So, Programming Language are good suitable for what they do best, even Delphi exist also as a Pascal OOP proposal, so don’t missunderstand what a good engineer has to deal with some “evangelist” statements that fit better with some educational enviroments. In the real world you have to pick the best tool you can have to solve the real problem, so don’t hessitate even on using perl, ruby, python and other programming language available somewhere in the World Wide Web.

Regards, TG

For algorithmic code, I believe that “non-idiomatic” Python (no list comprehensions, or other features that don’t resemble those in C/Java) would be more concise and readable in most situations. It’s much closer to working pseudocode. For code involving the design of object hierarchies, Python might be more confusing to those who don’t know the language because of its syntax for object initialization, its syntax for static methods (or penchant for eschewing this in favor of functions at the module scope), etc.

I am not a Ruby expert, but I expect that language could also be quite readable to those who don’t know it if–again–you avoided features (such as x.each {}, etc.) that have a less obvious correspondence to features of C/Java.

Scheme probably does the most with the least syntax, so it is actually a good choice for expressing ideas in a syntax-neutral way. However, it’s even less likely to be considered “readable” by anyone doesn’t know a dialect of Lisp. And of course you would not want to to touch code that uses macros.

Haskell might be a very “pure” functional language, but its syntax choses purity and simplicity over similarity to more “mainstream” languages. Ideas are not likely to be expressed similarly in both Haskell and another language, and it’s probably easier to translate from another language to Haskell than from Haskell to another language (with the important exception of output–monad-based output tends to be much trickier than the side-effect based output found in most languages).

And the more specialized the domain, the less likely that a general purpose language will beat out a domain-specific language. The above poster’s comment on Prolog for logic programming makes sense. Similarly, if you’re doing parallel programming, something like Cilk or whatever they call the Java version of Cilk would be a better choice.

There is a best language for every situation. Unless you only refer to enterprise and `basic’ applications, than Java is probably it. But if you want to make yourself understood when showing off neural nets, and you spend 50 minutes fighting through the syntax, when you could use something simpler (such as plain pseudo-code, or even a more appropriate language, such as Lisp), Java just doesnt do it.

“When I need to teach cross-platform subjects like DOM or data structures to mixed language audiences, I can easily use Java to show examples and everyone can follow them, even if they’re not Java programmers.”
That is an example where it would be a acceptable language to describe the problem. But if you stray too far from that path, even if they know Java, you’re probably missing the point.

The fact that we even consider a programming lingua franca is repulsive. The fact that java is even considered as one, is even more desgusting.

The language that I found easiest to read (not write) for modern audiences is actually Python. And in terms of documenting complex real-code well, Smalltalk’s syntax is unbeatable.

http://wgrosso.wordpress.com/files/2006/04/2003_Java_Better.pdf

quote: “…That Java itself is inevitable from a social point of view…. A whole new style of programming
Write Half, Steal the Rest”

Java copies better than most other languages.