Live data from Hacker News

Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

news.ycombinator.com

171–180 of 231 posts

Re: Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

#171

Earlier quoted context omitted.

Make the publishing of anything scientific independent from Journals and freely available to anyone.

Isn't that Arxiv.org?

Do tenure committees and other such parts of the scientific establishment give publication on arxiv.org the same weight they give to publication in traditional scientific journals? They didn't used to -- has this changed?

Re: Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

#173

A better, no hassles international payment gateway (I know Paypal exists, but I am not their fan)

Spreedly? Or maybe that's a layer on top of gateways? I don't know much about payment gateways, so feel free to enlighten me on why it's not a solution to this, if it's not.

Re: Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

#175
post #143

Earlier quoted context omitted.

I think by definition we already have such. Perhaps you should increase the specificity of your desire.

How about a re-usable electricity storage method with a power density equal to or better than petroleum?

Now we're talking.

Re: Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

#176

Online elections. Think of how many more people would actually vote in the US if it were online . The 2008 election had 63% voter turnout, which is historically high, but is nothing compared to countries like Italy with 90%. Obviously, there are huge security concerns and it would be a logistical nightmare, but that's what someone needs to solve.

If not proper elections at least a whole population simulation before the polls. That way people could express their true preferences without fearing that their vote will not count if it goes to a fringe candidate and would reassure them that they're not as few as regular polls and eventual results usually show.

That way people could express their true preferences without fearing that their vote will not count if it goes to a fringe candidate

Isn't that exactly what instant-runoff voting solves?

Re: Ask YC/HN: What's a problem, any problem, you'd like to see someone solve?

#179

A practical, robust method for separating essential complexity from accidental complexity in software engineering. Trevor Blackwell has written about something like this: http://tlb.org/busywork.html

Thanks for the pointer. Basically, Trevor wants the distraction of low-level tweaking so he can think about the high-level problem. That's the weakest argument for low-level languages I've ever heard. But Python fits the bill: Python for high-level, C libs for performance and low-level tweaking.

The big solution I was thinking of would have to go a bit further than Python/C. Basically you should be able to write a description of the actual essential functionality of a program without going into data structure and algorithm implementation that isn't relevant to the problem domain. Something like Prolog or Haskell here, Python isn't all that declarative compared to them. Then you could write the bits that need tweaking in a lower level language, but you should be able to automatically verify that the low-level implementation matches the high-level description, at least through extensive QuickCheck-style automated testing if a full automatic proof isn't feasible. You don't get anything like this with a Python/C style solution.
Post reply on HN