OSNews

Palm didn't like it because it was not fully compatible with older Palm apps. However, I am sure that their new implementation is not 100% compatible either, I guess they learned that the hard way if they didn't believe Palmsource.

Short of total hardware virtualization, there is no way they could run all old Palm apps. Programming for the classic palm was more like programming for Dos. Apps had very low-level access to things. Classic apps were written in C with our friends Malloc() and Free(). Pointers everywhere, low-level database record access. I think the same can be said for classic Dos apps running on Windows 3.1. Many could, but certains apps couldn't because of their low-level access to system resources.

The market decided. Cobalt wasn't what customers wanted when they wanted it.

Possible Cobalt customers cited a wide number of reasons for why they didn't buy in, but mostly they seemed not to think that the new feature set was worth the risk.

By "the market" you must not mean actual PDA users, because Cobalt-based PDAs were never put into production. I used the Cobalt simulator, and I rather liked it. I've seen some apps on PalmGear.com that were developed for PalmOS 6, so the only PDA they have to run on is the simulator. I've read a lot of people's excited comments about "Cobalt is ready! It's cool! Alright!" and then "Where is a Cobalt device?" and then "There's gonna be a Cobalt device in November!" (Garmin being the company that actually announced one) and finally, "It's March and still no Cobalt device... what happened?!?"

Someone here explained to me in another thread some month ago that the problem with Cobalt is that only four people knew how to write a device driver for it, and they all worked at PalmSource. Someone else said that the problem was that PalmSource wanted a lot of money up front to license the OS. At this point I don't think anyone really knows what happened, except the management at companies that decided not to implement it.