From an end-user perspective, bricked. I simply have a hard time calling something bricked if it has a working bootloader.

I'm not exactly certain why the kernel would be an issue here. Perhaps someone with an understanding of the workings of the T1 can address that point. (Outside of kernel modules, I've never run into difficulty replacing the kernel on other systems.) Yet it did look like the partition map was changed between different versions of the firmware. I was also running into difficulties because the firmware is encrypted.

This is completely contrary to my experiences with Kobo's firmware. I agree with you that Kobo is more open, albeit for different reasons.

First to elaborate on your comment about downgrading. Kobo devices contain two full copies of the firmware: the working copy, and the factory firmware. Updates are only applied to the working copy, so you can always force a factory reset to the version of the firmware that your device shipped then upgrade to the desired version. In other words, you have a downgrade path. It is also handy if you "brick" your ereader while modifying the software, since it is virtually always possible to return to the factory firmware.

Yet Kobo goes one step further: they don't encrypt their updates. You can unpack those updates to see what has changed and make your own changes. With the T1 this is much more difficult to do before applying an update since you need specialized tools to handle the special file format and encryption. In the sense that anyone with a knowledge of Unix utilities can modify Kobo systems, I would consider Kobo devices to be much more open.

As for the actual reader software, Kobo is definitely a closed system.