Suggestion for revising documentation on writing custom parsers

https://github.com/boostorg/parser/blob/f468d529fe440313d9d67d092f360b1a64a6f702/doc/tutorial.qbk#L3792

Thank you for your work on Boost Parser. It's an impressive library, and I've found it quite powerful.

While I appreciate the humor in the writing, I found the tone to be slightly dismissive of scenarios where writing custom parsers might genuinely be the most effective solution. For instance, I am working on parsing SystemVerilog code, which includes number literals with unique syntax, such as `5 'D 3`. Parsing such constructs character by character using char_ primitives seems neither convenient nor effective, and it would make a strong case for writing a custom parser akin to `uint_`.

I believe this use case illustrates that there are situations where users might reasonably need more guidance on creating low-level parsers. If this section of the documentation could include:

1. A more neutral tone when addressing the use of custom parsers.
2. An example of how to effectively handle complex cases (like SystemVerilog number literals) using the library's primitives or via a custom parser.

This approach would make the documentation more inclusive and better serve a wider range of users.

Thank you again for your efforts, and I hope my feedback helps in further refining the excellent resources you provide.