such as the horrendous indentation syntax used in Python.
What we really need is for the Standards groups to actively develop some sort of experimental compiler that can actually implement language features before they are permanently set in to an official standards document.