This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 |
@ Jadebenn:Since the above conversation has become unreadable and we seem to have come to an agreement concerning the Program Cost so far, I think starting a new section focusing only on Launch Cost is a good idea. For now I am going to delete the Lunch Cost altogether until we come to an agreement (ie you come to your senses). It's all estimates in the best of cases anyway. So you admit that the NASA administrator speech is "a weak citation by itself" and add that it should somehow be lumped with the other (weak) references to make your case. multiple anecdotal evidence do not amount to a good argument. This is a fallacy. So I'm gonna take your references one by one!
@ Jadebenn: I just realised how much you've been acting like you own this page, I am at least the 6th person in the last 6 months that tries to revert this figure to the real one :
1) https://en.wikipedia.org/?title=Space_Launch_System&diff=925620594&oldid=925395840
2) https://en.wikipedia.org/?title=Space_Launch_System&diff=925963019&oldid=925876896
3) https://en.wikipedia.org/?title=Space_Launch_System&diff=926359702&oldid=926277355
4) https://en.wikipedia.org/?title=Space_Launch_System&diff=927943530&oldid=927446682
5) https://en.wikipedia.org/?title=Space_Launch_System&diff=955609445&oldid=955286614
and me.
Which side is the consensus on? Moamem ( talk) 13:37, 29 May 2020 (UTC)
That "marginal cost" figure is the only launch cost we have. Again, SLS development will be complete the second SLS leaves the pad, just as Shuttle development was complete the second it left the pad. There are still many unrelated costs to launch. BOLE will be part of that $2B/year until the 8th launch for example. The RS-25E and EUS until its fourth. Things like eCryo also recieve funding from the SLS program. The cost to develop those technologies is totally separate from how much money it takes NASA to buy an SLS rocket and launch it. It's a similar situation with that "$100M/year" RS-25 cost. The unit price of an engine will be lower than that, exactly because that figure includes things like personnel costs and R&D. As Gopher65 said, there's a difference between program cost per launch and the cost of something itself. If SLS cadence increases to twice per year, for instance, the launch cost will remain roughly the same, but the program cost per launch will decrease significantly. Thus, using it as the "cost per launch" metric is misleading at best.
However, we seem to have once again reached an impasse. To break this stalemate, I suggest we solicit a third opinion from a neutral observer. Would you agree to abide by their decision if I will? – Jadebenn ( talk · contribs · subpages) 18:59, 31 May 2020 (UTC)
How's this for a compromise, set the cost per launch in the infobox to $500 million - $2 billion with a link to a note that explains total program cost vs. program operating cost per flight vs. marginal cost of adding a flight to the manifest? Eggsaladsandwich ( talk) 20:53, 31 May 2020 (UTC)
Moamem Are you OK with removing the "(2019 estimate)" from the cost per launch item in the infobox? The timing of the estimate is reflected in the reference. Eggsaladsandwich ( talk) 01:35, 3 June 2020 (UTC)
I, too, have come over here as a result of an invitation on my Talk page. I have read through the above comments in this section, from 28 Mar to early June. Do you feel you all reached consensus and the article now reflects that consensus? If not, I'd be happy to provide some input to what seems to be a complicated discussion. Pinging Moamem, Eggsaladsandwich, Jadebenn. Cheers. N2e ( talk) 16:35, 25 June 2020 (UTC)
@
N2e: I believe a consensus was reached. –
Jadebenn (
talk ·
contribs ·
subpages)
19:26, 25 June 2020 (UTC)
@ N2e::Hey, as I said the $500 million figure is very concerning to me, it was an aspirational figure that is and never will be achievable since the engines alone cost $400 million excluding development and production restart costs. I just got tired of bickering alone with jadebenn, if that's considered consensus, I don't know... Would love to hear your opinion tho. Moamem ( talk) 02:34, 27 June 2020 (UTC)
@ N2e:I think we've reached a workable compromise solution. Eggsaladsandwich ( talk) 20:56, 1 July 2020 (UTC)
N2e Regarding your comment on the Habex Final Report. The cost analysis is actually documented in Chapter 9 of that report. Appendix G contains the results of an additional independent cost estimate study. The PDF states that appendix is withheld "due to U.S. Export Regulations". It is unlikely that appendix will ever be publicly available. FWIW, the version in the article is identical to the document available here: https://www.jpl.nasa.gov/habex/pdf/HabEx-Final-Report-Public-Release.pdf — Preceding unsigned comment added by Eggsaladsandwich ( talk • contribs) 21:44, 5 July 2020 (UTC)
References
The follow-on contract to produce 18 engines is valued at $1.79 billion. This includes labor to build and test the engines, produce tooling and support SLS flights powered by the engines
So, according to the space agency, NASA has spent $3.5 billion for a total of 24 rocket engines. That comes to $146 million per engine.
WRT Launch cost...
The problem you are running into is that there are three different ways of calculating launch cost. There is the commercial concept - how much it would cost to launch a specific payload if you contracted with a commercial provider? NASA has no analog in their costing structure, so if you are trying to find something like that, I think you are destined to run into issues. Note that trying to compare NASA programs to commercial costs is a bit problematic as a NASA program on a commercial provider has many costs outside of launch costs because of NASA overhead.
The "per launch" price that NASA quotes is what is most useful for their budgetary planning; the incremental price of launching one more or one fewer times during a year. At the end of the shuttle, that number was $450 million, but the only number of this type for SLS was an aspirational number early in the program, back when they though things would be cheaper and were planning on launching twice a year. NASA has chosen not to provide an updated estimate of this type; it can be argued that that is a good thing as SLS is a bit weird in that the earlier launches that reused hardware leftover from shuttle are likely to be much cheaper than later flights. There's also the issue that NASA hasn't known main engine costs until recently.
Which leads us to what I call "yearly program-based cost", which is simply looking at the per-year program cost and dividing it by the launch rate. This is my preferred way of looking at program cost as the number is easy to get and hard to fudge (it's very hard to try to do costing from component program costs) and the number includes NASA overhead. It's a bit unfair to SLS as there are MSF assets in NASA that have been there for years that get charged to SLS because that's the only place they fit. On the other hand, there are ground support facilities (VAB, crawlers, launch pads) that are *not* in the SLS budget but are still part of the program cost.
And I guess there's a cost estimate that includes development costs, but that is problematic as that's not the way NASA thinks about things.
My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide. — Preceding unsigned comment added by 24.16.132.243 ( talk) 13:31, 13 July 2020 (UTC)
That will encapsulate most of the cost figures and annual support costs is always hard to figure out, as center personnel are billed in funny ways but it's the best you have.
Then put in Estimated flights per yea — Preceding unsigned comment added by Patbahn ( talk • contribs) 17:19, 16 July 2020 (UTC)
70.70.39.1 ( talk) 23:31, 22 August 2020 (UTC) Ok, you know what?
I'm just going to put 1-2 billion 'depending on how it's calculated' and leave it at that. We have no idea what the actual cost is.
Fredinno ( talk) 23:57, 22 August 2020 (UTC) Oh yeah, BTW, I'm 70.70.39.1 , I just forgot to log in.
People have been editing the cost per launch seemingly constantly depending on people's opinions. It's starting to get dumb.
Fredinno ( talk) 22:36, 23 August 2020 (UTC) But there IS no consensus. The reality is that we have no idea what it actually costs. The other dude didn't agree, he GAVE UP. Not the same thing.
In fact, the conclusion right above was: "My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide." I was trying to fit that. "Over 2 billion excluding development costs" is NOT accurate, even by the sources listed in the original version (neither MENTIONS development costs, you can't just assume that not mentioning development costs means not including development costs). The current version also presents no nuance.
There is no way to make everyone happy other than to list every single quoted price from the federal government (ie. 500 mil to 2 billion, with the generally quoted median from NASA being 1 billion).
I get that you're highly invested in this. I'm not trying to argue for or against, I'm just pointing out that $2 Billion with no if, and, or buts, is not accurate. Not everybody is going to read the talk page. Or hover over for a note. People are just going to take the $2 Billion and assume that's fact and that there's not a significant number of people who disagreee.
The reality is that we have no idea what it actually costs.We don't have "no idea", we have a very large number of estimates made by sources spanning the entire spectrum from reliable to unreliable for this particular figure.
In fact, the conclusion right above was: "My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide."I believe you may be misreading this page, or misunderstanding what consensus means (or both). That isn't a conclusion, it's just what one IP said. Please see the actual RfC here, which is closed. Please note that we cannot explain nuance in the infobox, that belongs in the article itself.
Not everybody is going to read the talk page. Or hover over for a note.Yes. There is discussion of that here, and admittedly I have not made time recently to cover the cost figures in the article itself. I hope I can get to it soon, or someone else can.
There is no way to make everyone happy other than to list every single quoted price from the federal governmentMaking everyone happy is not the goal here. The goal is to present the most accurate figure based solely on the reliable sources that we have. See WP:VNT. Leijurv ( talk) 03:14, 24 August 2020 (UTC)
@
Moamem:
Yeah, I don't believe in the 500mil cost either. You may notice the original edit was 1-2 Bil. I just added 500 million because some people were saying it.
This is in the SLS Wikipedia article: "In May 2019, NASA's Office of Audits reported that the SLS Block 1's marginal cost per launch is to be at least US$876 million."
Also, I've been seeing the $876 mil number quoted as the marginal cost in the Talk page as well. I don't see anywhere where $1Bil is the marginal cost.
I've seen the RfC. I agree with the non-use of the marginal launch cost, but the $1 Billion is not mentioned anywhere as the marginal cost, only as 'the cost'.
My issue with the 2 billion is the 'excluding development costs' part, which isn't stated in that source. 'Once development is complete' doesn't necessarily mean 'excluding development costs'. 'Once development is complete' could mean 'once the vehicle is fully operational'. Ie. After the 2 test flights- costs change after the development launches as the vehicle design and production lines are optimized- this happened with the Shuttle as well (eg. There's no paint on the Shuttle ET even though it was on STS-1.)
If so, both the 1 billion and 2 billion numbers are valid- with no indication of if it includes or excludes development costs, since it's not explicitly stated.
Ok, thanks. Not used to Wikipedia.
Fredinno ( talk) 02:32, 24 August 2020 (UTC)
OK, so... can I change the cost per rocket to 1-2 Billion without mentioning development costs? As the development costs part is not fully clear, and 1 Billion is not marginal cost? Fredinno ( talk) 22:03, 11 September 2020 (UTC)
'Once development is complete' doesn't necessarily mean 'excluding development costs'I disagree. Our role is to summarize information presented in reliable sources. If the cost per launch is $2B once development is completed, I do read that to mean once development is completed, and not this alternate interpretation of "once the vehicle has flown once or twice". Leijurv ( talk) 01:40, 12 September 2020 (UTC)
References
Well, it IS kind of a continuation off the previous discussion of SLS launch costs.... I was replying to someone else on this section of the talk page, and they didn't respond.
You're not understanding what that point I'm trying to make.
There is nothing in the quote, in context, or not, referring to development costs, or indeed even how the OIG arrived at the cost number.
That is my primary concern, I don't understand where and how "excluding development costs" could have been gotten from. It gives the impression the $2 Billion estimate is the minimum potential cost, bloating the cost estimate for the SLS. If it just said "2 Billion", or "Up to 2 Billion," I would be fine with the sidebar, as it gives a more accurate representation of the varied cost estimates for the SLS.
That is my main concern.
It's a cost estimate also on the high range for quoted SLS costs from the government as well. For example, from NASA Director Jim Bridenstine, later in 2019: > NASA Administrator Jim Bridenstine on Monday revealed his latest guess at how much SLS will cost each time it launches: $800 million per rocket for a bulk order and $1.6 billion if NASA purchases just one, Bridenstine, told CNN Business' Rachel Crane. That's just an estimate, Bridenstine noted, because the space agency "needs to sit down with its primary contractor, Boeing (BA), and negotiate the best solution to getting the right mix of the number of rockets and the cost per rocket." Boeing declined to comment on those negotiations and referred questions to NASA. From: https://edition.cnn.com/2019/12/09/tech/nasa-sls-price-cost-artemis-moon-rocket-scn/index.html
Hence, I also suggested having a cost estimate that also includes the lower cost estimates of $800 (low end cost estimate range) to 1.6-2 Billion (high-end cost estimate range)- or even "1-2 Billion" or "Up to 2 Billion" as less misleading and more fair options.
I understand that people here are probably sick of it, and that's part of the reason there's so much aversion to changing the sidebar. Fredinno ( talk) 23:41, 31 March 2021 (UTC)
I understand that people here are probably sick of it, and that's part of the reason there's so much aversion to changing the sidebar.Yes I am a little sick of it by now.
You're not understanding what that point I'm trying to make. There is nothing in the quote, in context, or not, referring to development costs, or indeed even how the OIG arrived at the cost number.I have no idea what you're talking about. The source says
At an estimated cost of over $2 billion per launch for the SLS once development is complete. The article says
Over $2 billion excluding development. Can you just spell this out for me because from my point of view this is just a tiny nitpick of words?? Over two billion US dollars per launch, after development. Why do you want it to say "Up to 2" or "2"? The source says "Over 2". So does the other ars technica source, which reached out to a NASA spokesperson, who did not deny that figure.
Hence, I also suggested having a cost estimate that also includes the ...Again I refer you to WP:VNT and WP:FALSEBALANCE. We do not need to give the subject the last word. Take a look at WP:MRDA which is a similar idea. It isn't a perfect analogy, but my thinking is along the lines of " WP:INDEPENDENT exists for a reason". If the head of NASA says that the program will be cheaper than independent sources indicate, I'm not so sure we should trust that. E.g. see #"first_two_flights" or #Coverage_of_Cost_in_the_Article_Itself. There's a lot of cost sketchiness here, see Space Launch System#Criticism (the sidebar image relating to the booster shenannigans), or Exploration Ground Systems which has a separately budgeted $600m/yr, but only supports SLS. If SLS launches once a year, and its launch pad costs $600m/yr to maintain, should we add that $600m to cost of launch of SLS? (well that is a rhetorical question - obviously "we", meaning "wikipedia editors" should not take it upon ourselves to do that arithmetic, but I'm making the point that all-inclusive estimates of the cost of launch of SLS are not as simple as just "how much do we pay Boeing to get one of them"). As I said earlier when you brought this up we simply do not need to list all the possible cost estimates. The infobox is not the place for nuance. Since the last time we talked about this six months ago, I have gone ahead and actually written into the article, see Space_Launch_System#Per_launch_costs. That is where we can go into depth as to who said what when about how much SLS will cost. The infobox should indeed remain with the best overall source according to WP:EDITDISC and common sense. Leijurv ( talk) 02:35, 1 April 2021 (UTC)
I was pointing out that the OIG is one source, and it's way out of line with other cost estimates for SLS.
Similar to how the payload for the Saturn V used to be listed as something like 140 mT to LEO- which was a vast overestimate out of line with others due to being the total mass to orbit, which is different than the total payload to orbit- and was also from 1 source. That lasted for years before it was changed.
That's my point. It's a data point that's an outlier.
Not whether the OIG says it is "over" or "under" 2 billion.
If anything, being "over" 2 Billion makes it even more of an outlier. If the OIG is getting the cost/launch by dividing the cost of the SLS program by the number of launches per year (which is the only way I can imagine you can get to over 2 Billion/launch), that's a horrifically inaccurate estimation that would bloat the cost of all similar programs.
For example, by that kind of estimation, the cost of a seat on a Dragon 2 paid to the ISS would be more like $80 Million per seat. (a 45% cost overestimation. (!!!))
We can both agree that's not true. And it's not.
Also, we use the price per launch for F9 Reusable as basically what Elon said to SpaceNews, so yes, I do think what the director of NASA says as the cost to NASA should be considered a reliable source.
...
But- even if we can disagree on the above- Can we at least remove the "excluding development costs" part? That was my original and biggest concern.
There is no part of the OIG report that even refers to development costs specifically.
I would be satisfied if we can at least resolve that aspect. Fredinno ( talk) 20:14, 1 April 2021 (UTC)
and it's way out of line with other cost estimates for SLS. I'm not so sure. This has been discussed many times on this page, but I'll get into this in the next two paragraphs.
If the OIG is getting the cost/launch by dividing the cost of the SLS program by the number of launches per yearI don't think we can conclude that they did that. I realize this may appear to be getting a little bit closer to WP:OR than WP:EDITDISC, but actually we can look in the same cited source of Ars Technica here to get clarification on this. See this passage:
The White House number appears to include both the "marginal" cost of building a single SLS rocket as well as the "fixed" costs of maintaining a standing army of thousands of employees and hundreds of suppliers across the country. Building a second SLS rocket each year would make the per-unit cost "significantly less," Hambleton said. What the White House cost estimate did not include, however, was development costs. Since 2011, Congress has appropriated approximately $2 billion per year for the "development" of the SLS rocket (this does not include hundreds of millions of dollars spent annually on ground systems "development" for the rocket at Kennedy Space Center). If these costs are amortized over 10 launches of the SLS vehicle during the 2020s, the per-flight cost would be approximately $4 billion per flight.So as you can see, if we take the cost and divide by the number of launches, we get a number twice as large, according to the cited source. If you look down on this page you can see there has been a lot of arguing about these specific words. Imagine SLS launches once a year, doesn't that mean the cost of one SLS launch includes the cost of paying one year of salaries to everyone who constructs SLS? Why would it not include that? If we want one more SLS, we do indeed need to pay one more year of those salaries. The same goes for Exploration Ground Systems (although the $2B number does not include EGS, I'm just using it to make a point), if the SLS launch pad (which has no other role but to launch SLS) costs $X per year to maintain, doesn't that count as a cost of launch? What about the other EGS operations, which are to
assemble, integrate, prepare and launch the SLS and its payloads? NASA budgets all these things separately. Again, take a look at the image in the Criticism section of the main article relating to how NASA has very inventively moved this budget around.
assemble, integrate, prepare and launch the SLS, that's EGS (which, again, has a separate budget).
a concept study report for a space telescope was advised by NASA HQ in 2019 to budget US$500 million for an SLS launch in 2035, that absolutely does not mean we can pull out that $500M number as the cost of launch of SLS. Believe it or not, that source was cited as the lower range of what the cost of launch figure used to say (it used to be $500M to $2B, citing that for the lower range). I hope we can both agree that that is absurd. It's been discussed many times on this page how much of an impossibility that is.
We can both agree that's not true. And it's not.Well hold on there. Taking the total cost of a program divided by how many times it launches isn't, like, "lying". It isn't "untruthful". It's just a different metric. Average total cost versus marginal cost are just two different economic metrics you can calculate. But again as I said earlier, this $2B figure is not an average total cost, as it is stated to be after development in the original source, and the Ars Technica article explains further exactly what is and isn't included. And I want to make sure to point out that that Ars Technica article is not just Eric Berger's personal interpretation of the OMB number, it actually includes quotes that he got from reaching out to a NASA spokesperson on the topic, so it actually is new information and not just a restatement of the OMB letter.
Also, we use the price per launch for F9 Reusable as basically what Elon said to SpaceNews, so yes, I do think what the director of NASA says as the cost to NASA should be considered a reliable source.SpaceX is a private company. The US Government is a higher authority over NASA that provides its funding. It isn't really comparable. Additionally, SpaceX has no real reason to distort budgeting. (referring only to internal costs, not to external statements, those have an incentive to be low). SpaceX has customers, NASA does not. NASA doesn't even need to figure out how much SLS costs because it's going to happen regardless, SpaceX needs to present to its customers a number which is how much they need to pay for a launch, then SpaceX has to honor that contractually. NASA has nothing like that for SLS (yet?). But anyway, WP:OTHERSTUFFEXISTS and such. We have a better source which is a specific estimate for the marginal cost of a SLS launch, which is clearly superior to an offhand / informal statement from an administrator on how much he imagines he can negotiate the cost down to in the future, followed up by an "I don't know, honestly" and an explanation of his obligation to get the cost as low as possible.
There is no part of the OIG report that even refers to development costs specifically.What? I'm going to paste in exactly what the source says, yet again:
At an estimated cost of over $2 billion per launch for the SLS once development is complete. Estimated cost ... once development is complete. It does talk about cost, and it does say that the number is after development. Looking back at your explanation you wrote 7 months ago, you wrote
My issue with the 2 billion is the 'excluding development costs' part, which isn't stated in that source. 'Once development is complete' doesn't necessarily mean 'excluding development costs'. 'Once development is complete' could mean 'once the vehicle is fully operational'. Ie. After the 2 test flights- costs change after the development launches as the vehicle design and production lines are optimized. I think this is a little silly. You're assuming that "development" means "the first 2 flights". Why? I thought the first flight was uncrewed testing and the second flight was crewed operational? You're assuming that development costs will still be included, but only after some development is complete? So you interpret
once development is completeto mean "still including development, but only development that occurs chronogically after the first two flights". Why? See my response from 6 months ago:
If the cost per launch is $2B once development is completed, I do read that to mean once development is completed, and not this alternate interpretation of "once the vehicle has flown once or twice".
:, adding one more per level, so
::in response to this) Leijurv ( talk) 20:53, 1 April 2021 (UTC)
"Once development is complete" does not mean "excluding development costs."
I have no idea how you reached that conclusion.
Fredinno ( talk) 21:19, 1 April 2021 (UTC)
Ars Technica is considered generally reliable for science- and technology-related articles.), here, that synthesized and explained what that cost number means, with the article being based upon, and sharing quotes from, comment and explanation given by a NASA spokesperson in response to the author's questions asking for clarification about that OMB primary source document. Quote from the article:
What the White House cost estimate did not include, however, was development costs.Leijurv ( talk) 21:57, 1 April 2021 (UTC)
Except that article also says the cost per launch including development costs would be $4 Billion, which is factually untrue.
The cost for the $20 Billion development cost number includes Orion (as also noted in the SLS Wikipedia article itself), which bloats the total cost of development per launch.
SLS does not require Orion to launch (Clipper was manifested for it at one point), and thus including the costs for Orion in the costs for 1 SLS launch is at best misleading. At worst, a complete lie.
So, no, I think this might be an exception to Ars being reliable. Especially reliable enough to state that without any reference to any original or secondary source- considering it can be proven it contains fake/misleading data to begin with. — Preceding unsigned comment added by Fredinno ( talk • contribs)
Except that article also says the cost per launch including development costs would be $4 Billion, which is factually untrue. The cost for the $20 Billion development cost number includes Orion (as also noted in the SLS Wikipedia article itself), which bloats the total cost of development per launch. SLS does not require Orion to launch (Clipper was manifested for it at one point), and thus including the costs for Orion in the costs for 1 SLS launch is at best misleading. At worst, a complete lie.What are you talking about? I'm finding it harder and harder to WP:AGF on this, as it seems like you're not really reading the article / the sources / what I've written. But I'll go for it yet again. In this case, you have completely misread the article. It actually says the literal opposite of what you claim. Please go to the Funding history section of the main article, and read the text underneath the table. The table says a total cost of 20.314 billion dollars, and the text underneath says:
On top of this, the costs to assemble, integrate, prepare and launch the SLS and its payloads are funded separately under Exploration Ground Systems,[97] currently about US$600 million[98] per year. Excluded from the above SLS costs are: •Costs of payloads for the SLS (such as Orion crew capsule)(emphasis added). You can also look at the funding section of the Orion article to see that the funding amounts per year are completely different. If we add them together we get 20,314+21,477 = forty one billion dollars, which is not twenty. So your tirade about how "$20 billion" supposedly includes Orion, therefore meaning that Ars Technica is an unreliable source, is completely unfounded. You yourself wrote
as also noted in the SLS Wikipedia article itselfwhile it says the exact opposite so I'm really not sure what to make of this. I have now written >2500 words of explaining this just to you, no one else, on assumption that we are both acting in good faith, so do you see why this is frustrating?
Especially reliable enough to state that without any reference to any original or secondary sourceSecondary sources can cite primary sources, but don't have to. That is what it means to be a reliable secondary source: if Eric Berger learns information from a direct interview or other communication with a NASA spokesperson, we just have to trust that he has correctly quoted, summarized, contextualized, etc that interaction he had with NASA. This is something that cannot be cited or "proven", if you suspect he is making
complete lies and fabrications. We trust in "the institution" of Ars Technica - so if Eric Berger tweets something, we don't care, but if it's a published Ars Technica article, it has gone through editors, if it were proven wrong they would publish a correction, if he consistently wrote incorrect things he would be fired, etc. This is what journalistic reliability is. See WP:RSEDITORIAL, which Ars Technica meets, according to current consensus on WP:RSPSRC. Leijurv ( talk) 02:17, 2 April 2021 (UTC)
The unit conversions are wrong in the lift capacity of the Block 2 version. This is the only place where Tonnes and Long Tones have the same value. And, it a large value, so it can't be a rounding issue if the others are to be believd. "130 tonnes (130 long tons; 140 short tons)" I would correct, but I'm not sure which value is authoritative and which are the conversions. — Preceding unsigned comment added by 69.47.45.106 ( talk) 17:53, 31 August 2020 (UTC)
This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 |
@ Jadebenn:Since the above conversation has become unreadable and we seem to have come to an agreement concerning the Program Cost so far, I think starting a new section focusing only on Launch Cost is a good idea. For now I am going to delete the Lunch Cost altogether until we come to an agreement (ie you come to your senses). It's all estimates in the best of cases anyway. So you admit that the NASA administrator speech is "a weak citation by itself" and add that it should somehow be lumped with the other (weak) references to make your case. multiple anecdotal evidence do not amount to a good argument. This is a fallacy. So I'm gonna take your references one by one!
@ Jadebenn: I just realised how much you've been acting like you own this page, I am at least the 6th person in the last 6 months that tries to revert this figure to the real one :
1) https://en.wikipedia.org/?title=Space_Launch_System&diff=925620594&oldid=925395840
2) https://en.wikipedia.org/?title=Space_Launch_System&diff=925963019&oldid=925876896
3) https://en.wikipedia.org/?title=Space_Launch_System&diff=926359702&oldid=926277355
4) https://en.wikipedia.org/?title=Space_Launch_System&diff=927943530&oldid=927446682
5) https://en.wikipedia.org/?title=Space_Launch_System&diff=955609445&oldid=955286614
and me.
Which side is the consensus on? Moamem ( talk) 13:37, 29 May 2020 (UTC)
That "marginal cost" figure is the only launch cost we have. Again, SLS development will be complete the second SLS leaves the pad, just as Shuttle development was complete the second it left the pad. There are still many unrelated costs to launch. BOLE will be part of that $2B/year until the 8th launch for example. The RS-25E and EUS until its fourth. Things like eCryo also recieve funding from the SLS program. The cost to develop those technologies is totally separate from how much money it takes NASA to buy an SLS rocket and launch it. It's a similar situation with that "$100M/year" RS-25 cost. The unit price of an engine will be lower than that, exactly because that figure includes things like personnel costs and R&D. As Gopher65 said, there's a difference between program cost per launch and the cost of something itself. If SLS cadence increases to twice per year, for instance, the launch cost will remain roughly the same, but the program cost per launch will decrease significantly. Thus, using it as the "cost per launch" metric is misleading at best.
However, we seem to have once again reached an impasse. To break this stalemate, I suggest we solicit a third opinion from a neutral observer. Would you agree to abide by their decision if I will? – Jadebenn ( talk · contribs · subpages) 18:59, 31 May 2020 (UTC)
How's this for a compromise, set the cost per launch in the infobox to $500 million - $2 billion with a link to a note that explains total program cost vs. program operating cost per flight vs. marginal cost of adding a flight to the manifest? Eggsaladsandwich ( talk) 20:53, 31 May 2020 (UTC)
Moamem Are you OK with removing the "(2019 estimate)" from the cost per launch item in the infobox? The timing of the estimate is reflected in the reference. Eggsaladsandwich ( talk) 01:35, 3 June 2020 (UTC)
I, too, have come over here as a result of an invitation on my Talk page. I have read through the above comments in this section, from 28 Mar to early June. Do you feel you all reached consensus and the article now reflects that consensus? If not, I'd be happy to provide some input to what seems to be a complicated discussion. Pinging Moamem, Eggsaladsandwich, Jadebenn. Cheers. N2e ( talk) 16:35, 25 June 2020 (UTC)
@
N2e: I believe a consensus was reached. –
Jadebenn (
talk ·
contribs ·
subpages)
19:26, 25 June 2020 (UTC)
@ N2e::Hey, as I said the $500 million figure is very concerning to me, it was an aspirational figure that is and never will be achievable since the engines alone cost $400 million excluding development and production restart costs. I just got tired of bickering alone with jadebenn, if that's considered consensus, I don't know... Would love to hear your opinion tho. Moamem ( talk) 02:34, 27 June 2020 (UTC)
@ N2e:I think we've reached a workable compromise solution. Eggsaladsandwich ( talk) 20:56, 1 July 2020 (UTC)
N2e Regarding your comment on the Habex Final Report. The cost analysis is actually documented in Chapter 9 of that report. Appendix G contains the results of an additional independent cost estimate study. The PDF states that appendix is withheld "due to U.S. Export Regulations". It is unlikely that appendix will ever be publicly available. FWIW, the version in the article is identical to the document available here: https://www.jpl.nasa.gov/habex/pdf/HabEx-Final-Report-Public-Release.pdf — Preceding unsigned comment added by Eggsaladsandwich ( talk • contribs) 21:44, 5 July 2020 (UTC)
References
The follow-on contract to produce 18 engines is valued at $1.79 billion. This includes labor to build and test the engines, produce tooling and support SLS flights powered by the engines
So, according to the space agency, NASA has spent $3.5 billion for a total of 24 rocket engines. That comes to $146 million per engine.
WRT Launch cost...
The problem you are running into is that there are three different ways of calculating launch cost. There is the commercial concept - how much it would cost to launch a specific payload if you contracted with a commercial provider? NASA has no analog in their costing structure, so if you are trying to find something like that, I think you are destined to run into issues. Note that trying to compare NASA programs to commercial costs is a bit problematic as a NASA program on a commercial provider has many costs outside of launch costs because of NASA overhead.
The "per launch" price that NASA quotes is what is most useful for their budgetary planning; the incremental price of launching one more or one fewer times during a year. At the end of the shuttle, that number was $450 million, but the only number of this type for SLS was an aspirational number early in the program, back when they though things would be cheaper and were planning on launching twice a year. NASA has chosen not to provide an updated estimate of this type; it can be argued that that is a good thing as SLS is a bit weird in that the earlier launches that reused hardware leftover from shuttle are likely to be much cheaper than later flights. There's also the issue that NASA hasn't known main engine costs until recently.
Which leads us to what I call "yearly program-based cost", which is simply looking at the per-year program cost and dividing it by the launch rate. This is my preferred way of looking at program cost as the number is easy to get and hard to fudge (it's very hard to try to do costing from component program costs) and the number includes NASA overhead. It's a bit unfair to SLS as there are MSF assets in NASA that have been there for years that get charged to SLS because that's the only place they fit. On the other hand, there are ground support facilities (VAB, crawlers, launch pads) that are *not* in the SLS budget but are still part of the program cost.
And I guess there's a cost estimate that includes development costs, but that is problematic as that's not the way NASA thinks about things.
My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide. — Preceding unsigned comment added by 24.16.132.243 ( talk) 13:31, 13 July 2020 (UTC)
That will encapsulate most of the cost figures and annual support costs is always hard to figure out, as center personnel are billed in funny ways but it's the best you have.
Then put in Estimated flights per yea — Preceding unsigned comment added by Patbahn ( talk • contribs) 17:19, 16 July 2020 (UTC)
70.70.39.1 ( talk) 23:31, 22 August 2020 (UTC) Ok, you know what?
I'm just going to put 1-2 billion 'depending on how it's calculated' and leave it at that. We have no idea what the actual cost is.
Fredinno ( talk) 23:57, 22 August 2020 (UTC) Oh yeah, BTW, I'm 70.70.39.1 , I just forgot to log in.
People have been editing the cost per launch seemingly constantly depending on people's opinions. It's starting to get dumb.
Fredinno ( talk) 22:36, 23 August 2020 (UTC) But there IS no consensus. The reality is that we have no idea what it actually costs. The other dude didn't agree, he GAVE UP. Not the same thing.
In fact, the conclusion right above was: "My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide." I was trying to fit that. "Over 2 billion excluding development costs" is NOT accurate, even by the sources listed in the original version (neither MENTIONS development costs, you can't just assume that not mentioning development costs means not including development costs). The current version also presents no nuance.
There is no way to make everyone happy other than to list every single quoted price from the federal government (ie. 500 mil to 2 billion, with the generally quoted median from NASA being 1 billion).
I get that you're highly invested in this. I'm not trying to argue for or against, I'm just pointing out that $2 Billion with no if, and, or buts, is not accurate. Not everybody is going to read the talk page. Or hover over for a note. People are just going to take the $2 Billion and assume that's fact and that there's not a significant number of people who disagreee.
The reality is that we have no idea what it actually costs.We don't have "no idea", we have a very large number of estimates made by sources spanning the entire spectrum from reliable to unreliable for this particular figure.
In fact, the conclusion right above was: "My best advice WRT cost is that the article should deal with the nuance and present the different costing models and let the reader decide."I believe you may be misreading this page, or misunderstanding what consensus means (or both). That isn't a conclusion, it's just what one IP said. Please see the actual RfC here, which is closed. Please note that we cannot explain nuance in the infobox, that belongs in the article itself.
Not everybody is going to read the talk page. Or hover over for a note.Yes. There is discussion of that here, and admittedly I have not made time recently to cover the cost figures in the article itself. I hope I can get to it soon, or someone else can.
There is no way to make everyone happy other than to list every single quoted price from the federal governmentMaking everyone happy is not the goal here. The goal is to present the most accurate figure based solely on the reliable sources that we have. See WP:VNT. Leijurv ( talk) 03:14, 24 August 2020 (UTC)
@
Moamem:
Yeah, I don't believe in the 500mil cost either. You may notice the original edit was 1-2 Bil. I just added 500 million because some people were saying it.
This is in the SLS Wikipedia article: "In May 2019, NASA's Office of Audits reported that the SLS Block 1's marginal cost per launch is to be at least US$876 million."
Also, I've been seeing the $876 mil number quoted as the marginal cost in the Talk page as well. I don't see anywhere where $1Bil is the marginal cost.
I've seen the RfC. I agree with the non-use of the marginal launch cost, but the $1 Billion is not mentioned anywhere as the marginal cost, only as 'the cost'.
My issue with the 2 billion is the 'excluding development costs' part, which isn't stated in that source. 'Once development is complete' doesn't necessarily mean 'excluding development costs'. 'Once development is complete' could mean 'once the vehicle is fully operational'. Ie. After the 2 test flights- costs change after the development launches as the vehicle design and production lines are optimized- this happened with the Shuttle as well (eg. There's no paint on the Shuttle ET even though it was on STS-1.)
If so, both the 1 billion and 2 billion numbers are valid- with no indication of if it includes or excludes development costs, since it's not explicitly stated.
Ok, thanks. Not used to Wikipedia.
Fredinno ( talk) 02:32, 24 August 2020 (UTC)
OK, so... can I change the cost per rocket to 1-2 Billion without mentioning development costs? As the development costs part is not fully clear, and 1 Billion is not marginal cost? Fredinno ( talk) 22:03, 11 September 2020 (UTC)
'Once development is complete' doesn't necessarily mean 'excluding development costs'I disagree. Our role is to summarize information presented in reliable sources. If the cost per launch is $2B once development is completed, I do read that to mean once development is completed, and not this alternate interpretation of "once the vehicle has flown once or twice". Leijurv ( talk) 01:40, 12 September 2020 (UTC)
References
Well, it IS kind of a continuation off the previous discussion of SLS launch costs.... I was replying to someone else on this section of the talk page, and they didn't respond.
You're not understanding what that point I'm trying to make.
There is nothing in the quote, in context, or not, referring to development costs, or indeed even how the OIG arrived at the cost number.
That is my primary concern, I don't understand where and how "excluding development costs" could have been gotten from. It gives the impression the $2 Billion estimate is the minimum potential cost, bloating the cost estimate for the SLS. If it just said "2 Billion", or "Up to 2 Billion," I would be fine with the sidebar, as it gives a more accurate representation of the varied cost estimates for the SLS.
That is my main concern.
It's a cost estimate also on the high range for quoted SLS costs from the government as well. For example, from NASA Director Jim Bridenstine, later in 2019: > NASA Administrator Jim Bridenstine on Monday revealed his latest guess at how much SLS will cost each time it launches: $800 million per rocket for a bulk order and $1.6 billion if NASA purchases just one, Bridenstine, told CNN Business' Rachel Crane. That's just an estimate, Bridenstine noted, because the space agency "needs to sit down with its primary contractor, Boeing (BA), and negotiate the best solution to getting the right mix of the number of rockets and the cost per rocket." Boeing declined to comment on those negotiations and referred questions to NASA. From: https://edition.cnn.com/2019/12/09/tech/nasa-sls-price-cost-artemis-moon-rocket-scn/index.html
Hence, I also suggested having a cost estimate that also includes the lower cost estimates of $800 (low end cost estimate range) to 1.6-2 Billion (high-end cost estimate range)- or even "1-2 Billion" or "Up to 2 Billion" as less misleading and more fair options.
I understand that people here are probably sick of it, and that's part of the reason there's so much aversion to changing the sidebar. Fredinno ( talk) 23:41, 31 March 2021 (UTC)
I understand that people here are probably sick of it, and that's part of the reason there's so much aversion to changing the sidebar.Yes I am a little sick of it by now.
You're not understanding what that point I'm trying to make. There is nothing in the quote, in context, or not, referring to development costs, or indeed even how the OIG arrived at the cost number.I have no idea what you're talking about. The source says
At an estimated cost of over $2 billion per launch for the SLS once development is complete. The article says
Over $2 billion excluding development. Can you just spell this out for me because from my point of view this is just a tiny nitpick of words?? Over two billion US dollars per launch, after development. Why do you want it to say "Up to 2" or "2"? The source says "Over 2". So does the other ars technica source, which reached out to a NASA spokesperson, who did not deny that figure.
Hence, I also suggested having a cost estimate that also includes the ...Again I refer you to WP:VNT and WP:FALSEBALANCE. We do not need to give the subject the last word. Take a look at WP:MRDA which is a similar idea. It isn't a perfect analogy, but my thinking is along the lines of " WP:INDEPENDENT exists for a reason". If the head of NASA says that the program will be cheaper than independent sources indicate, I'm not so sure we should trust that. E.g. see #"first_two_flights" or #Coverage_of_Cost_in_the_Article_Itself. There's a lot of cost sketchiness here, see Space Launch System#Criticism (the sidebar image relating to the booster shenannigans), or Exploration Ground Systems which has a separately budgeted $600m/yr, but only supports SLS. If SLS launches once a year, and its launch pad costs $600m/yr to maintain, should we add that $600m to cost of launch of SLS? (well that is a rhetorical question - obviously "we", meaning "wikipedia editors" should not take it upon ourselves to do that arithmetic, but I'm making the point that all-inclusive estimates of the cost of launch of SLS are not as simple as just "how much do we pay Boeing to get one of them"). As I said earlier when you brought this up we simply do not need to list all the possible cost estimates. The infobox is not the place for nuance. Since the last time we talked about this six months ago, I have gone ahead and actually written into the article, see Space_Launch_System#Per_launch_costs. That is where we can go into depth as to who said what when about how much SLS will cost. The infobox should indeed remain with the best overall source according to WP:EDITDISC and common sense. Leijurv ( talk) 02:35, 1 April 2021 (UTC)
I was pointing out that the OIG is one source, and it's way out of line with other cost estimates for SLS.
Similar to how the payload for the Saturn V used to be listed as something like 140 mT to LEO- which was a vast overestimate out of line with others due to being the total mass to orbit, which is different than the total payload to orbit- and was also from 1 source. That lasted for years before it was changed.
That's my point. It's a data point that's an outlier.
Not whether the OIG says it is "over" or "under" 2 billion.
If anything, being "over" 2 Billion makes it even more of an outlier. If the OIG is getting the cost/launch by dividing the cost of the SLS program by the number of launches per year (which is the only way I can imagine you can get to over 2 Billion/launch), that's a horrifically inaccurate estimation that would bloat the cost of all similar programs.
For example, by that kind of estimation, the cost of a seat on a Dragon 2 paid to the ISS would be more like $80 Million per seat. (a 45% cost overestimation. (!!!))
We can both agree that's not true. And it's not.
Also, we use the price per launch for F9 Reusable as basically what Elon said to SpaceNews, so yes, I do think what the director of NASA says as the cost to NASA should be considered a reliable source.
...
But- even if we can disagree on the above- Can we at least remove the "excluding development costs" part? That was my original and biggest concern.
There is no part of the OIG report that even refers to development costs specifically.
I would be satisfied if we can at least resolve that aspect. Fredinno ( talk) 20:14, 1 April 2021 (UTC)
and it's way out of line with other cost estimates for SLS. I'm not so sure. This has been discussed many times on this page, but I'll get into this in the next two paragraphs.
If the OIG is getting the cost/launch by dividing the cost of the SLS program by the number of launches per yearI don't think we can conclude that they did that. I realize this may appear to be getting a little bit closer to WP:OR than WP:EDITDISC, but actually we can look in the same cited source of Ars Technica here to get clarification on this. See this passage:
The White House number appears to include both the "marginal" cost of building a single SLS rocket as well as the "fixed" costs of maintaining a standing army of thousands of employees and hundreds of suppliers across the country. Building a second SLS rocket each year would make the per-unit cost "significantly less," Hambleton said. What the White House cost estimate did not include, however, was development costs. Since 2011, Congress has appropriated approximately $2 billion per year for the "development" of the SLS rocket (this does not include hundreds of millions of dollars spent annually on ground systems "development" for the rocket at Kennedy Space Center). If these costs are amortized over 10 launches of the SLS vehicle during the 2020s, the per-flight cost would be approximately $4 billion per flight.So as you can see, if we take the cost and divide by the number of launches, we get a number twice as large, according to the cited source. If you look down on this page you can see there has been a lot of arguing about these specific words. Imagine SLS launches once a year, doesn't that mean the cost of one SLS launch includes the cost of paying one year of salaries to everyone who constructs SLS? Why would it not include that? If we want one more SLS, we do indeed need to pay one more year of those salaries. The same goes for Exploration Ground Systems (although the $2B number does not include EGS, I'm just using it to make a point), if the SLS launch pad (which has no other role but to launch SLS) costs $X per year to maintain, doesn't that count as a cost of launch? What about the other EGS operations, which are to
assemble, integrate, prepare and launch the SLS and its payloads? NASA budgets all these things separately. Again, take a look at the image in the Criticism section of the main article relating to how NASA has very inventively moved this budget around.
assemble, integrate, prepare and launch the SLS, that's EGS (which, again, has a separate budget).
a concept study report for a space telescope was advised by NASA HQ in 2019 to budget US$500 million for an SLS launch in 2035, that absolutely does not mean we can pull out that $500M number as the cost of launch of SLS. Believe it or not, that source was cited as the lower range of what the cost of launch figure used to say (it used to be $500M to $2B, citing that for the lower range). I hope we can both agree that that is absurd. It's been discussed many times on this page how much of an impossibility that is.
We can both agree that's not true. And it's not.Well hold on there. Taking the total cost of a program divided by how many times it launches isn't, like, "lying". It isn't "untruthful". It's just a different metric. Average total cost versus marginal cost are just two different economic metrics you can calculate. But again as I said earlier, this $2B figure is not an average total cost, as it is stated to be after development in the original source, and the Ars Technica article explains further exactly what is and isn't included. And I want to make sure to point out that that Ars Technica article is not just Eric Berger's personal interpretation of the OMB number, it actually includes quotes that he got from reaching out to a NASA spokesperson on the topic, so it actually is new information and not just a restatement of the OMB letter.
Also, we use the price per launch for F9 Reusable as basically what Elon said to SpaceNews, so yes, I do think what the director of NASA says as the cost to NASA should be considered a reliable source.SpaceX is a private company. The US Government is a higher authority over NASA that provides its funding. It isn't really comparable. Additionally, SpaceX has no real reason to distort budgeting. (referring only to internal costs, not to external statements, those have an incentive to be low). SpaceX has customers, NASA does not. NASA doesn't even need to figure out how much SLS costs because it's going to happen regardless, SpaceX needs to present to its customers a number which is how much they need to pay for a launch, then SpaceX has to honor that contractually. NASA has nothing like that for SLS (yet?). But anyway, WP:OTHERSTUFFEXISTS and such. We have a better source which is a specific estimate for the marginal cost of a SLS launch, which is clearly superior to an offhand / informal statement from an administrator on how much he imagines he can negotiate the cost down to in the future, followed up by an "I don't know, honestly" and an explanation of his obligation to get the cost as low as possible.
There is no part of the OIG report that even refers to development costs specifically.What? I'm going to paste in exactly what the source says, yet again:
At an estimated cost of over $2 billion per launch for the SLS once development is complete. Estimated cost ... once development is complete. It does talk about cost, and it does say that the number is after development. Looking back at your explanation you wrote 7 months ago, you wrote
My issue with the 2 billion is the 'excluding development costs' part, which isn't stated in that source. 'Once development is complete' doesn't necessarily mean 'excluding development costs'. 'Once development is complete' could mean 'once the vehicle is fully operational'. Ie. After the 2 test flights- costs change after the development launches as the vehicle design and production lines are optimized. I think this is a little silly. You're assuming that "development" means "the first 2 flights". Why? I thought the first flight was uncrewed testing and the second flight was crewed operational? You're assuming that development costs will still be included, but only after some development is complete? So you interpret
once development is completeto mean "still including development, but only development that occurs chronogically after the first two flights". Why? See my response from 6 months ago:
If the cost per launch is $2B once development is completed, I do read that to mean once development is completed, and not this alternate interpretation of "once the vehicle has flown once or twice".
:, adding one more per level, so
::in response to this) Leijurv ( talk) 20:53, 1 April 2021 (UTC)
"Once development is complete" does not mean "excluding development costs."
I have no idea how you reached that conclusion.
Fredinno ( talk) 21:19, 1 April 2021 (UTC)
Ars Technica is considered generally reliable for science- and technology-related articles.), here, that synthesized and explained what that cost number means, with the article being based upon, and sharing quotes from, comment and explanation given by a NASA spokesperson in response to the author's questions asking for clarification about that OMB primary source document. Quote from the article:
What the White House cost estimate did not include, however, was development costs.Leijurv ( talk) 21:57, 1 April 2021 (UTC)
Except that article also says the cost per launch including development costs would be $4 Billion, which is factually untrue.
The cost for the $20 Billion development cost number includes Orion (as also noted in the SLS Wikipedia article itself), which bloats the total cost of development per launch.
SLS does not require Orion to launch (Clipper was manifested for it at one point), and thus including the costs for Orion in the costs for 1 SLS launch is at best misleading. At worst, a complete lie.
So, no, I think this might be an exception to Ars being reliable. Especially reliable enough to state that without any reference to any original or secondary source- considering it can be proven it contains fake/misleading data to begin with. — Preceding unsigned comment added by Fredinno ( talk • contribs)
Except that article also says the cost per launch including development costs would be $4 Billion, which is factually untrue. The cost for the $20 Billion development cost number includes Orion (as also noted in the SLS Wikipedia article itself), which bloats the total cost of development per launch. SLS does not require Orion to launch (Clipper was manifested for it at one point), and thus including the costs for Orion in the costs for 1 SLS launch is at best misleading. At worst, a complete lie.What are you talking about? I'm finding it harder and harder to WP:AGF on this, as it seems like you're not really reading the article / the sources / what I've written. But I'll go for it yet again. In this case, you have completely misread the article. It actually says the literal opposite of what you claim. Please go to the Funding history section of the main article, and read the text underneath the table. The table says a total cost of 20.314 billion dollars, and the text underneath says:
On top of this, the costs to assemble, integrate, prepare and launch the SLS and its payloads are funded separately under Exploration Ground Systems,[97] currently about US$600 million[98] per year. Excluded from the above SLS costs are: •Costs of payloads for the SLS (such as Orion crew capsule)(emphasis added). You can also look at the funding section of the Orion article to see that the funding amounts per year are completely different. If we add them together we get 20,314+21,477 = forty one billion dollars, which is not twenty. So your tirade about how "$20 billion" supposedly includes Orion, therefore meaning that Ars Technica is an unreliable source, is completely unfounded. You yourself wrote
as also noted in the SLS Wikipedia article itselfwhile it says the exact opposite so I'm really not sure what to make of this. I have now written >2500 words of explaining this just to you, no one else, on assumption that we are both acting in good faith, so do you see why this is frustrating?
Especially reliable enough to state that without any reference to any original or secondary sourceSecondary sources can cite primary sources, but don't have to. That is what it means to be a reliable secondary source: if Eric Berger learns information from a direct interview or other communication with a NASA spokesperson, we just have to trust that he has correctly quoted, summarized, contextualized, etc that interaction he had with NASA. This is something that cannot be cited or "proven", if you suspect he is making
complete lies and fabrications. We trust in "the institution" of Ars Technica - so if Eric Berger tweets something, we don't care, but if it's a published Ars Technica article, it has gone through editors, if it were proven wrong they would publish a correction, if he consistently wrote incorrect things he would be fired, etc. This is what journalistic reliability is. See WP:RSEDITORIAL, which Ars Technica meets, according to current consensus on WP:RSPSRC. Leijurv ( talk) 02:17, 2 April 2021 (UTC)
The unit conversions are wrong in the lift capacity of the Block 2 version. This is the only place where Tonnes and Long Tones have the same value. And, it a large value, so it can't be a rounding issue if the others are to be believd. "130 tonnes (130 long tons; 140 short tons)" I would correct, but I'm not sure which value is authoritative and which are the conversions. — Preceding unsigned comment added by 69.47.45.106 ( talk) 17:53, 31 August 2020 (UTC)