Hacker News Reader

Comment by pnathan | Hacker News Reader

A solid solution is to have a professional union/association with a meaningful exam for entrance, meaningful certifications for specialties, and the mandate to avoid committing professional malpractice (strike/walkout enforced). The ACM could do that, but has chosen not to: the IEEE's attempt is sort of a joke.

Having spent the last 3 months totally involved in the US medical field regarding a family member's life-critical condition, I can say with confidence: the professional credential/certification process works really well when built well, and software engineers with an interest in seeing our field should strongly consider the medical community as a working example - remember - "rough consensus and running code".

View on HN

And aviation. Not pretty but it definitely works.

I definitely agree that good professional organizations and well-functioning unions and good worker protection laws gives a good total system.

However, one thing that must be kept in mind when comparing to e.g. healthcare or aviation is that progress in those fields happens at a pace that's positively glacial as compared to software development. The development time for new airplane tech or healthcare tech can easily be a decade or more. At that pace, employers don't perceive "you're old and don't understand the new stuff anymore" to nearly the same degree.

Some of the stuff that happens regularly here in software land is like if suddenly accupuncture became the best cure for most diseases, or if pilots were given new planes with flight controls manipulated by moving your eyelids in a special way.

Can you cite specific improvements in the past 5 years that aren't more reinventions of previous wheels?

Yes, this old saw that software changes so quickly that only kids can keep up seems BS. I've been a developer for 30+ years and I still do the same thing I've been doing for years. There's some new frameworks that come along that arrange things a little differently, but if you know C (well) then all the other things are just variants imho... and yes I've built web apps, coded industrial micro controllers, written SQL databases and so on. The things I learnt last century I still use every day. (Though I would say the level of quality and documentation has decreased and unnecessary complication has increased since the 90's)

Purely functional UI (React) stopped being fringe academic topic and became a mainstream.

But tech like Flux and Redux are clearly derived from the Command pattern (of course i've never seen or heard reference to these patterns in the docs or videos for for flux and redux), and also seem pretty similar to Java enterprise patterns so arguably were already in the mainstream 10 years ago (at least). See http://www.corej2eepatterns.com/Patterns/ServiceToWorker.htm http://www.corej2eepatterns.com/Patterns/FrontController.htm http://www.corej2eepatterns.com/Patterns/DispatcherView.htm

I work daily with javascript and enjoy it but it frustrates me that people treat aforementioned tech as cutting edge but these ideas were already around a long time ago. The counter to this is that javascript is finally maturing and starting to use these patterns which is a positive thing, I just wish the people promoting these "new" technologies would give recognition to what came before.

I won't say anything as to what is good progress vs. reinventing the wheel. For the purpose of this discussion, I don't think it matters much; if employers expect you to code using React/Angular/Bootstrap/whatever, it doesn't matter if you think those are inherently good progress on previous tech or not.

And there are definitely software devs working in fields that change more slowly. But if you take web development as we frequently do around here, e.g. the joke article posted yesterday about "Hello world" in PHP does provide some good examples of stuff that's changed dramatically in the last five years.

https://kukuruku.co/post/hello-real-world-in-php-in-2017/

Convolutionnal neural network (called IA in the news, will automate previously safe jobs)

Upgrade in hardware

Bitcoin technology (made it's way into traditionnal banking)

And if you want more thing into programming:

In the last 15 years the Racket team has came with contracts (something with high order function that didn't existed before), gradual typing which generate contracts, macro composition. None of them existed before.

I'll take bitcoin and NN advances as highly valid and true responses to my query (although tbh I think btc is a bit older than 5 years). To answer my own question, I think Rust is a substantive improvement.

But, that really doesn't affect the vast bulk of programming: what we do isn't changing that fast.

We changed many, many details. Polished many tools, even completed replacing a few more fringe ones.

Now, take a look of what changed in aviation tech in the past 5 years. I don't even think you'll be able to. Aviation does not change on that speed - your 5 years time frame is nonsense there, like asking what changed on IT in the last half hour.

Aviation tech isn't stagnant because of their unions and professional associations. It's stagnant because the FAA mandates it be stagnant. Go read up on what you're allowed to do with a Cessna you just bought. You can't so much as run a USB charger out to the front panel without the FAA coming down on you like a ton of bricks. Disruptive avionics start-up? Get real.

And imnsho that's a good thing. Experimental aircraft are labeled experimental for a reason, and anything that gets sold for general aviation purposes should do what it says on the tin, no more, no less.

If aviation adopted the 'move fast and break stuff' paradigm it would be the end of aviation as we have come to rely on it. No doubt there would be a huge amount of technical progress in a short time but for the purpose of enabling commerce and having reliable transportation it would be an absolute disaster in every sense of the word.

I'm just pointing out the root cause being regulatory and not the existence of professional organizations. I personally find the dual-world "experimental" vs. certified regulatory environment to be a pretty good compromise.

> I personally find the dual-world "experimental" vs. certified regulatory environment to be a pretty good compromise.

Agreed. The accident rates in both categories are also a good indicator.

Agreed. But that means that progress in aviation does happen in a slower pace, and that must be taken into account when comparing with other areas.

Reply on news.ycombinator.com