Hacker News Reader

Comment by WalterBright | Hacker News Reader

> The drive to build another Silicon Valley may be doomed to fail

This reminds me that there have been many attempts to duplicate Lockheed's "Skunkworks". They've all failed. I suspect they failed because they always tried to create a skunkworks by fixing perceived "shortcomings" of the original, not realizing that those shortcomings were what made it work.

It takes a lot of guts to create a skunkworks, because it defies all conventional notions of how to do things. Lockheed is very commendable for having those guts, but they remain the only one.

It also reminds me of Caltech, which remains the only university to have a pervasive honor system. Nobody else has the guts to even try it.

View on HN

> This reminds me that there have been many attempts to duplicate Lockheed's "Skunkworks". They've all failed. I suspect they failed because they always tried to create a skunkworks by fixing perceived "shortcomings" of the original, not realizing that those shortcomings were what made it work.

> It takes a lot of guts to create a skunkworks, because it defies all conventional notions of how to do things. Lockheed is very commendable for having those guts, but they remain the only one.

Except for Boeing's Phantom Works, NASA EagleWorks, various "advanced concepts" or "preliminary design" groups across other aerospace companies, and various small aerospace firms operating with minimal overhead and extremely talented engineers (Karem Aircraft's head actually specifically cited Kelly Johnson's original rules for a SW-type organization, plus there's Joby Aviation, Makani/Google, Zee.Aero). Sure, none of them have produced anything like an SR-71 yet, but neither has the original Skunkworks since then. The SW legacy of a tightly knit group of elite engineers operating with minimal overhead and freedom to try radical concepts is alive and well in aerospace though.

It's good to hear of others. I've never heard of any, and when I worked in aerospace nobody knew of any examples.

The JSF, sadly, is certainly not an example. (SW produced a lot more than the SR-71, they produced an unparalleled series of unique and spectacular aircraft.)

You've piqued my curiosity. Got any references to these comparing them to the original?

Rice also has a very successful Honor Code, though I'm not sure how it differs from that of Caltech. At Rice, it means that many (most?) exams are not even in-class. Students are trusted not to cheat and are allowed to take exams home. If the professor says that it is a one hour closed-book exam, you are expected to adhere to that. Of course, there are in-class exams as well; it's all up to the discretion of the professor, but my experience was that many teachers embraced the Honor Code.

Violations are punished very harshly. Most violations are reported by other students. And students serve on the Honor Council for deciding how violations of the Honor Code should be handled (and if a violation had in fact occurred).

I'm glad there's another example of this. At Caltech, it was institute policy that exams were not proctored. If they were in class (a rarity) the professor was required to leave the room for the duration.

Another unusual Caltechism is that all classes could be placed out of by taking an exam - and if you passed the exam, you even got credit for it.

Professors were not allowed to take attendance nor have attendance be any part of the grade (except for PE). This had the nice effect of the only people in the lecture hall were people who wanted to be there. Cut ups and disruptive students simply didn't attend.

How did they handle stuff like foreign language courses where there was a speech component? (I remember doing one-one conversations for assessment, don't see how you could escape that).

I had exams like that at my school and I hated them. Because I didn't cheat, but others did, obviously, though not provably.

Great comment.

Can you point to any more info about Caltech's honor system? I haven't heard of it before and you piqued my curiosity. (Couldn't find much on Google.)

I don't have a link, but here's one aspect of it that'll give an idea:

Exams are not proctored. They are usually take home. It is up to the student to honor the rules of the test, such as time constraints and whether it is open or closed book.

This has interesting results. It turns the students and professors into collaborators rather than adversaries, it's simply a much nicer atmosphere.

It's so easy to cheat that there's no status for it. In fact, if you're going to cheat, you'd better be damn sure your fellow students don't find out about it, because they'll ostracize you. The students like the honor system, and they're not going to tolerate anyone undermining it.

Also, students often simply left their dorm rooms open. In my four years there, I never heard of any incidents of students stealing. Not once. It was pretty obvious that the students simply would not tolerate other students stealing. Policy was that the campus security was not allowed in the dorms unless invited by the students.

I hope that gives some idea of how it worked.

https://deans.caltech.edu/documents/55-hch2012.pdf

Just to emphasize how pervasive this is among the community: I have had several locked bikes stolen from Pasadena City College which is just blocks away, while at Caltech I left my bike unlocked every day for six months without a single incident.

Regarding the exams, the professors (and I believe TAs as well) are straight up forbidden from being near the students during any exams and the result is one of the lowest rates of cheating across all surveys.

The vast majority (if not all) of the classes I sat in on very much encouraged collaboration between students (even on exams in some aerospace classes) and the only request from professors was that each student write the names of their collaborators on each problem set.

Absolutely incredible culture that shows how treating someone as a responsible adult can reinforce mature and honest behavior.

> Absolutely incredible culture that shows how treating someone as a responsible adult can reinforce mature and honest behavior.

The neat thing is I've applied these ideas in my subsequent life, and by and large it's been very successful. People tend to rise to your expectations of them. For example, I refused to put copy protection on the software I sold in the 80's, and never had any problems with piracy.

I also run the D programming language forums, and the D Conference (along with Andrei). We refuse to put a code of conduct on either. We simply assume professional behavior, and that's what happens.

I went to a place that was very much not Caltech and had one of the worst bike stealing problems I've ever seen[x], but I had a physics professor who embraced that sort of exam. There were a couple of really hard problems and we had a week, and he basically said "here's the classroom everyone, go at it and see what you come up with".

[x] It was also really weird. People would steal your bike to ride across campus (it was very rural), then someone else would probably bring it back a week later.

Reply on news.ycombinator.com