Yes, the Archos still suffers from many, many issues (many of them can not be fixed in software) but, through 2 years of software development it has become a pretty polished piece of firmware

..where as Karma has been out for all of 1 day. I would have liked to have got v1.11 into production but we couldn't get it qualified in time. It was always intended to be online by the time any customer received a player, and that is a screw up, but it should be there very soon.

Beyond that, it is a wide audience and customer feedback that creates a great product. We've had several months of beta testing with a quite diverse team but day one of retail shipping will always generate a slew of reports and feedback. In your case the feedback was "this sucks, I'm returning it" so clearly there's nothing I can do to help with that.

But, it has nowhere the polish of the empeg

It doesn't have three years of incremental development behind it. Even so I think it outshines the car player (with which it shares a codebase) in most areas.

Many people are mistaken when they think that a UI can just be thrown on top of some code

The UI was designed before anything else, although our codebase *is* exceptionally modular as it happens. You haven't really made it clear what you hate about the UI so much, except some trivial stuff like a misaligned control. It makes me think you're totally missing the point (which by implication is still our problem) as the player is so patently simple to use. The only bit I really dislike is the playlist editor (which most other players have chosen not to implement) but since the Now Playing view provides a really nice method for building playlists out of the running order I'm less concerned about that.

It may be true, though, that in terms of final polishing more work has gone into the filter graph than anything else. The core audio functionality is #1 priority and already has many additional capabilities which are just waiting to be exposed.

The UI developer needs to work with the “engine” developers throughout the entire process

In fact we use a short iterative development discipline (eXteme Programming) to ensure that all members of the team work closely.

Just look at the synching animation. It’s beautiful. Why isn’t the screen used to its potential. I am thinking about embedded album covers that are viewable on the device, full screen animations, etc.

Toby will be pleased to hear that he is responsible for the only part of the product you actually like. It took him all of half an hour to write. Album art, visualisations etc are gloss features that get held back or added in to iterative releases. For example, right now the only visual is the VU meter, but all the car player visualisation core is there ready for them to get plugged in. I think your comments have proved that there are *already* too many 1.0 features and launching with even more of them would have pushed the project out indefinitely. The QA regression plan already takes several days to run, and must be run for each and every RC (1.11 has taken two RC's which is why it's not online today).

I truly believe that with all the bugs present that you should have held off on the release until closer to the holiday season

I think you misunderstand retail purchasing. Christmas is the start of October - there is no opportunity to get into a store after that (well, not this year anyway).

Rob