This paper discusses code-alternations in the English language-learning blogging community “Lang-8”, in which users write blog posts in the target language and get feedback and corrections from native speakers. About 2/3 of the examined blog posts evaded all code-alternations. Among the additional third, the most often observed type of code-alternation included partial or complete blog post translation (interlinear translation and en bloc translation), or switches prompted by lexical need (intricate lexical gaps). Quoting, other sentence-level switches and word-level switches for meta-linguistic discussion were scarce. It is argued that technological affordances, varying comprehsion of one’s audience and of the language learning process, and, to a restricted level, community norms, mold code-alternations and their avoidance.