I think the Saleae Logic software should have a kind of DIY Logic Decoder Module "built in," that would allow anyone to easily design their own decoder without resorting to C.
The protocols included in Saleae Logic software are pre-set patterns that have been set down by an official international organization such as ANSI. If you're dealing with a protocol such as CAN (very big in the automotive industry), then of course that protocol is 100% compliant to industry standards. It's easy for Saleae to incorporate that into their software because it's easy to get a device that outputs the CAN protocols to test it. There can be no deviation.
However, if CAN were missing from Saleae, you could get the datasheet and sit down and write out some C code in accordance to the standard and perhaps within a few days you'd be up and running. That is, if you're a expert at Microsoft Visual C++.
However, real life has thrown me some curve balls that no logic analyser (as far as I know that is) has ever reasonably dealt with. There are protocols out in the concrete jungle that mother never told me about.
Take for instance, I2S audio. It dates back to the 80's. It's as old as the midi music standard. However, modern day engineers have resolved to get a few more miles out of this old dog by using something known as Time Division Multiplexing. TDM allows you to use the same two basic channels that the I2S data stream uses to add in multiple channels of audio by slicing up the audio data into very fine pieces and sending them down the pipe in a very neat row. However, each company has it's own unique way of doing this and it all starts on the hardware level with the ADC. Some IC's have multiple ADC's and some cut costs by only having two ADC and doing the TDM within the pre-mixer/multiplexer buffer. They all have one thing in common, the data sheets are pretty poor. You really need access to the engineers at the company in order work with these chips and they don't have much time for anyone else.
I can't begin to tell you how amazingly difficult it is working with Time Division Multiplexed audio streams. DSP can be just as difficult. There are no resources available. It would be nice if Saleae took the initiative to deal with such unknown protocols. It would really open a world of possibilities. Perhaps it could start experimentally with pattern recognition routines?