Discussion: Sage, Macaulay 2, and other Mathematical Software for Algebraic Geometry

What are the absolutely critical features that you '''must''' have in the mathematical software you use for '''your''' research? (E.g., fast linear algebra, Groebner basis, sheaves?)

What are the '''killer features''' that your dream mathematical software would have? (e.g., good mailing list, free, super fast algorithm for XXX, latex output?)

Polytopes

Linear algebra

What are some things that disturb you about the direction in which Sage is going? (E.g., too big/ambitious? not open enough or too open? too many bugs? changing too quickly? referee process for code inclusion too onerous?)

What are some things that disturb you about the direction in which Macaulay2 is going? (similar e.g. as above)