- "100% of the software would be free software, and third parties would be encouraged to set up their own installations" - I'm planning on open sourcing it under AGPL soon, once I've got documentation, testing etc. ready. Plus it's easy to set up your own installation (git clone; mkdirs for data; docker-compose up -d).
- "I would not have it crawling the entire web from the outset" - That's one of the key features of my approach, only crawling submitted domains. I'm focussing on personal websites and independent websites at the moment, primarily because I don't currently have the money for infra to crawl big but useful sites like wikipedia, but there's nothing to stop people setting up their own instances for other types of site.
- "who’s going to pay for it? Advertisements or paid results are not going to fly" - A tough anti-advert stance is another key differentiating feature to try to keep out spam, e.g. I detect adverts on indexed pages and make sure those pages are heavily downranked. Planning to pay running costs via a listing fee, which gives access to additional features like greater control over indexing (e.g. being able to trigger reindexing on demand).