Comment by msla | Hacker News Reader

Comment by msla | Hacker News Reader

It's the people who make laboratory glassware, which is the classic skilled craft taught by an apprenticeship in addition to formal classwork:

> To master scientific glass blowing, proper training and apprenticeships are key. Only one school in the nation, Salem Community College in New Jersey, offers a degree program.

> In addition to the hands-on training, which requires a knack for precision as well as coordination, students must take courses in organic chemistry, math and computer drawing.

> "You need to know enough about everything, about mechanics, about chemistry, about physics, about thermodynamics — whatever a chemist can come up with, you need to know just a little bit to get that chemist through," said Dennis Briening, instructional chair of Salem's two-year program. "And of course, you need to be very skilled, technique-wise. So it really takes a long time to get to a position like Rick's."

Sounds like a job for people who are bright, good with their hands, and like making money. Really, the whole piece is about how few people are going into this trade, and I can't see how people can't be brought into it if the demand's there.

> Really, the whole piece is about how few people are going into this trade, and I can't see how people can't be brought into it if the demand's there.

Because you may spend years getting a degree in making lab glassware, only to find yourself put out a job because someone invents a glassblowing machine... Or trains up a few thousand people in China to do it for one tenth your wages.

Not exactly motivating me to drop everything, move to Salem, and spend the next few years becoming a glassblower.

> Fellow chemists are relieved Caltech will keep a glass blower on campus. For them, it's more convenient than ordering from a catalog and preserves the intimacy researchers should have with the experiments they design.

Let me get right ahead on staking my livelyhood on a university's budget indulging the 'intimacy' between its chemistry lab and their resident glassblower. Who do you think will be canned when Caltech will need to scrounge up some cash to build a swanky new dorm? "Sorry, Doctor Weinberg, you're going to have to get your glass on Amazon, like everyone else."

The piece is about these people doing very small-batch custom runs, precisely the kind of thing you don't go to China for, and it's hard to fully automate skilled trades.

Today, you can already get small-batch custom runs of PCBs shipped straight from Shenzhen. Nobody keeps a local PCB lithographer on staff, when you can order a batch of 20 for $50.

I'd be surprised if, in a decade, you couldn't do the same for glassware.

With custom-built mechanical equipment it absolutely doesn't work that way. Doubtless, someone in China will build you your piece from drawings, and six weeks later it arrives, and you find that it needs modifications to work as intended. Wash, repeat. The slow feedback loop is the biggest problem with outsoucing. It already doesn't work particularly well with custom synthesis.

Also, when you order PCBs from China, you're doing all the CAD work yourself, and they're building those parts exactly according to the engineering drawings (called "gerbers").

With glassware, the chemist isn't using AutoCAD to create detailed engineering drawings of what he wants. The glassblower is taking on much of the design role. I suppose you could just get the university to hire a mechanical engineer to do the CAD work, and send CAD drawings to China to manufacture the glassware according to spec, but that's not going to be cheap.

You can make the same argument with software development. You can get people overseas to do it at cheaper price.

If software problems were as well-defined as lab glassware specs, that argument would hold some merit.

Also, millions of engineer hours have been put into making the software equivalent of a glassblowing machine. No CS department employs an in-house 'Kernel writer' or 'TCP/IP Implementer'.

The actual OP article suggests that lab glassware specs aren't as well-defined as you imply at all.

> "You need a person who has some idea about the scientific process and how to translate what a student wants to do into a piece of apparatus that really works," said Bob Grubbs, a 2005 Nobel laureate in chemistry and professor at Caltech. "How do you get from a couple scratches on a piece of paper, to something you actually go use in a lab that you trust and is not going to blow acid in your face?"

Doesn't mean those who hold the purse strings will pay for it. (Doesn't mean they'll pay for custom software either).

The problem, according to the article, is that there are few opportunities for apprenticeship. There is demand for experienced professionals but very few job offers for young glassblowers.

See, I have a different take on this. I think that the job should be split up, it's a huge problem and a risk to try to wrap up all of these qualifications into one person.

>"How do you get from a couple scratches on a piece of paper, to something you actually go use in a lab that you trust and is not going to blow acid in your face?"

Engineering. That's how. Instead, the academic community seems to want to rely on rockstar glaziers the same way a startup relies on one or two developers to handle everything from design to implementation to release.

Have the student sit down with someone who can draw proper technical drawings and specs. Then hand those proper technical drawings off to a glazier. Then give the finished glass to a QA person who will verify it against the specs.

That's a lot of overhead when you just want one of something.

But even then someone needs to know a lot about chemistry, glass, and glass manufacturing to understand if something is both feasible to build and useful.

Over time, you could compile a book of technical drawings that could be reused and modified for future products. I'm having a hard time believing that every person who needs glass blown needs something totally unique that has never been done before; it's probably slight modifications of existing products and the occasional odd pipe/manifold.

I think the most common 98% ends up being mass produced and they just buy that. This is only really about covering edge cases so they probably all end up fairly unique.

Even knowing what is and is not slight modifications of existing products requires a fairly deep understanding when you're dealing with highly toxic substances.

Custom work is expensive--one can't just hand-wave it away and hold out for a polymath glassblower, no?

That level of engineering isn't economically feasible for a one-off piece of something, unless you're NASA and have enormous amounts of money to spend on a one-off.

If universities were willing to spend that much money on glassware, they'd just pay glassmakers $750k or so, and wouldn't have any trouble hiring them.

>Full-time university glass blowers are considered tops in their field, but few institutions still offer such positions or give young glass blowers the chance to hone their craft.

Seems like they've created the issue themselves.

Reply on news.ycombinator.com