Since this does not appear to have been asked here before, I would like to solicit suggestions and recommendations, ideally with rationales, for a textbook in a first course in linear algebra. In my university, students will not necessarily have taken any proof-based course (we do not have an introduction to proofs course). Our course is expected to cover basic computational issues including Gaussian elimination and eigenvectors/eigenvalues. We also cover theoretical material including an introduction to vector spaces over $\mathbb{R}$ (as well as over $\mathbb{C}$ to deal with diagonalization) culminating, ideally, with a statement and proof of the Spectral Theorem. We expect students to learn how to carry out fairly straightforward proofs that primarily follow from the definitions or directly from one of the theorems covered in class. Because we have a large number of engineers and other science majors in the course, I do plan to spend some time on practical issues as well. I like to discuss rotations and reflections in $\mathbb{R}^3$ using an eigenvector/eigenvalue analysis. I will definitely cover the singular value decomposition and would like to spend some time discussing modified Gaussian elimination to handle numerical problems due to rounding error, though the latter is likely to be omitted because of time constraints.
I've tried several different books before (Strang, Anton) and yet to have found one that had a presentation close to what I have in mind for the course. Any suggestions for books whose aims are aligned with the goals above would be appreciated.