Sometimes I'll instant the components in order to test them out, then delete them and research them the proper way. I especially do this for missile engines, because I want to get them just right. Otherwise I research my components normally. It adds another level of challenge: at any given moment you have to decide whether to "commit" to a technology or to wait for "one more level...". If you commit too early, your designs will be inferior to your enemies', but if you keep researching for too long, you won't get anything into production in time.