Hm? I can't think of any recent projects with good design but bad presentation. The first half is true, but the second half seems mistaken in the context of the web.
In my experience, the opposite is often the case -- The argument: "We can't start over because we don't have enough time left" grows in strength as the project approaches the deadline and is used to justify shoddy work. (You paint yourself into a corner.)
I guess one aspect why "agile" stuff is effective, is that the "last day" (on which it becomes surprisingly easy to make all the decisions that couldn't be made until now) happens multiple times and earlier in the project.
To me this is more of a NASA thing. If a program ends up going over budget too much it gets cancelled. But the need is still there, so new program is often started up with a similar set of requirements. This problem is endemic across government aerospace programs.
It took 24 months of effort to get a particular system to barely run every hour by some contractors. After 3 weeks, I rewrote most of it and still have a month left, but it runs in 5 seconds. I'm having trouble right now convincing my boss it's better to work on the 5 second version even if he spent hundreds of thousands of dollars on the 1 hour version. It's the most annoying sunk cost ever.
I took him as being sarcastic when saying there was always time to do it over. He's ironically stating the bad logic that people use to justify introducing new technical debt.
At least, that's what I assume, since I've heard this sentiment expressed in the much clearer phrasing "If you haven't got the time to do it right, when will you find the time to do it over?".
That's apparently the title of a popular book by a management consultant named John Wooden, but I'm not sure of the provenance of the quote.
Expanding: it should be known when failure is tolerated. Situations where it isn’t (e.g. maiden flight with astronauts) should follow ones where it is.
Being the one learning through failure in an organisation not tolerating it isn’t fun.
Depends what kind of payload you have on your rocket. People or a billion-plus dollars of equipment? Or is it just a test rocket to see if it can land itself?
Requiring no risk of failure for people is a large part of why we don’t launch astronauts much.
Space is risky. Setting the expectation that astronauts can never be lost, as opposed to that they are on an exciting frontier that has its dangers, is out of line with astronauts’ own risk preferences. This has been a failure of expectations setting by NASA.
36:Any run-of-the-mill engineer can design something which is elegant. A good engineer designs systems to be efficient. A great engineer designs them to be effective.
This statement reminded me of the popular quote on teacher, “The mediocre teacher tells. The good teacher explains. The superior teacher demonstrates. The great teacher inspires.”
Elegant engineering work is work that has good properties as evaluated by other engineers and that don't directly impact the properties non-engineers care about (what doesn't mean that aren't impacts, just not direct ones).
I don't think you will find this in any dictionary, but the meaning is incredibly consistent, even on different engineering areas and different languages.
I would say, if engineers can even identify what the desired outcome is, they are already on the path to being excellent (or maybe, if they even attempt to identify what the desire outcome is!)
I thought the quote on teachers are 'those that can't do, teach'. I have definitely had a few teachers that inspired, but a far greater number that were much more appropriately lumped into the "can't do" category.
"39. (alternate formulation) The three keys to keeping a new human space program affordable and on schedule:
1) No new launch vehicles.
2) No new launch vehicles.
3) Whatever you do, don't develop any new launch vehicles."
Notably, SpaceX did not develop the Falcon 9 as part of their human spaceflight program, and it was in fact not rated for carrying humans for many years. They instead developed their launch vehicle as part of a standalone "launch vehicle program" then later committed to building a human spaceflight program on top of their existing rocket.
Starship _is_ being developed as part of a human spaceflight program, and we have yet to see whether this violation of Akins Laws will be justified.
I had the opposite thought. SLS/Artemis is a “new human space program” that includes a new launch vehicle and it is hopelessly unaffordable and off schedule. SpaceX developed one of the most affordable human launch systems ever made, in a reasonable amount of time, by using their pre-existing cargo launch vehicle. Even Boeing will likely have Starliner, which also uses an existing workhorse launcher, flying humans before SLS launches anything.
SLS reuses almost everything (with modifications) from the shuttle program. These are 40 year old designs. Falcon Heavy reused a 10 year old design (Falcon 9).
edit: If we count back to the first successful propulsive landing, the technology was only 5 years old. Falcon Heavy had been planned since way back in 2005.
This is not really the case. Everything about SLS was specified by Congress to be similar enough to Shuttle to require all of the same contractors, but different enough to require everything to be redesigned from scratch. No part of it should be thought of as having flight heritage.
The central tank is kind of like Shuttle’s, to justify building it in the same factory in Louisiana, but it’s a different diameter, so it had to be a clean-sheet design and all of the tooling had to be created from scratch. The solid rocket boosters are similar to Shuttle (the good Senator from Utah, with all his engineering expertise, required SLS to use solid fuel boosters that only one company in Utah can make), but a different number of segments in length, requiring them to be designed from scratch.
Early in the process of SLS (under its original name, Ares V), a group of NASA engineers lobbied for a true Shuttle-derived version, which would have been much cheaper and quicker to create, with the benefit of lots of flight heritage. Of course they got nowhere, because none of that was why SLS was the way it was. Everything about SLS is designed for the sole purpose of funneling the maximum amount of money to the right contractors in the right states for as long as possible. Thus, it is acceptable that it has never flown even after so many years and so many billions—indeed it’s desirable! If the costly design phase goes on for as long as possible, the money spigot will dispense much more than if it proceeded into operations.
But, just in case, SLS will undergo two costly redesigns after coming into service: a whole new upper stage and new boosters. That should keep the gravy train running for a good long while.
In addition to the Shuttle-industrial complex which must be kept running with make-work, there is now a Station-industrial complex which must receive the same treatment, in the form of an utterly useless lunar-orbit station called Gateway. I’m not sure if you can tell but I’m fairly bitter about all this.
Not mentioned in the above link was an amateur group developing VTVL tech around the San Francisco bay area in the 90's. IIRC, it was EPRS. FWIW, they also invented a multi-rotor platform to test their conrol system that evolved into the modern drone.
Valid point. Also, if one wants to stretch the envelope, Harold Graham, flying the Bell Rocket Belt, performed the first rocket powered landing, April 20th, 1961 at Bell Aerospace, upstate New York. This development footage opens with that flight:
https://www.youtube.com/watch?v=gxmxbMdToR4
Not quite, spacex changed quite a few things every launch or at least every version, that's why the first falcon 9 is very different compared to one from today. The one carrying humans is called block 5 of the full thrust version (version 1.2).
SpaceX was neither affordable nor on schedule. If you wanted someone up there cheaply and on time 5 years ago, you should just buy a seat on a Soyuz.
However, #39 doesn't say “don't ever develop new launch vehicles”, rather “don't develop new launch vehicles if staying in budget and on timeline is your priority”.
SpaceX did not develop a new launch vehicle for a human space program. SpaceX developed two new reusable launch vehicles for transfering cargo, whilst not cutting the corners that could not be cut if it were used for a human space program. Then when it came time to put humans on it, it already had a legacy.
Starship, however, is in fact designed for humans. But it is not part of a "human space program", rather it is a multi-purpose vehicle. It is yet to be seen towards which human space programs it will be applied, but even with the Artemis bid many aspects of Starship were clearly designed without the Artemis bid as a specific target.
Starship is intended to be a dual-purpose cargo and human flight vehicle and Falcon 9 + Dragon were iteratively improved through many cargo launches before Dragon 2 for human flight.
What has likely changed is that there is sufficient demand for cargo flights to bootstrap an affordable human flight program on top of it.
"29. (von Tiesenhausen's Law of Program Management) To get an accurate estimate of final program requirements, multiply the initial time estimates by pi, and slide the decimal point on the cost estimates one place to the right."
I wish I'd learned this a long while ago. Going on some of my current projects, it seems that a good portion of my brain still hasn't!
“ 7. At the start of any design effort, the person who most wants to be team leader is least likely to be capable of it.”
That’s interesting. I generally have strong opinions about things I care about and would much rather be the one making decisions so I can own the outcome, ESPECIALLY negative outcomes. Maybe it’s just a control thing... I’d rather fix my own mess
“ To summarize: it is a well-known fact that those people who most want to rule people are, ipso facto, those least suited to do it.
To summarize the summary: anyone who is capable of getting themselves made President should on no account be allowed to do the job.” Douglas Adams
I think it's largely a function of the intrinsic reason that someone wants to lead.
Are they in it for the title or the prestige? Their desire will as some point run in conflict with the larger program goals.
Do they do it because they enjoy bringing a diverse set of viewpoints together to create something that couldn't be created in isolation? That might have a better chance of success.
Unfortunately it's very difficult to determine this externally without a significant investment in time or otherwise.
> Do they do it because they enjoy bringing a diverse set of viewpoints together to create something that couldn't be created in isolation? That might have a better chance of success.
That reminds me of the phrase:
If you want to go fast go alone. If you want to go far go together.
Let's consider this in the context of engineering, rather than politics.
It's not clear that this is a good rule. If you want to get mass from ground to orbit, the rocket equation means that you want the least quantity of infrastructure mass that will work, maximizing the payload fraction. If you're in charge of safety equipment on an orbital station, you want the most effective things you can cram in to your mass budget.
Life support? Whenever possible, eliminate it in favor of automation. But if you need it, you either want minimal (emergency only) or maximal (resilient against disasters) to the extent of your available space/power/mass.
Particularly as “the middle” can be dependent on where the ends are. The middle ground of American politics is different to the middle ground of European or Chinese politics for example.
Your name is familiar. Were you there in the 99-02 time frame? I was an undergrad employee then working on SCAMP and later NBV2/RTSX. I helped design and build the big crate on the deck housing the power, comms, and air supply.
Why is almost every spacecraft design in reality and sci-fi “pointed” like airplanes and sea ships, as if it’s only going to move in one direction on one plane?
I’m guessing it’s because of the need to go through our atmosphere first.
Wouldn’t an omnidirectional saucer-like shape (or spherical for larger ships) be more practical in 3D vacuum?
> Why is almost every spacecraft design in reality and sci-fi “pointed” like airplanes and sea ships, as if it’s only going to move in one direction on one plane?
Because they have directional main engines. That does not strictly necessitate them to be pointy on the opposite end, but it at least requires some design that can distribute the force through the ship's body with the minimal amount of structural mass. E.g. an axial design that only has to withstand compressive forces, not shear forces.
A bigger issue is the lack surface area for cooling. Not even The Expanse gets that one right.
Scifi ships tend have engines with stupendous power ratings (Gigawatts and higher). Even with 99% of efficiency they'd have to shed megawatts of heat. Since there is no convective or conductive cooling in space your steady-state option is radiative cooling which is constrained by surface area, temperature difference and emissivity. You don't have much flexibility on the latter two. So you need to increase surface area. The ISS' radiators take up about 200m² and can reject 70kW. There are non-steady-state options such as sacrificial coolants or plain thermal mass but that would require extra mass, resupplies or additional downtime. In The Expanse the ships just don't have enough area considering they're powered by fusion reactors burning constantly.
What about lasers for emitting waste heat? Obviously there are practical issues with having powerful-enough lasers essentially powered by a heat engine, but are there thermodynamic reasons it couldn't work?
This question is not 'Expanse'-related but inspired by David Brin's 'Sundiver'.
I think you're going to run into problems turning waste heat into usable energy, because it implies there's a cold sink already that you can use to convert the temperature differential into electrical energy to drive your laser. High end laser efficacy ranges from 30% to 50%?[1]
But regardless, you're asking to get low-entropy energy out of high-entropy energy, and that's one of those "you can't win or break-even at thermodynamics in a closed system" kind of thing. Since spaceships come conveniently jacketed in a vacuum-insulating outer layer (i.e. space :) you're still stuck with some variant of radiative heat transfer, and "a big piece of metal being used as a infrared heat radiator because we pushed heat into it using a refrigerator loop) is still the cheapest option.
The last option techie that I've ever heard for rejecting heat already onboard on a spaceship (and this is pure technobabble from proposals about Star Wars ships massive power generation needs) would be some magical way to convert heat into neutrinos, which, since neutrinos can pass though most matter, could be inside the ship and still function as a heat rejection mechanism. [2]
The author suggests creating a fusion reaction well behind the ship, but channeling the useful parts of it in an electromagnetic field tunnel to create thrust while dissipating the heat more effectively. Vaguely like Project Orion, but with a more-or-less continuous external fusion reaction instead of a series of nuclear explosions. I am not a physicist, but they seem to have done their homework.
Whatever you do with internally generated energy, you either get it out or it accumulates, overheating you.
You need 3-10 m^2 of radiators for every kW of power you use. If your ship is powered (I mean internal power, not engine exhaust) by a 1MW reactor, then you need radiators close to the scale of a football field pushing out waste heat, no matter what tech that reactor uses.
Delightfully counterintuitive given the low temperatures of space - would it help said ships to hide on the dark side of the moon and avoid solar radiation?
Where does that 3-10 m^2 per kW estimate come from? The best I could guess based on my shaky grasp of radiation is to equate internally generated power to total radiated power using the Stefan-Boltzmann Law:
Based on that, you need 3-10 m^2 of radiative surface area per kilowatt, but that's assuming your equilibrium temperature is 203 to 275 K (-70 to 2 °C). Assuming I haven't made some basic mistake, couldn't you decide to heat part of your surface to some much higher temperature and radiate most of your internal power out of that part?
This is what our current spacecraft radiators manage. You also have to take into account that radiators also interact with the sun and with the rest of the spacecraft.
Of course, you could have a heat pump pushing heat to a higher temperature surface, some current radiators do that, but it has its limitations requires more energy for the pump the higher temperature difference you want to sustain.
> A bigger issue is the lack surface area for cooling. Not even The Expanse gets that one right.
This depends on the working temperature of the radiator too. In theory if area needed to be optimized a (series of) heat pumps could shed more power from a hotter radiator.
If future designers want to be cute about it they could stick a high temperature radiator at the front of the ship as a headlight and lower temperature radiators in the back for a tail light.
Saturn Run, is a really interesting and fun book - by John Sandford. I highly recommend it, as a successful modern science fiction book. That topic is really well covered, with engineers solving all sorts of problems, including that one.
Banner of the Stars addressed it by cooling peak loads via disposable coolant instead of solely relying on radiative cooling. It became a limiting factor during a large-scale fleet battle.
Sci-fi shows need designs that are "readable" by the viewer in split secons. That means unique, distinguishable silhouettes that also tell you at a glance where everything is moving in a shot. Pointy designs make that work excellently.
As for the real world: there are plenty designs that are not pointy and flying: every satellite, the ISS, the lunar lander... however, all the designs that interact with earth's atmosphere have to respect aerodynamics. Thus, just about every ascent vehicle becomes pointy. Landers are different. The Soyuz return module doesn't look particularly pointy, for instance.
They’re not? Obviously many of the spacecraft we’re used to spend a fair bit of time in the atmosphere, but even real manned spacecraft have come in different shapes and sizes (usually variations on spheres and/or cones).
If you are going fast enough, my understanding is that the interstellar medium is actually not that empty and such a shape might make sense again. At least that is the reason given in the Revelation Space saga.
We don't yet need to go through interstellar medium. Currently vehicles which need to go through atmosphere are pointy. Vehicles which do not are not pointy (all current satellites).
In SciFi many ships have a military purpose and should be designed around the guns and the engines. Everything else used cramed in where there is space.
lists of advice like this from experienced practicioners are always great, but does anyone have any good methods for actually incorporating just pieces of wisdom into their own work? I find myself bookmarking stuff like this but never coming back to it or unsure how to fully utilize them
I think there are two ways you can go about incorporating this.
1. You can look back on projects you’ve done or projects you’re working on and try to see how they fit with your work.
2. Trim the list down and add the remaining items to a list or documented ordered by what you work on. Keep the remaining items in a handy place you can reference when you’re in appropriate steps of your project.
For example, the rules in the article can possibly be grouped as follows: 1. Project timelines, 2. project accountabilities, 3. Engineering requirements and design. I would then refer to these in the corresponding phase of a project.
It would be nice if there was a store attached to each law. Knowing the source of the law would be helpful in knowing how to apply it.
Not sure why you are downvoted, unless we are in math land, you might be right.
Numbers that come from measurements are often not precise enough, measure additional things, massaged until they fit the preconceived etc. So its still the opinion pretending to be engineering.
These lists can be fun because they’re a list of inside jokes, but there’s nothing useful or actionable here. I’m not a fan of this type of self-congratulating stuff about budgets and schedules never being right and launch vehicles being too hard. This type of mentality is why most of the industry is stuck in the 60s. Honestly it sounds analogous to wondering why anyone would want a personal computer 40 years ago.
