Linear optics quantum computing (LOQC) has become one of the leading candidates for the implementation of large-scale quantum computation. However, it is extremely technologically challenging - an optical quantum computer with post-classical capabilities would require technology well beyond what is presently technologically possible. Recently, Aaronson & Arkhipov presented a much simpler model for LOQC, known as 'Boson-sampling'. It does away with some of the more challenging requirements, such as quantum memory and fast-feedforward, while still implementing an algorithm that cannot be efficiently classically simulated. In this talk I will present a review of Boson-sampling. I will then present some recent research results, which indicate that Boson-sampling is highly robust against photon loss, source and detector inefficiency, photon distinguishability, mode-mismatch and spectral impurity. I will also present an interesting encryption algorithm allowing for secure computation via Boson-sampling. These results suggest that a Boson-sampling computer may well be within reach of present-day technology, and an optical quantum computer with post-classical capabilities may occur much sooner than previously expected.