| Feature | Target |
| Port to Intel C++ | ?? |
| Automated comprehensive performance timing suite | Q2 2000 |
| Interfaces to more LAPACK routines (banded, packed, symmetric) | ?? |
| Reference implementation of the new BLAS on MTL | Q2 2000 |
| More blocked algorithms (Banded/triangle mat-mat mult(), tri_solve()) | Q1 2000 |
| Expression templates ala Blitz++ or PETE | December |
| Sparse direct solvers | ?? |
| Parallel execution (for shared and distributed memory) | Q2 200 |
| Scripting interface to MTL | ?? |