Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Not to be that guy, but I often say software engineering as a field should have harsher standards of quality and certainly liability for things like these. You know like civil engineers, electrical engineers and most people whose work could kill people if done wrongly.

Usually when I write this devs get all defensive and ask me what the worst thing is that could happen.. I don't know.. Could you guarantee it doesn't involve people dying?

Dear colleagues, software is great because one persons work multiplies. But it is also a damn fucking huge responsibility to ensure you are not inserting bullshit into the multiplication.



Some countries such as Canada have taken minor steps towards this, for example making it illegal to call oneself a software engineer unless you are certified by the provinces professional engineering body, however this is still missing a lot. I also don't wish to be "that guy" but I'll go further and say that the US is really holding this back by not making using Software Engineer as a title (without holding a PEng) illegal in a similar fashion.

If we can at least get that basis then we can start to define more things such as jobs that non Engineers can not legally do, and legal ramifications for things such as software bugs. If someone will lose their professional license and potentially their career over shipping a large enough bug, suddenly the problem of having 25,000 npm dependences and continuous deployment breaking things at any moment will magically cease to exist quite quickly.


I'd go a step farther and say software engineering as a field is not respected at the same levels as such certified/credentialed engineers, because of these lacks of standards and liabilities. Leading to common occurrences of systemic destructive failures such as this, due to organization level direction being very lax in dealing with software failure potential.


I don't know, I get paid more than most of my licensed engineer friends. That's the only respect that really matters to me. Not saying there might not be other advantages to a professional organization for software.


I feel the same way but do agree there’s a general lack of respect for the field relative to other professions. Here’s another thread on the subject https://news.ycombinator.com/item?id=23676651


Respect has to be earned.


I believe instances like this will push people to reconsider the lax stance. Humans in general have a hard time regulating something abstract. The fact that people can be killed is well-known since the 80s', see https://en.wikipedia.org/wiki/Therac-25


I once worked on some software that generated PDFs of lab reports for drug companies monitoring clinical trials. These reports had been tested, but not exhaustively.

We got a new requirement to give doctors access to print them on demand. Before this, doctors only read dot matrix-printed reports that had been vetted for decades. With our XSL-FO PDF generator, it was possible that a column could be pushed outside the print boundary, leading a doctor to see 0.9 as 0. I assume in a worst worst case scenario, this could lead to an misdiagnosis, intervention, and even a patient's death.

I was the only one in the company who cared about doing a ton more testing before we opened the reports to doctors. I had to fight hard for it, then I had to do all the work to come up with every possible lab report scenario and test it. I just couldn't stand the idea that someone might die or be seriously hurt by my software.

Imagine how many times one developer doesn't stand up in that scenario.


This is why I made that point, similar to you I would not stand for having my code in something that I can't stand behind, especially if it potentially harms people.

But it should not hinge on us convincing people.


I'd endorse this. That way when my hypothetical PHB wants to know why something is taking so long I can say "See this part? Someone could die if we don't refactor it."


Related talk by Alan Kay: https://youtu.be/D43PlUr1x_E


It’s important not to disregard that software engineers are often overruled by management or product when strict deadlines and targets exist.


"If only we asked harder problems for our leetcode interview!"




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: