Lists: Sorting list by title but ignoring articles

I agree and have just added my agreement to your comment/question.

After all the years of database sorting your suggestion should be the default.

It is a good idea. I'd say post it as a suggestion (if it is needed - this might be good enough).

Just to provide some historical perspective/context: until a few years ago, titles used to be stored internally with the article appended at the end, e.g.

Blues Brothers, The (1980)
Dolce vita, La (1960)
Boot, Das (1981)
Mariachi, El (1992)
Clockwork Orange, A (1971)
etc.

This was better when generating/displaying alphabetized lists (as it eliminated the inconvenience you are describing) but it was not devoid of complications.

For display purposes. the title needed to be converted to the more readable format (e.g. The Blues Brothers, Das Boot, La dolce vita). This made sorting better in many cases, but presented its own set of challenges -- the article needed to be put back at the beginning when displaying the title, but different languages have different rules for that (e.g. French titles beginning with indefinite article needed to alphabetized under the letter U, but those beginning with the definite article needed to be alphabetized by the letter of the first following noun); capitalization rules also vary, creating all sorts of complications (when converting an Italian or Spanish title you'd have to change the first letter of the title to lowercase when prefixing an article, but you need to keep it uppercase for English and German).

Also, detecting which titles had an article in need to be moved to the front versus those which didn't was a challenge in itself, requiring the need for built-in exceptions (otherwise a title like "Die, Mommie, Die" would be rendered as "Die Die, Mommie" unless the system was made aware that 'Die' here was not a German definite article)

Last but not least, a non-trivial amount of users objected to this format and complained about the sort order being inaccurate because of titles 'La dolce vita' appearing under 'D' and not 'L' where they wanted/expected it to be.

To be honest, I don't remember exactly when we switched to the current format (possibly around 2005). But I do know that apart from the 'alphabetical list' inconvenience of having many titles grouped under the letter T for 'The' etc. (which is not necessarily a problem for some users, as explained above) it did not generate significant complaints or feedback.

I'm not saying that the current system is necessarily better, although it did simplify the way we approached certain technical choices; just pointing out there is no clear-cut solution and that each approach has its shortcomings. Ideally of course we would be able to store both versions of a title (one for sorting, one for display) but adding/backfilling all that information that would be a non-trivial task.

I must admit, I assumed you'd kept the original sorting, so you could sort against that and display the other. You could have even offered people the option of sorting with or without the article.

The backfilling shouldn't be that difficult - you could hand-fill out some of the weird exceptions (as you mention, those that start with "Die") and then run some bespoke SQL to process the various titles that start with the various articles.

The longer this isn't corrected the more of a 'non-trivial' task it becomes.

Possible to enlist the help of users?

Articles being included in sorting is absolutely dreadful. 

I suggest that IMBd contact Collectorz.com, specifically Alwin Hoogerdijk and see if they'll sell you the solution. They have solved all the IMBd issues that I've ever read about. Of course, they are a paid service specializing in books, movies, music, comics and games with a great feature set. Problems are usually corrected within a couple of days. I switched to their products almost a year ago after not getting any response to the article issue here.

Either IMDB backfilled to combine the article with the title in the db's title field or it is still separate in the db and they are combining it before it gets sent to the client. If the latter is the case then simply separate it in the presentation bean and combine it in the client, at which point offering the option of prepending or appending would be easy enough. If they actually modified the title field in the db then they would be left with the decision of which articles to move and which to leave alone. Giving an option to leave it as is (default) or move the article to the end would put the choice in the user's hands and we'd get what we asked for.

I don't sort A-Z because of the clump at "The ...", which I think is ridiculous.

I am a very amateur programmer, but here is the regex I use in my perl sort subroutine.

s/\s*\b(A|a|An|an|The|the)(_|\s)//xi; # Strip off leading articles (in English).

I'm sure your programmers could take into account the title language and sort accordingly.