One thing I like about Lycan is that he doesn't only give the corporate line, he gives his own opinion too. Too many companies don't let their folks do that and as a result, customers and interested parties feel like they're talking to a wall or a robot.

We're all here to learn, and reasonable people can be agreeable and yet still disagree.
John,
Your expectations are ridiculous.
No, his expectations are sane. He may not be aware that his expectations are not optimum, we'll get into that next...

First powering on and off your hardware is a NO NO. If the device was meant to be powered off, it would have had a switch.
He's probably not alone -- often-repeated advice in the media has been to power down your devices when not in use to save money and energy, especially if the device has a power adapter or a stand-by mode, unplug it even if the device has a power-switch!
While I've never tested the consumption either way, D-Link has been responding to the world's desire to conserve power. The DIR-655 has many ground-breaking, energy-saving features not enjoyed by D-Links earlier routers like the DI-624 -- but none of them had power switches.
Even if this does shorten the DIR-655's lifespan (it might), it certainly is reasonable to expect that if the router and modem are left on, that after a power failure the router will boot up in a state allowing communications to continue. If it doesn't, that would affect my purchase decision. This is also a real simple test to perform, and the firmware development team ought to be running a test suite pre-release that includes that test.
(I am not confirming that this does or does not occur for some. My configuration is unique and after a power failure, my system seems to be fine.)
Second, you're complaining about the modem and the router not booting in sync with one another, of course they don't. Ones a modem and the other is a sophisticated router with lots of things going on.
Today's modems are quite sophisticated themselves! Most are hybrid devices (modem/VOIP phone/router/coffee-maker). Some of the newest ones are now joining multiple data streams to get faster peak speeds.
There's nothing wrong with the firmware (well there is, but it's nothing related to your issue), you need to understand that this router is designed to be powered on and LEFT on. the firmware was built around that idea.
Yes, I agree. This isn't the device you should unplug to save power or cycle on/off regularly to save power.
You're comparing a 624 (old school) router and another manufacturers device to ours and making a judgment based on dissimilar boot sequences?
Requirements and features are assessed in terms of MUST, SHOULD, and COULD. Everyone's grading scale will be different. Friendly to turning off/on frequently would be a SHOULD requirement for me, but recovering from a power failure into a working state would be a MUST requirement.
The 624 is a busybox device, the 655 is a Ubicom device and their YEARS apart from one another.
The DI-624 isn't a busybox device (busybox devices would require source releases under the GPL). But your main point is valid, the DI-624 is a much simpler device and capability-limited device.
I'm glad that John moved on to try something else -- it's what he should have done given the circumstances and his requirements. They are important requirements to him, and what he was trying to do -- whether the best way to do it or not -- really should have worked.