Hacker News Reader

Comment by vectorEQ | Hacker News Reader

killing things is bad mkay. prevent instead of cure. solve the root cause.

" “The tidal wave of data is still teaching us fundamental concepts about the interaction of the human immune system and human cancer.” "

So, if they still learning fundamentals. -> impact of actions based on inaccurate or incomplete knowledge, is unknown.

in programming this kind of practice would lead to bugs / undefined behaviour etc. - imagine what kind of shit that would mean for human health...

It'd be nice if they had cancer prevention studies instead of cancer cure studies. kthx

View on HN

Many of the DNA mutations which cause cancer are essentially random and impossible to prevent. The human body is not a simple deterministic computer system.

If this were true, all I could express would be gratitude biologists don't think like programmers.

Often the alternative to trying a treatment that you don't fully understand is death.

While we're still trying to sort out the root causes, a lot of people are dying. So we're trying to stop them from dying, even while trying to find root causes. (Society as a whole can work on more than one thing at once.)

When dealing with a software bug you have to reproduce the issue before you can definitively "fix" it.

Not true. (It is often possible to understand the cause by just looking at the damn code.)

Maybe for superficial bugs in small toy programs. But a real bug in any sophisticated program requires a bit more than looking at the code most of the time. User feedback, log/trace data, performance data, system events data, memory dumps, etc.

Of course the bug could be in proprietary software your code is using. How would you fix that by purely looking at your own code?

Humans, and every other life form have an undocumented, undesigned self executing specification that consists entirely of the most reticulated, self referential spaghetti code compatible with successful reproduction and it’s not just riddled with Heisenbugs, they’re core parts of the machinery of how everything works. We’ll get there eventually but it’s going to take an exceedingly long time.

You are looking at human readable and annotated code, not what is compiled for the computer to run. Imagine finding bugs by looking at binary. DNA is more complex than 1 or 0, it could be ATCG. There are 3.2 billion nucleotides in the human genome, and we are still a long way from understanding the nuances of the DNA 'programming language.'

Reply on news.ycombinator.com