Since there seems to be some interest in hearing how folks organize their local media, here's my $0.02  

For my part, I store movies as 1234567.iso (using the imdb number as the filename) because:
(a) it makes the imdb lookup from a script later unambiguous and
(b) it is easy to have a parallel directory of hard or soft links that are The Movie Title (year) 

Movie posters are great, but the ones at the imdb are pretty low quality. At the same time that I put the movie in the server, I find a good quality (at least 750x500 and mostly 1500x1000) poster or artwork and save it as 1234567.jpg. Again, using the imdb number as a filename / database key makes it unambiguous to match the right poster and movie title. Because some titles get reused, this can otherwise be a problem. 

  Google Images makes the search and the specification of a "large" image easy. www.impawards.com is another good source. 



At one point, I hit the imdb.com site quickly with a script and my IP address got banned for a while. 

So, now, I keep a local lookup table of imdb number:::The Movie Title (year) 
and only the site for an unknown movie. Checking this file first is also significantly faster than a remote lookup. 

Having some local resources makes it straightforward to use MP4Box to add metadata like the cast info, poster, et al into a transcoded version of the file that can be thrown on an iPad, played by a Roku box, put on an Android phone, etc. HandBrake is, I think, the most straightforward solution to this. 

Keeping the ISO online is best for quality, special features, etc. but few devices will play it directly. 

If I were starting over, I'd use tmdb instead of imdb. 

Thanks for the pointer to the API site. 

Thomas 


On May 21, 2012, at 3:46 AM, Mike Miller <mbmiller+l at gmail.com> wrote:

> I wrote a little script that reads in a DVD ISO filename, which for me is a movie title with a year, and it creates an IMDb query string, launches firefox and shows the result.  It does pretty well at that, but I have to make sure it grabs the right page.
> 
> To do what I think you are doing, I guess I would just need to supply a list of IMDb accession codes in a hash table with the filenames.  But then there is the tricky part of parsing the IMDb pages.  I did a lot of that with allmusic.com data using bash scripts (that had a lot of perl code inside).
>