AVaughan

I could see Elon aiming for a 2026 launch of their first attempt at landing a ship in Mars. Then aiming to send two Starships to Mars in 2028, each with prototype mining/ISRU fuel facilities, with the goal of demonstrating extraction of ice from Mars, and conversion to methane and oxygen, plus methane and oxygen transfer between two Starships, then maybe test fires of Raptor engines (one engine at a time, just to demonstrate that they still work after sitting on Mars for 12+ months). I'm sceptical of them achieving that timeline, but I could see Elon setting it as a goal. After that, them maybe you are ready to send a manned mission to the same site next launch window. (I guess they could be aiming for an ISRU type mission in 2026, but that seems ambitious, even for SpaceX). Of course this is SpaceX, whose stated goal is colonising Mars. So maybe they skip the second mission, and send their first colonists with 2+ years worth of supplies without demonstrating fuel production or a reasonable chance of a successful return to earth. Unless they are aiming for a prototype ISRU type mission in 2026, then this indefinite colonists approach seems like a possible (aspirational) option for them to achieve a 2028 manned landing.

Large language models generally don't understand anything about either mathematics or the concepts they sprout. I wouldn't trust them to actually calculate anything more complex than a straightforward unit conversion. Most of them can't even correctly multiply two large numbers together. ChatGPT once said that one of the advantage of an underwater telescope would be to study the stars without atmospheric distortion. Consider them good at tasks like "When the training data talks about topic X, this word tends to come next after this other word". That means they can often parrot back information that is already in their training data, but makes them terribly unreliable once you get to anything that is outside their training data. In general, they do not have enough understanding to actually calculate anything, they just parrot back whatever information was in their training data.

Isn't the plan for a south pole landing site? If so, then why would orbit timings mean they need long durations? Can't they launch from the pole to the correct orbit at pretty much any time?

Did they replace any engines between the extended test fire of the ship and the launch? An extended test fire at atmospheric pressure, (and within a flame trench which potentially reflects at least some of the acoustic energy back in the general direction of the engines) is probably pretty hard on the engines (and especially the engine bells of the Vac Raptors). I expect they did a visual inspection afterwards, but they might have missed something.

So is that a launch to the same inclination as other Starlink launches, but using a launch trajectory that cuts closer to land? (Presumably less of a dogleg, and so saving, and enabling larger payloads)? Or is it the same trajectory, just with a different landing location of the booster?

Then their use of release date varies from game to game. Eg see https://store.steampowered.com/app/427520/Factorio/ (On that page they use the 1.0 release date, but Factorio was in steam early access before that. Using the same term to mean different things is definitely bad, and I'm not sure that a judge wouldn't agree with my interpretation of the section I quoted. Do you have any source that says that an early access game is available for purchase through steam prior to early access release? ie is it possible to buy an early access game though steam before the early access release date? If it isn't, then the only way that paragraph I quoted makes sense (at least to me) is if the release date it mentions is the 1.0 release date. (And if Steam/Valve did not intend it to be read that way, then they should not have mentioned early access under the heading of titles purchased prior to the release date).

Quoting https://store.steampowered.com/steam_refunds I read that as saying that for an early access title, the 14 days does not start until the release date.

Is that the devs/publishers (internal to Valve) account with Valve/Steam or the devs/publishers actual (external to Valve) bank account. How would that work if Valve attempted to direct $50,000 worth of refunds to an external bank account with only $100 left in it. (Lets assume the dev expected this was coming eventually, and basically closed the company, and moved as many assets as possible to another company in preparation). I still think that there is no way Valve will willing assume liability to pay for mass refunds, unless they expect they can collect the cost of those refunds from the publisher/dev. (Valve might be fine just offsetting refunds against payments owed for sales in the current accounting period. Valve might even be willing to tolerate the occasional case of a dev deciding to discontinue development of a game triggering more refunds than can be covered by current and expected future sales. But automatic refunds for all abandoned early access games will mean a lot of games will suddenly generate large amounts of refunds that will not be able to offset against current or future sales. I can't see Valve willing assuming that financial burden without expecting to be able to recover the money from publishers/devs. And I expect that many small devs that abandon games just won't have the money to cover the cost of mass refunds. So at the very least attempting to force Valve to do mass refunds for abandoned early access games would force Valve to be careful which companies they allowed to run (paid) early access programs. In some ways that might be good, but it would most likely mean that (paid) early access is only available to companies/publishers which Valve is confident they can recover costs in the event of a mass refund, and to other companies if they agree to some sort of escrow service. (Valve could also allow other devs to run free early access/demo/unpaid beta type programs, which in some ways would be better for consumers than early access, but that won't allow devs to support themselves with early access). If I'm reading https://store.steampowered.com/steam_refunds correctly, then the Steam refund window for early access games extends to 14 days after 1.0, provided you have played less than 2 hours. (There are also plenty of reports on the internet of Steam processing refunds for people who have played more than 2 hours).

I do not see any realistic way they can guarantee that early access games will be finished. The only realistic way I can see for them to offer refunds for all abandoned early access games is to keep the developers funds in escrow until the title is released. In practise, any time I have attempted to do that, all I would achieve is giving the developer the extra 30%. (Which is fine, but publishers and developers do go out of business, and unless they offer a Steam key or a similar arrangement with a store I already use, I would prefer to use Steam rather than run the risk of not being able to download the game at some future time). Often, buying from Steam is cheaper than buying direct from the developer, once you consider Steam's regional pricing, and that developer storefronts typically don't price games in my local currency. Convenience also sells, and having most of my games on Steam is just more convenient than buying from multiple stores. (Yes GoG Galaxy might offer the same convenience as the Steam client, but I was already invested in the Steam ecosystem before GoG launched).

Seriously, I just don't get the "automatic refunds" bit. How do you think that would work? Do you expect Valve to assume that sort of financial risk? The only way Valve would willing offer time unlimited automatic refunds for all unfinished EA titles is if they were certain they could recoup that money from developers. If Valve tried to recoup money that developers had already received, that would lead to too many developers either marking games as finished (even if they aren't), or draining the companies bank account, and winding up the company/declaring bankruptcy before Valve could recover any money. In practice, the only way that sort of strategy would work, would be for Valve (or someone else) to hold developer's EA money in escrow, until the game was out of early access, which again would just mean that if a dev wanted to abandon a game, they would just mark it as finished, when it is not. (Now that might work for a large publisher that wants to test out an idea in early access, but absolutely would not work for a developer that is trying to work on a game full time, and is relying on early access sales to pay the bills). And if you are going to argue that a dev shouldn't be able to mark an unfinished game as finished, how are you going to enforce that? Lots of expensive court cases? With Valve (or someone), having to pay legal fees, and probably being unable to recover any funds even when they won, because an expensive court case would just push many small devs into bankruptcy anyway? I'm sorry, I just don't see ideas like that working in practice, and I don't see anyone convincing Valve to do that. If the US or the EU, or any other region tried, then I think Valve is more likely to simply stop offering EA games in those regions. They certainly would not willingly accept the financial risk of automatic refunds every time an early access game is abandoned by its developer. Isn't the current Early Access disclaimer on Steam clear that there is a real risk that the title may not receive any further updates, and that you should only buy if you want to play it in its current state? And if people ignore that and buy anyway? Well they have access to refunds for a limited time after purchase. (Also has the entire concept of caveat emptor been forgotten?)

So what did you buy? A flight proven Raptor engine maybe?

Weren't they intending to keep all the centre engines running in the initial plan for the flip manoeuvre? So a two burn test would have been appropriate. (And they probably knew they needed actual flight tests, since I doubt they could accurately replicate on test stands the actual conditions of the flip, including things like fuel slosh).

Why on Earth should the FAA require any such thing?. Do they require any other rocket company to release videos of any of their engines whilst firing or during restarts? Why should you expect them to require SpaceX to release such videos? And whilst I'm about it, why are you even surprised (let alone reacting like it is part of some conspiracy), if a rocket engine during shutdown might run some fuel through its turbopumps as part of its normal engine shutdown? Why do you act surprised if a rocket that contains liquified gas vents some of that gas? You do know that boiloff is a normal process for liquefied gas, right? And that tanks are designed for a certain maximum safe pressure and will at times need to vent to make sure that they never exceed that safe pressure? You know that methane gas is flammable, and, in the right mixture, even explosive? So why do you treat it as a conspiracy if any vented methane gas catches fire? Isn't that something you would expect might happen? Indeed, given the risk of explosion, isn't even conceivable that maybe SpaceX wants any vented methane gas to burn, to be sure it does not explode? Now I am not saying that I think Raptor is perfectly reliable, or that I think there aren't problems with the booster and ship design, but no-one should expect any commercial company want to release every detail of their design and all their test results to the public. Virtually no commercial enterprise is ever that open. On balance I think Elon and SpaceX are more open than most other commercial rocket companies.

Maybe I should have said "... as simple as you make it sound." To me, landing legs and a droneship or a concrete pad somewhere are simpler, lighter and less technically challenging. (Any the simple approach is personally very appealing). But ultimately they either will want either permission for re-entry/approach over land or a new launch + catch location where they can re-enter and approach from the West over water or uninhabited land, then launch eastward over water. Has anybody been spending $ buying land at the very south end of Florida lately? (Maybe around Key Lago?)

Whilst that might be technically possible, I'm not sure it is as simple as you suggest. If the ship breaks up during re-entry at least some parts probably decelerate faster than the intact ship, if you plan a re-entry with enough velocity that all the debris makes it to the coast, in all possible re-entry failure scenarios, what velocity would the ship still have when it crosses the coastline? How much dV would you need for the boostback burn? When would you start the boostback burn? What would be the ship velocity at that point? (Is the ship still hypersonic? Is it supersonic?) Can it handle the aero loads involved in that sort of flip, at that velocity? Is fuel slosh going to be a problem for engine start? Consider the reduced area of tiles on the sides of Starship where the catch arms might damage tiles during the catch. During the last re-entry, the steel rings that comprise Starship body seem to have deformed slightly suggesting that the steel might have softened during re-entry. If the steel is still hot in those regions, is that steel up to the aero loads involved in the flip? (Since Starship needs to be fully re-useable, is this deformation going to be an issue for future flights?) It might even be easier to aim the re-entry for the Gulf west of Florida, then do a burn to hop over Florida to Cape Canaveral. But if something goes wrong during that burn, you suddenly have debris landing in Florida, so that doesn't seem any better. I'm not sure how much extra fuel would be required, but I think it would be a lot more than just the few hundred m/s they reserve for the flip and landing burn. I doubt the header tanks hold enough fuel for both the boostback burn plus the flip and landing burns. I'm not sure they can use the main tanks to start the engines for the boostback burn, and if they start the engines from the header tanks, I'm not sure they can transfer fuel to refill the header tanks during the boostback burn, so they might need to change their tank/fuel flow/valve arrangements. (This is starting to sound like a redesign of Starship. And then they will want to remove all this extra mass eventually, since they won't want to make a boostback a permanent feature of re-entry). I'm not sure the current design has the margins during re-entry for all the extra fuel required for the boostback burn. They came close to burning through one of the flaps during at least one of the previous flights, and extra mass of more fuel would mean more momentum and kinetic energy to burn off, and that probably means a higher total heat load during re-entry. If they really want to do something like a boostback burn, then I would expect to see them attempt it during a re-entry before they do it for real. However I am confident that Musk would prefer not to do a boostback burn on a regular and on-going basis, so I expect him to attempt to push straight to a re-entry/approach over land. If he can't get that approved then I expect them to either build a catch tower somewhere else, or add landing legs and attempt to land somewhere else. (I will add that a catch tower just for Starship could be shorter and probably lighter and cheaper to build that what is needed for the full stack). Does he need FAA approval for a re-entry of Mexico? (I assume he does, since the FAA needs to approve the entire flightplan before they approve the launch). Does he need approval from Mexican authorities? Assuming he has approval from Mexican authorities, would a re-entry over Mexico be more acceptable to the FAA than one over Texas? What would be the liability situation if something did go wrong and debris killed someone in Mexico? Would the damages payout be cheaper?