Whilst the aim to produce one-logging-framework-to-rule-them-all is laudable, it has been tried and failed many times.
Producing yet another logging framework (yalf would be a good name if you do), will not reduce the need for bridges, but will just create yet another set of bridges, so that the logging library of choice for each dependency can be mapped to yalf.
So how about taking a slightly different approach. Rather than specifying how logs are collected, via API, why not instead come up with a standard of how to configure an arbitrary logging system.
This could then be directly implemented by all/most existing logging frameworks and configuration bridges written for those that do not adopt it.
Such a standard would need to include naming, formatting and configuration. It need not include an actual implementation.
