Reply to thread

Message: <blockquote data-quote="fforde" data-source="post: 543434" data-attributes="member: 52082">The biggest problems with what you are suggesting RoChess involve usability. To begin with, your plan would require two rounds of importation for one movie, meaning the user would in some cases have to approve a match twice, once for each language. This I think would be counter intuitive to most people and the UI we have for the importer does not really support this kind of work flow very well. Then you get into error conditions that have to be dealt with and exposed to the user so that they can correct things. Instead of a file getting auto approved as the wrong movie, now you might get an English language film imported in German because of a bad match in phase one. Or a German language film that matched correctly in round one but incorrectly in round two, meaning you'd have the right language but the wrong movie details. How do we expose these situations to the user? Currently the user can just click the drop down in the importer to select the correct movie when something goes wrong. But how should this be handled with a two phased import process? What possible matches should be listed in the drop down, search results from phase one or phase two? And when the user selects a movie from the drop down will this restart the whole process? Should the importer attempt to switch to a new scraper script if necessary again? So what you are suggesting is not very easy. It would be far simpler to select a language based on .NFO details, filename or mediainfo. Even a setup like this though would require a sophisticated rules system to allow the user to define what scripts should be used when. This could easily be just as complicated as the filtering system (although perhaps a few components could be reused). Ultimately though I am a bit skeptical that the payoff would be worth the effort. I suspect most users will want all their movie details in the same language. Meaning a lot of development time would be spent on a feature that would not be used by a large number of users.</blockquote>

Verification