Hacker News Reader

Comment by rp1 | Hacker News Reader

Almost none of the reasons have anything to do with releasing the source. You can release the source but still maintain central control. They shouldn’t even have posted this. Obviously they want to make money from their products, and that’s fine.

View on HN

Yea, it's embarrassingly bad.

It sounds like their idea of open source is something along the lines of "slap a MIT license on it, throw it over fence, and hope random nerds on the internet finish it while we wait idly."

> You can release the source but still maintain central control.

Until someone with a larger budget than you forks it and uses their immense resources to force you out of the market. Then you liquidate your company and someone else gets the rights to maintain your fork.

I think this is an underrated point that has been more noticeable with AWS in particular. Their own CentOS spin, their own ElasticSearch distro, their own elastic cloud, etc.

There is zero chance that Wolfram's products would be "stolen" in this way. No one is interested in developing them but Stephen Wolfram.

I’m sure some cloud provider would love to slap wolfram features on to their existing service offerings as a value-add.

How about using AGPLv3?

When people say "I'm sure" I'm always just floored. Why are you so sure? Have you spoken to these supposed cloud providers and had them tell you this?

Here's one I'm sure about: no one cares about Mathematica except a couple of schools, professors, and finance wonks who've clung to it since the early days of CAS. The language is wacky enough that no mainstream user would prefer it (even with all of its built-ins) over Matlab or scipy. And I say this as someone that has developed things in Mathematica.

Even if you’re worried about this, there are source-available licenses like SSPL or Elastic License v2 explicitly designed to prevent this, and which have seemingly been successful at doing so. The exact choice of license would depend on the kind of thing you’re worried about.

I mostly agree, at least in the current day where the powerful companies in context are [x]aaS companies, but now you’re going down a road where people will argue whether OSI/FSF are legitimate gatekeepers of the definition of “Open Source”

That’s exactly why I used the phrase “source-available”, not “open source” - because those licenses aren’t recognized by the OSI. I’m actually of the opinion that if AGPL is open source, SSPL is too, no matter what some dorkwads as the OSI say. But I’m employed by a company that uses the SSPL, so I try to use the generally accepted phrasing to avoid being accused of “sowing confusion”.

The irony certainly is that wolfram uses open source libraries (and there were some twitter threads that they might be aying a bit fast and loose with the acknowledgement of LGPL licenses)

Indeed: https://twitter.com/hypergeometer/status/1470505437934067725

I think what he has in mind is not "open source" per se, but the way open source projects tend to organize development—distributed development, unpaid volunteers, open communities, etc. He wants the Cathedral, not the Bazaar—and that's fine; there's no reason everything should be developed by a community of volunteers. But, as you say, there's nothing to stop them centralizing development but releasing source code.

(Plus, I think putting Stephen Wolfram at the top of a Linux kernel-type project is unlikely to work well at all.)

Reply on news.ycombinator.com