Behrooz

THAT explains it. I quit playing last night after going to do some upgrades and seeing that the level one upgrades cost 270,000 each and resolved to quit playing until the no-grind mod was updated.

This edge case already exists within the current system. Ideally, closing the cargo bay around something that fits entirely inside auto-docks and simulates it as part of the cargo bay until the bay is re-opened, same as attached parts inside the bay. Worst case, ignore it and accept the clipping/collision problems you'd encounter with the same scenario in the existing implementation.

As I was reading the dev notes today, one of the specific notes Harvester had is that v1.0 cargo bays will become more useful because they will now properly shield their contents from aerodynamic forces when closed. I think this concept should be extended to include cargo tie-downs. It makes sense that part of the standard gear of a cargo bay should be equipment for securing a load under any reasonable conditions-- anything up to where the cargo bay itself is in danger of structural failure. This leads to the following conclusions: While a cargo bay is closed, it should no longer be necessary to simulate the individual parts that it contains. All of its contents can and should be treated as physics-less, and their mass distribution summed and aggregated, calculated for physics purposes as a single virtual part locked to the cargo bay part they are contained within. Fuel use or other mass redistributions for shielded parts should be calculated to affect the aggregate. We don't need to render or calculate collisions for parts that are contained within cargo bays. The potential for performance improvements is excellent, especially during launch when part counts and lag are highest, and cargo bays will nearly always be closed. The above improvements can also apply to mods with ejectable fairings or other parts that aero-shield. After all, ejectable fairings are just a special case where the cargo bay starts closed and opens irreversibly. Procedural fairings is a must-have mod, and would become even better with these features. Linked open-ended cargo bays should operate their doors as a single unit, and should be closed off at their ends. Cargo bays should be tweakable to be open-ended or not, and link or not at either end. Right now, you have to terminate a cargo bay with another similar-shaped part or the end will be open to space-- make everyone's lives easier. Cargo bay structural failure could become substantially more exciting. Under the right kinds of stress or after taking enough damage, the bay doors should fail, popping open and causing the contents to Kerbalize rapidly. Used properly, implementation of cargo bays has the potential to reduce micromanagement worry about game-y structural problems, improve performance, and make game mechanics more intuitive. I think these features would be worth the coding time.

This has already murdered the forums. Steam, however, is entirely unconcerned, as the difference between current usage and peak usage for the last 24 hours gives us 20Mbps worth of headroom for each of the 115,000 users on the KSP forums. So, if everyone on the forums re-downloaded the entire 1.5GB KSP install in ten minutes, we'd almost make Steam hit its 24-hour peak. Steam is the Kerbal equivalent of 20.0 meter parts. That are entirely made of tubes. That are entirely made of bacon.