> but there're no tests to check whether my adjustments are correct.
Start here. Start writing tests. Lots of tests. This is part of what makes it engineering and not just craftsmanship.
Once you have enough code coverage (this will take a long time), start rewriting small parts of the system. You will have your tests to fall back on to verify you new design is correct. For each part of the old system, once you are confident, replace with your new code.
This all needs to be communicated up front as a long term goal. It's not about a rewrite. It's about the tests and having high quality software.
Something you need to understand, they may not want or need high quality software. Quality costs money. Sometimes good enough is all they can afford right now, even though we know the tech debt will accumulate interest. A part of your pitch is to communicate how important quality is to the long term viability of the software.
Start here. Start writing tests. Lots of tests. This is part of what makes it engineering and not just craftsmanship.
Once you have enough code coverage (this will take a long time), start rewriting small parts of the system. You will have your tests to fall back on to verify you new design is correct. For each part of the old system, once you are confident, replace with your new code.
This all needs to be communicated up front as a long term goal. It's not about a rewrite. It's about the tests and having high quality software.
Something you need to understand, they may not want or need high quality software. Quality costs money. Sometimes good enough is all they can afford right now, even though we know the tech debt will accumulate interest. A part of your pitch is to communicate how important quality is to the long term viability of the software.