Comment by shadowmint | Hacker News Reader

Comment by shadowmint | Hacker News Reader

Common sense speaking. Patents are indefensible nonsense.

That said, I think the point he raises is interesting. There are some ideas which are entirely algorithmic and are hard.

Specifically complex technical standards like video compression and encoding, machine vision, wireless signal processing, etc.

These are things that require a significant capital outlay to invest in research to solve the problem, but once it's solved... well, someone else can clone the logic and make a free version. So you patent the process.

If you dont, you cant guarantee a return on investment in the project, and thus you dont get investment in the project, and ultimately the problem doesn't get solved.

Its an interesting thought experiment, and it makes for a good argument for patents existing. ...but then, where do you draw the line between "complex and costly" and "stupid and trivial".

It's an argument I've heard a lot; I think it's nonsense too, but it convinces a lot of people. It'd be interesting to see a thoughtful critique of that argument, rather than the 'stupid patents are bad' argument, which, basically everyone already agrees on.

> If you dont, you cant guarantee a return on investment in the project, and thus you dont get investment in the project, and ultimately the problem doesn't get solved.

Well, it doesn't get solved by for-profit actors as a way to make money.

But it can get solved by government-sponsored research. You can look at patents as a complicated tax system intended to fund innovation. We might be better off just making it an actual tax system that funds education.

It can also get solved by foundations, or by for-profit actors working in non-profit-seeking ways. Linux is a great example of innovation in that model. Industry consortia are an interesting model, as is something like Underwriters Laboratories.

In such cases, you try to convince people to pay for your idea up front. Depending on how revolutionary/badly needed this idea is, it will get funded.

Kickstarter and clones are the answer.

If you can already prove the idea isn't nonsense then there is no need to fund the R&D. If you can't prove the idea isn't nonsense, no one will fund you. Research tends to be either self-funded or backed by funding organizations that specialize in evaluating high-risk R&D projects.

Even among the organizations that specialize in it, there is a problem that there are only a handful of people in the world that can even begin to evaluate the potential of a proposal. And in those cases, the people doing the vetting do not fully understand it, but they are the best outside experts capable of detecting any subtle red flags that suggests it will fail. This is the normal state of affairs for advanced research; no normal person could hope to evaluate advanced research.

Kickstarter type models would be poor for funding most types of serious research. Leaving this to popularity contests like Kickstarter is a recipe for a lot of clowns and wannabes with good PR getting funded in the absence of serious (and expensive) diligence and serious research being ignored because it can't be wrapped into a shiny soundbite that everyone understands.

When your whole model revolves around it, the effort that goes into evaluating the prospects of serious research topics will perhaps undergo a massive change as well.

>That said, I think the point he raises is interesting. There are some ideas which are entirely algorithmic and are hard.

Clever and/or sophisticated perhaps but not hard. The hard parts are the mathematical foundations, done by mostly dead people who are not getting a penny out of it.

>If you hold a software patent you should feel bad about it.

The problem is that they are sunning their arse in the Bahamas and are feeling great about it.

Reply on news.ycombinator.com