I've been Newton developer since the launch in 1993. I can't speak for all Newton programmers, but I can tell you my perspective. My apologies if this sounds whiney; it is. I'm still mad about the whole thing.
We are looking at the alternatives to the Newton platform, but the alternatives are a real step backwards.
The Newton programming application, although flawed, gave us a taste of what software development could be - faster, more efficient, allowing us to focus more on ideas and design and less on mechanics.
The current alternatives, the Pilot and WinCE devices, lag significantly behind; while porting our application to WinCE may be possible, it may not be feasible; instead, we'll have to come up with an uglier program with less functionality. Worse, I have the sense that because the WinCE platform is less technology-driven and more market-driven (read: contains less innovation), its development environment will continue to be based around old, ugly technology, and Java, which shows some promise because of its open nature (however, this promise is being willfully ruined by Microsoft in its refusal to maintain the open nature of Java.)
I think the end result is that we will continue to pursue handheld applications, but not with the gusto, and certainly not with the enjoyment, that we once did.
Although the Newton programming tools took a lot of criticism for being quirky and non-standard, there were good reasons that they were. The original Newton team was trying to find a better way to write applications. They succeeded, but at a cost: developers couldn't write applications with C++, or Visual BASIC -- the most popular tools for PC applications. These tools are widely used, but many programmers also believe they are too awkward, difficult, limited, and time-consuming: they make the software development cycle excessively long and expensive, and tend to lead to hard-to-find bugs.
To programmers with a background in object-oriented programming, especially those with some exposure to Lisp or SmallTalk, NewtonScript was quite easy to learn. Its beauty was that it combined a very small set of powerful operations and built-in handling of data structures with a minimal set of tools necessary for constructing elegant object-oriented applications. The Newton design supports component reuse, with a far, far simpler design than other component-software systems such as Microsoft's COM or Apple's now-defunct OpenDoc.
Because NewtonScript is dynamic, and your programs can actually compile and execute NewtonScript code itself on the fly, or change the context and structure of its own code, it is exceedingly useful for building programs that change their behavior over time. A good example of this is an expert system, which allows the user to specify a set of rules governing the program's operation. Another is a program that contains its own scripting language.
NewtonScript provides much of the power of Lisp with a fraction of the complexity for the programmer. The extensive use of components in Newton Toolkit give you similar functionality to Visual BASIC, where much of your application can simply be drawn.
I don't think management at Apple ever quite realized what it had, and customers didn't care, but developers did. We also realize what Apple could have had, if the platform had been marketed, understood, and positioned correctly: NewtonScript could have had part of Java's thunder; Windows CE could have not gotten off the ground; Apple could have owned the Pilot's market share. Instead, we have mediocrities, corporate infighting, and the realization that technical excellence rarely gets the nurturing it needs to truly succeed in the marketplace. That's probably the hardest part to get over.