People keep forgetting there's a Montreal studio. You know, one that was advertised awhile back looking for a multiplayer programmer.
And 6 months is entirely possible. Say the game is scheduled for a Christmas release, meaning out in November. You want the game to go into content complete lockdown and spend say, 4 months to do bug fixing, polishing, and any porting to multiple platforms. Then add about 2 months for it to go through certification to the consoles (does it load levels fast enough off the disc? Fits within memory? Does it crash if left on over night? If someone sends you a message off XBL, will it kill your game?). I suppose you can call the lockdown phase alpha - completely playable but buggy, and the cert phase beta, complete playable and mostly bug free.
Why is certification phase so long? Because you have to submit your release candidate to Sony, Nintendo or Microsoft to validate it works on their consoles (granted, could be skipped if only on the PC and released faster). And if you're under an existing publisher, you submit to them first for their testing, then to the consoles. You get into a cycle of waiting for their testers to go through their testing, get a pass or fail (and very rarely the first one ever passes), then another couple days or so to fix existing bugs, verify they've been fixed before another submission, and another submission. So you're looking at a turn around time of 1-2 weeks and a lot of release candidates.
So you're looking at a lead time of 4-6 months. Manuals and inserts should have been printed already, so you're only just looking at 2 weeks for manufacturing and pressing of the gold master discs, and the final delivery of the game to the shops. Meaning the game should be going into lockdown in the March-April time frame, entering certification from August to September, manufacturing in October, to make your November release date.
You only need a skeleton crew for those stages, obviously most of the programmers, tech designers for quest, gameplay and balancing bugs, a couple (tech) artists to handle art bugs, and of course the QA department to test the game.
So... what do you do with the other content creators for those 6 months? Fire them (bad), roll off to another project (if you're large enough a studio to have several going in an overlapping schedule), get started on the sequel (not always good, don't need that many during prototyping), or: work on DLC.
Often a studio can create several DLC modules from scratch in a couple months, allocating several 5-10 person teams who should be familiar with the game's setting and story, the toolset, art style and such, and they should be able to crank out stuff very quickly. Sometimes it's cut content from the lockdown that has been deemed salvageable, other times it's completely new ("Well now that we know what the engine can do, what the game is about, what should or shouldn't be done, we can better design our levels and quests!").
Since DLC modules are much smaller in scope compared to the main game, they can be tested in a much faster iteration. During the lockdown phase, any changes to the main game might require a complete playthrough from start to finish to make sure a fixed quest bug doesn't affect anything else. So you're looking at a turn-around time anywhere from 5 to 20hrs depending on the game. For a DLC module, you're looking at a smaller turn around duration of a couple hours.
And that's why they can't simply insert it back into the main game and be on the disc if it's done at the same time, it essentially resets the certification process phase since it's new content, as opposed to a fixed bug. Thus, scheduled to be released at the same time as when the game is released as Day 1 DLC.
. I shouldn't really post when I'm dead tired.. I misunderstood your text. And well yes 10 hours is a bit short now that I think about it. Lets just hope that future Bioware games will not be five hours long 
