Back before the download manager, the most downloaded OXP on that list was almost certainly BGS - at less than 10% of the number of Oolite 1.80 downloads we could track (e.g. not counting ones installed via Linux distributions, or the various mirror sites outside our control).
The download manager certainly made OXPs more visible - some of mine tripled or so in downloads - but I'd be very surprised if any OXP, even the most popular, made it over 20%.
As regards adding OXPs to the core functionality, this has happened occasionally, but generally with fairly small self-contained OXPs and not necessarily by just inserting the OXP code. e.g. there used to be a few "tell me what the ASC is targeting" OXPs which were fairly popular and then were superseded by adding a new core HUD indicator to do just that. Considerations for doing so in future would probably include things like:
- is the OXP code and assets suitable for inclusion (licensing, quality, etc.)
- is the code itself extensible or possible to make extensible for future OXPs to modify (may require converting JS code into Obj-C code for the core engine)
- does it significantly increase the amount of in-game text for translations? (Is it even practically translatable?)
- is it written to the latest OXP best practice standards or would it need a significant rewrite?
- does it work well on non-shader computers or low-spec ones?
- would it cause incompatibilities for existing OXPs? (and what happens if someone already has the OXP installed?)