I agree, though the one published in dedicated "software editions" is usually okay. But it's a question on what you're optimizing for. Scientific publications mostly optimize for good (or at least impressive) science and novelty, not software quality. But if you don't want to publish in scientific journals, why publish at all?