HTC is a much bigger company than Archos with a lot more resources to throw at problems to make them go away sooner.
Also, Archos attempted to do something with Android that was more ambitious than what HTC does. Android is designed to be a phone OS and that's what HTC uses it for. Pretty straightforward. Archos, on the other hand decided to use it to power a media tablet and had to write it's own graphics subsystem running within Android to do so. The most HTC does is to write a customized Home app.
I don't mean to take anything away from HTC. They do some fantastic HW design (the Nexus One is a thing of beauty). But their FW is a lot less of a deviation from the basic Android OS and is, therefore, much less of a coding challenge than something like the A5A is.
That being said, Archos bit off far more than it could chew in shooting for an October release for the A5A and the device came out with an almost non-functional firmware.
I understand the practical and financial reasons that forced their hand (by the time the firmware reached a shippable state, it was already Christmas Eve, they would have missed shipping their only profitable product for the critical Q4 in 2009 and we'd be discussing how it was too bad Archos is out of business now). But that's only an explanation, not an excuse. The FW should have been what it is now on release.
And now that Archos has a more mature Android codebase, let's hope we don't ever see a repeat of that situation (or at least not any time soon!
Anyway, good conversation!