I am principally interested in development languages for scientific computing and file-system management (FORTRAN, Python, Octave, R, C/C++, bash, ...). In my way of using an IDE, compiling and building is purely instrumental in running calculations after editing a source code, and not an area of investigation in and of itself. I might just play around with a couple of optimization flag.
I have eyed up two related IDEs Eclipse and LiClipse because of their GUI support. Either website naturally praises its own strong points. In another postpost EMACS has been recommended.
I would rather have a feel of where IDEs are expected to touch their limits --- which I find more useful to anticipate the long-term pros and cons. Such 'limits' could result from (arbitrary order)
- performance (sluggish management of views and windows)
- scarcity/superabundance of options and information (the tree-and-wood effect for the novice)
- steep learning curves to navigate content and select actions
- lack of support for features that are not needed for 'Hallo World' programs but normally turn out to be pretty essential when the task size increases.
Any recommendations/words of caution in that respect?