@PeterJones ,
This is the hardest part of doing the UDL’s I think. Even after we’ve done the original dBASE UDL you helped with, I’m finding that doing the earlier versions now, and fine-tuning the keywords, operators and the nesting precedents was the most confusing and finer control over the language. Doing the earlier versions I’ve figured out how to keep an apostrophe, from screwing up my code highlighting, while still keeping it as a delimiter by simply NOT allowing it to be included in the other delimiters nesting options. In turn, dBASE also has a JSON object, and with a little fine tuning, I was able to make the delimiters in dBASE also work to outline a JSON structure, so far. If there are other options in a JSON structure I’ll have to learn that as I go, but for now, the simple W3C version showing it, it works right, as well as the dBASE delimiters without messing things up.
Obviously a quick UDL is possible with just the categories to put things in, but I’m finding that the real power of the UDL system is in the finer use of the mechanisms like the nesting, differentiation of forward search/backward search, prefix mode, etc works much better when combined to get a desired result. It’s actually much more powerful than I originally figured. I wish I’d had a good explanation/example like you’ve done, back then…although, with so much to include to make a UDL almost a complete language system, it would have probably been too much to absorb at the time and would not have availed me much improvement at the time. It’s something that requires a lot of trial and error and understanding of the mechanisms and time playing with it, I think, to get things just right, if it’s at all possible to do. :-)