(Argh, I just lost a 20-paragraph post because I chose to change my post style to plain old text so it would format nicely. Thanks slashdot...)
I'll be brief this time. Disclaimer: I'm the author of morituri, a CD ripper for Linux with support for AccurateRip, modelled after Exact Audio Copy but command line.
I was in the same situation as you a few years ago. I had originally ripped my collection to Ogg/Vorbis, and thought that this time I wanted everything 100% lossless so I would never have to rip again, but just transcode from the rips. The main issue I wanted to solve, besides going lossless, is to make sure I had no bad tracks with skips in my collection. (You detect those skips over the years listening to the songs, not as you do the rip, and you're never sure if there are glitches in the tracks or not, and it drove me crazy).
But when I researched what it meant to get it right this time, my mind got blown at everything that could go wrong. Here's a condensed version of the results of my research.
The biggest eye opener to me was that the fact that each drive model reads samples with a different offset. That offset is always the same for that model, but different across models. I have no idea why it is so (does anyone know), and we're lucky that it's constant for a model, otherwise I wouldn't even be able to solve my main concern - the detection of skips and bad rips. Nowadays people use AccurateRip, a database of checksums for ripped tracks that people upload. If your rip matches several other people's rips, you can be reasonably sure that you have a correct rip.
Since at the time there wasn't a single Linux-based ripper doing this, I created morituri.
There are several other issues that make ripping a fragile activity. I recommend you get a drive that is able to rip Hidden Track One Audio (The audio in Track 01 but between Index 00 and 01). Maybe you don't care, but I have a few gems in my collection with good stuff there (two Soulwax albums and Luke Haines's Das Capital spring to mind). Some drives are simply not able to get at this data. Most software doesn't get it either. EAC can be told to do so, but it's a manual and fragile process. morituri's goal is to create a perfect image so that you can burn a bit-exact copy; so it rips the HTOA tracks always.
I suggest you rethink whether you really want to go quick and dirty. You're going to rip the cd's once and then listen to the result many times. Are you sure you don't want to get it right on the first try this time ? Is your time spent changing the discs not valuable enough to not have to repeat it ?
morituri is probably slower than less accurate rippers, as the focus is accuracy. I would argue that the time spent ripping and encoding really is not the biggest issue. The real trouble is having to change disks, which is going to take time no matter how much time it takes for your computer to do its thing.
I made a quick calculation of how much time I would be spending to put in my 1600 CD's, and decided to spend that time on creating a LEGO CD changer instead (I had checked the price of disc changers, and the cheapest I could find was around $800, with no real guarantee of whether I'd be able to control it from Linux).
Friends visiting shared their scorn and admiration in equal doses, but the robot was able to do around 20 CD's reliably in one go, so I would just load them up in the morning before work. 3 months later my 1600 CD collection was digitized.
morituri interfaces with MusicBrainz to get the metadata, and you can retag albums later on based on a different release id or when the data is updated on MB. There's also options to do an encode of lossless rips; I regularly run a simple shell command to transcode the flacs to mp3s, and it only transcodes what wasn't done before.
Give it a try, let me know what you think.