85 comments

[ 5.4 ms ] story [ 140 ms ] thread
Engineers are capable of launching things into space, but the hamburger menu is only half visible on my IPhone XS running safari. What’s harder and more complex these days, rocket science or front-end development?
SpaceX has one deployment target, how many do front-end devs have? :P
Variables: time, weather, fuel temp, manufacturing tolerances, turbopump response time to commands, gimbal response time to commands, etc.

Targets: inclination, semimajor axis of altitudes, orbital eccentricity, longitude of the ascending node, true anomaly adjusted for launch time and ascent profile, argument of the periapsis.

So that's six deployment targets per satellite deployed, on hardware that is exposed to vacuum, variable 0-3 G's acceleration between both stages, heavy vibration and acoustical forces, temperature swings, radiation, etc.

However, I concede, SpaceX spaceflight engineers do _not_ have to target IE 6 anymore.

How much money is SpaceX putting into rocket science vs front-end development?
I doubt that all the problems in the world are fixed with money :) I've been to rich companies with crappy websites way too many times.

I assume the company has put the focus on the rockets ( as it should ) instead of the marketing, although they ( probably ) pay enough for developers / QA that can deliver the task right.

This problem in particular is directly related to the resources you are putting into solving it. More/better FE devs, better QA, more time to develop etc. There's a reason you will rarely see a broken hamburger menu on a Google or Facebook site, even though those companies aren't launching rockets.
The Engineers are clearly busy making rockets :P
I'll bite; the cpu power needed to operate and render that hamburger menu is many times more than was used on the Apollo missions to the moon. :p

We've been going into space for 60 years, we've only gone to cyberspace for 30.

I'll bite; the cpu power needed to operate and render that hamburger menu is many times more than was used on the Apollo missions to the moon. :p

Is that because the problem is hard, or because the programmers have iteratively built up their laziness?

At the time of Apollo missions computers were heavy so a simple one was put on the craft and a larger one on earth did the bulk of the computations. And even the ones on earth were limited due to technology at that time.
We don't code to the metal anymore, because we're optimizing for developer time (supposedly). So, there are many more moving parts and it's harder for someone to understand the full stack. Also, the whole dev process for a non-life-critical website is different than for safety-critical software (hopefully).

  > there are many more moving parts
It's actually many more layers of abstraction.

My current PM would code in ABL in the 1980s. Adding a field to a database and a CRUD app would take about two months, running on terminal hardware that the actual worker using it could not afford if he were to save up for an entire year.

Today, I add a field and an interface in about an hour, including automated testing for regressions, on hardware that the actual worker using it could purchase on a whim.

Css is hard, by proxy, ui is hard :)
Haha. I had the exact same thought. Lots of images with blocks of white behind white text. And the font is horrible. I guess sacrifices have to be made somewhere when you’re changing the world.
The hard part about rocket science is fluid Dynamics, and it's only hard to land.

But my God do the engineering equations for fluid Dynamics are nuts when you have an open, changing, turbulent system.

i mean material science, computer science/ee (guidance) and many other disciplines are important to launching a useful rocket
There's a great video from a few years ago on how SpaceX accelerates their CFD with image compression based algorithms and some GPU tricks.
They are definitely in the need of extra cash from smallsat market.
The transferable nature of the ride seems like a big gamble pared with their statement that launches won't be held up by co-passengers. Guess they're just counting on the odds that enough passengers will drop out to make the launch uneconomic will be fairly low.
It makes me wonder if they have any projects planned that they can throw in if there are empty slots.
Maybe that would help the economics quite a bit but I'm not sure what they'd be planning to use it for. SSO is a lot of earth observation (or sun observation for some specific SSO orbits that ride the terminator) and that doesn't really fit anything they're doing and there's already a couple companies in the low cost earth observation game.
Maybe they're planning to fill vacancies with their own payloads that are not time sensitive?
Maybe, I'm not sure how useful a SSO is to SpaceX though.
Starlink sats can maneuver to different orbits. Their original test sats launched on an SSO launch.
The Starlink satellites are using hundreds of m/s raising and lowering their orbit. The sort of plane change you're talking about would take something close to a thousand, conservatively.
Not if it's a near-polar orbit. They'll probably need some in that orbit eventually.
They're only going up to 84 degrees inclination [0]. Not surprising there's an absolutely tiny number of customers that would use it if they added enough satellites to make 100% coverage.

[0] https://en.wikipedia.org/wiki/Starlink_(satellite_constellat...

Hmmm, that is a relatively high delta-v to change inclination that much (~1.5-2 km/s for a ~12-15 degree difference). On the other hand, if the shared payloads are light enough (and they very well may be, particularly if the launch is undersubscribed), the upper stage could do the orbital transfer and should have plenty of delta-v to do so (particularly with a droneship landing) considering F9 regularly does recoverable flights with many tons through 2.5km/s above LEO.
They can but doing so reduces their operational life time so they'll want to minimize the amount of the thrust budget they expend on slot filling launches. As someone else pointed out their prices are high enough to cover the cost of a reusable launch with maybe half the ports filled so they're probably just playing the numbers game assuming there won't be a large coordinated drop out of customers. And those prices don't include the main top port which will likely cover a fair portion of the operation cost of the launch by itself.
This seems likely. If customers drop out, they can just toss in a couple starlink sats.
SSO != Starlink orbits.
They launched tintin A and B on SSO-A. They maneuvered to the correct orbits.
another comment mentions starlink launches. If spaceX has 12000 satellites of their own that they need to send up, they can fill up a fair bit of unused capacity.

it sounds like they're essentially selling the opportunity to take the place of a starlink satellite in any given launch. if the customer cancels, no big deal - they'll just send the originally scheduled starlink satellite.

(comment deleted)
This looks very different from their Starlink launcher though and doesn't look like it could live above or below it either. Sun synchronous orbit is a pretty specific orbit too.
Yeah, there aren't any plans to put Starlink satellites in SSO.
Yeah, the Starlink launcher was a pretty specific device designed around the Starlink satellites. The attachments shown in the attached link are the ESPA (EELV (Evolved Expendable Launch Vehicle) Secondary Payload Adapter) [1] rings. They work well for a variety of customers using it because you can put out a generic document for an interface and a box you need to fit within. The first dedicated rideshare program like this was the STP-2 mission. [2]

[1] https://en.wikipedia.org/wiki/EELV_Secondary_Payload_Adapter

[2] http://www.planetary.org/multimedia/space-images/spacecraft/...

In their graphic, I count 7 15" ports and 8 24" ports, which at their before L-12 pricing is over $50 million. There's also the large port on top, which doesn't list pricing, but is presumably quite a bit more. The L-12 to L-6 pricing would be more as well. I believe their cost for a launch is something in the neighborhood of $30 million so they have quite a bit of margin to make money on this. Especially now that they're routinely reusing first stages and will probably use these flights as the final send-off for stages that are end-of-life.
This is one launch per year. They already launch about 20 per year for the last few years. The extra cash here is minimal.
What does L-12 and L-6 from the pricing section refer to?
Launch date minus 12 or 6 months.
Not Lagrange points?
I was thinking Launch number.
There are only five Lagrange points, so L-6 and above are not Lagrange points. :-)
Haha can't argue with that! :) Pardon my space-rookieness.
(comment deleted)
With these prices it looks like they're not targeting cubesats, which I'm sure is a relief to the existing rideshare companies. It will be interesting to see how good they are at launching on time. Since SpaceX can shift around Starlink launches to help their schedule, they might have a bit of an advantage over conventional rideshare companies.

>For payloads who run into development or production challenges leading up to launch, SpaceX will allow them to apply 100% of monies paid towards the cost of rebooking on a subsequent mission (rebooking fees may apply).

Without knowing the magnitude of the rebooking fees this isn't a very informative statement.

How large is the CubeSat launch market?

Also, they do say this:

> Cubesats can be aggregated and launched on a customer provided, ESPA-compatible deployer.

Relatively small but very rapidly growing (like the smallsat market in general). A decade ago there were a dozen or so cubesats launched per year. Now there's a few hundred, and that number is projected to continue to grow [1].

Personally I do think that the small launcher market is a bit of a VC-subsidized bubble— I suspect that a lot of the companies in the space are going to fold over the next few years.

[1]: https://en.wikipedia.org/wiki/CubeSat#/media/File:Nanosatell...

Heliosync orbit suggests focus on visual Earth observation and surveillance type missions.
I think you are thinking of geosync?
No. Geosynchronous orbits are so far away that cameras can't see much detail. Sun-synchronous low earth orbits are much closer and they always pass over the light side of the planet allowing for consistent observation.
Ah I misread and was thinking this was heliocentric orbits.
(comment deleted)
Because they are posting pricing, I'm curious about the details. What happens if the rocket fails and loses the payload? Is there insurance included/available? Does the customer get a refund?
As with a regular satellite launch, the customer usually buys their own insurance? SpaceX would be on the hook for the launch costs though.
I think that's on a contract-by-contract basis. IIRC NASA still had to pay 80% of the costs for CRS-7, but Spacecom got a free launch after their payload was destroyed (that was kind of a weird case since the explosion happened before the point where their launch insurance would've kicked in, and the damage to the company was much larger than just the cost of replacing the payload)
If I'm not mistaken, the company supplying the payload usually has to insure the payload, and depending on the launch companies track record (and therefore risk) they get better/worse conditions. Of course the insurance company will then try to get their money from the launch company in the case of failure.
I don’t think insurers ever get money from launch providers.
That's what any insurer worth their salt at least tries to do, isn't it?

They try to sue and prove negligence of the party creating the damage. The damaging party then tries to prove that they were not negligent and/or try to prove that someone else was (e.g. the supplier producing the struts in that one SpaceX launch failure).

Well, insurance is a social game, usually. Everyone pays into the pot and if there is a launch failure, the cost is covered out of this pot. They might go after the launch company to get some of the pot back full but ultimately, so many things can go wrong in rocket science that tracking down negligence is hard and it's usually a sum of 20 different factors that are hard to call "negligence" in most cases. Of course, most of the famous disasters usually are a result of that but for example most cases of a SpaceX booster being lost on landing is simply one of a billion things being 1% out of alignment.
Can you point out an example? The amounts of money are large enough that I’d think it would be in the press.
Satellite insurance is a fairly big industry. The owner of the satellite being launched is responsible for getting said insurance against the satellite. This can cover anomalies during launch [1] or anomalies on orbit where the satellite is lost or can no longer perform its mission [2]. Pretty much all commercial companies will have insurance on their satellites, government satellites are typically "self-insured".

[1] https://www.insurancebusinessmag.com/us/news/breaking-news/s... [2] https://spacenews.com/digitalglobe-loses-worldview-4-satelli...

(comment deleted)
Coincidental that you mention this. AMOS-6, launching tomorrow on an F9 is getting a free ride, since the previous version of the satellite blew up on the pad.
The insurance for that one got complicated. It was destroyed in a static fire test that would have been a few days before the launch. Their launch insurance wouldn't cover it, because even though it was destroyed by a rocket exploding on the launch pad, it wasn't actually going to launch. They also had marine cargo insurance to cover the satellite before then. However, pre-launch static fires with payload attached are pretty unusual in the launch business and the insurer had no idea this was going to be done. Arguing that they were unaware of the increased risk and that the insured failed to notify them as required, they only made a partial payment on the claim. A lawsuit ensued (no surprise with ~$100 million at stake) and is apparently still ongoing.
What is the rough breakdown of applications for all these cubesats and other satellites?
Also curious what the rough breakdown of applicTions is... somehow you have several replies and most about a hamburger menu.

What are satellites even used for these days besides navigation and espionage (limited to a handful of large players, I assume)?

Communications is a huge one. Think satellite TV, telephone, and Internet services.

From the governmental side, another major application is Earth observation, both for science and for weather forecasting.

Sun-Synchronous Orbit... huh, that one you can't do in Kerbal Space Program.

Turns out SSO is exploiting the fact that Earth isn't a perfect sphere, and so you can use the non-spherically-symmetric gravity field to induce precession in your orbital plane. If you tune the precession go 360° per year, you get a Sun-Synchronous Orbit.

Decent.

--

[0] - https://en.wikipedia.org/wiki/Sun-synchronous_orbit#Technica...

Indeed it doesn't work in stock KSP, but it works using a mod called Principia[1] which implements N-body physics!

--

[1] https://forum.kerbalspaceprogram.com/index.php?/topic/162200...

> This initial configuration also includes J2 for the Sun, the planets, the Moon, and Vesta, so the resulting effects are felt (precession of Earth orbits, the possibility of heliosynchronous orbits, etc.).

Wow <3. I though the mod only did n-body physics, but it seems it goes beyond that. Thanks!