In this project, I explain a framework for the evaluation of testing procedures in web-based language learning programs containing four dimensions: construct validity, reliability, authenticity and impact. Qualitative, open-ended criteria are proposed and applied in an n=1 case study format. The evaluated program shows little validity with its stated objectives, tending towards informal, pre-communicative assessment that reduces the status of teachers and frames student performance in terms of discrete language functions. The framework itself, while informing the analysis, is still awkward to use due to the complexity of the subject. Appended are: (1) Brief annotated international bibliography of IT integration into education; (2) Research venture proposal; and (3) Interview questions. (Contains 1 table and 7 figures.)