The most egregious cases of violating this involve trying to create your own CMSs, search engines, and God forbid - databases. There are of course, exceptions. AirTable wrote their OWN database engine, but they knew what they were getting into, and they did it right.
I can't speak to CMS, but are search engines and databases really such a big deal? If postgres or influx or something kinda fits your workload then it's probably not worth the time, but on the other hand I've seen both custom databases and the cludges people have written to work around off-the-shelf DBs, and I vastly prefer the former both in terms of performance and maintainability in a problem space (similarly with search engines).
The problem with search engines is that people think it's just tokenizing the search query and then doing an AND query for the words, but then the fun starts. First of all, there is specific search theory that one has to dive into and have the Knuth tomes nearby. Then, different languages, thesaurus, taxonomies, "did you mean". And we are not even at indexing speed, replication, performance, etc. There is absolutely no chance you would be able to create and maintain a reasonably good search engine part time.
