Live data from Hacker News

Why Wolfram tech isn’t open source (2019)

blog.wolfram.com

21–30 of 316 posts

Re: Why Wolfram tech isn’t open source (2019)

#21

Most of these are about wanting control over the design of the software - but releasing their code doesn't have any impact on that - I don't get it? Java is open source - Oracle still tightly controls and manages the design of Java. For example 'Bad design is expensive' - how releasing your source lead to bad design? That's a complete non-sequitur.

They are arguing it does have impact on that.

The company that created Java went out of business and was sold to Oracle for a fraction of its original worth. So Java was a business failure -- or at least I'd be interested to see someone defend Java as a business success.

So I think it's realistic to at least engage with their argument seriously rather than dismiss it.

The argument is that open source licenses are incompatible with the kind paid license they're offering their customers. These licenses pay for the salary of their skilled employees.

Oracle makes money EXACTLY the same way -- by selling commercial licenses to a non-open-source database. That's why Java can be open source and not bankrupt the company.

Again, the argument here seems to be Just come up with a different business that makes hundreds of millions of dollars and use that to pay for Wolfram !

-----

edit: Since this comment is highly upvoted and visible, I'll mention that I'm seeking funding for a compiler engineer for the Oil shell, which I just mentioned on the blog:

http://www.oilshell.org/blog/2021/12/backlog-assess.html#mor...

So I know from over 5 years of direct personal experience that compiler engineers don't just show up to your open source project and start implementing things! Even if you have tens of thousands of readers on Hacker News for years. (Oil has had 48 contributors but not very consistently, which is understandable.)

To be clear, I did not expect help! I have asked for it, but I thought I would be able to finish it myself. So it looks like hiring someone with specialized skills is the most realistic way to accelerate the project (now that the language design is in good shape).

This funding is to finish the Python to C++ translation, which is already half done. Example:

Oil's Parser is 160x to 200x Faster Than It Was 2 Years Ago http://www.oilshell.org/blog/2020/01/parser-benchmarks.html (lots of progress since then)

I'm applying for 50K euros of funding from NLNet. Also in the process of setting up Github Sponsors. E-mail me at andy@oilshell.org if you want to make a donation -- I think I will need at least $250K to properly pay for this position. If you're a professional programmer and spend 2 hours fussing with a shell script, that's probably worth a few hundreds dollars.

I'll start mentioning this in more posts when everything is set up, but I thought I would use this extremely relevant topic to raise awareness. Programming languages don't pay for themselves! https://www.oilshell.org/release/latest/doc/oil-language-tou...

Re: Why Wolfram tech isn’t open source (2019)

#22

Who are Wolfram’s customers? I recall back in 1993 a girl I knew did an internship at Black and Veatch and she enjoyed it, but is B&V really the target market? If so, I don’t see how “math” is the major draw.

Lots of universities and similar research/educational institutes have site licenses.

Re: Why Wolfram tech isn’t open source (2019)

#23

Most of these are about wanting control over the design of the software - but releasing their code doesn't have any impact on that - I don't get it? Java is open source - Oracle still tightly controls and manages the design of Java. For example 'Bad design is expensive' - how releasing your source lead to bad design? That's a complete non-sequitur.

I don't really understand the definition of open source. I used to hear people say stuff along the lines of "If you just release the source and don't accept patches then that's not real open source"

But the opinion on that has changed I guess?

Re: Why Wolfram tech isn’t open source (2019)

#24
The only point here that _may_ be considered valid is #5: Crowd sourced decisions may be bad for you.

The reason one could consider this to be true is that often open source design is hype driven, and not long term benefit driven. Exceptions can be seen in communities like Rust, but there's still lots of issues there (like async). However just because there's lots of hype does _not_ mean you have to bow to it. You can easily commit to a read-only release of the code which obviates all the need to actually participate in the community.

And of course there's successful models like VSCode, which manages to be _mostly_ open, but all the "cool parts" (like Copilot) are still not possible to reproduce with the provided source alone.

Many ways to open source your code and still have people pay for it!

Re: Why Wolfram tech isn’t open source (2019)

#25

Earlier quoted context omitted.

Even native python is not as slow as people expect. And scientific languages are terribly slow, see for example loops in Matlab, which are terribly slow.

FORTRAN, the OG of scientific languages, is not slow. In fact, that's why it's still used to this day.

That's true, i was more thinking of scientific interpreted languages, a la R, Mathematica, matlab

Re: Why Wolfram tech isn’t open source (2019)

#26

Earlier quoted context omitted.

Even native python is not as slow as people expect. And scientific languages are terribly slow, see for example loops in Matlab, which are terribly slow.

FORTRAN, the OG of scientific languages, is not slow. In fact, that's why it's still used to this day.

In fact it may well be the fastest language - although it's hard to compete with the sheer number of man hours spent on C/C++ compilers.

Re: Why Wolfram tech isn’t open source (2019)

#27
post #7

Most of their points are valid for community driven open source projects but the existence of large corporate ones like Chrome, Android, .net, VS Code, Tensorflow and lots more prove that exerting tight central control while being fully open source is very much possible.

This seems too obvious to repeat, but Google and Microsoft fund Chrome and VS Code through the profits from their other products like Search and Windows.

Wolfram doesn't have that, and most businesses don't. It's very easy for us say "just make it open source" if we don't have to come up with a way of paying people! Just come up with a different business that makes hundreds of millions of dollars and use that to pay for Wolfram !

Re: Why Wolfram tech isn’t open source (2019)

#28

TLDR: some types of software needs more than programming. omitted TLDR: some of the above can only be done by paying people money to do them. Company needs income for that. omitted exibit: why haven't Gimp thrown Photoshop, Pixelmater and others from the throne. If programmers were the only people needed, then all those commercial image editors would be gone from the market. (but a unix shell only needs programming b…

Blender is a counter example. They prioritized UX and are now eating into commercial use cases.

Re: Why Wolfram tech isn’t open source (2019)

#30

Earlier quoted context omitted.

At first I thought, "how could anything be slower than python?" And then I figured that you probably would've ported it to leverage things like numpy which makes a lot of operations into native code and takes advantage of highly optimized blas libs, etc.

Even native python is not as slow as people expect. And scientific languages are terribly slow, see for example loops in Matlab, which are terribly slow.

Mathematica is the worst when it comes to pure number crunching, because they use unstructured arrays. Much more flexible, but much slower.
Post reply on HN