Announcing algorithm development support for Python 3, JavaScript, Rust, and Ruby

Re:Announcing algorithm development support for Python 3, Ja... by codelitt on Thursday May 19, 2016 @09:40PM (#11735117)

I said this in a reddit thread[1], but I really just have a philosophical aversion to this. The industry has moved forward because of open source code. If I come up with a great algorithm, it is in my best interest to share it with others, have it improved upon, and contribute to the community. What if something like quick sort were only available if I paid someone for the snippet/service?

I might pay to use a service that allowed an implementation not in my projects language (and it's faster, I can't be bothered to write it, etc), but close sourcing or licensing of algorithms seems to go against the entire ethos of the community.

Especially in the wake of Oracle v Google and the implications that it could have on APIs, I really think to purposefully close off algorithms is just another step in a really bad direction.

[1]: https://www.reddit.com/r/ruby/comments/4k4fpr/algorithmia_no...

[ Parent | on HN ]

Re:Announcing algorithm development support for Python 3, Ja... by tomc1985 on Friday May 20, 2016 @12:12AM (#11735500)

This doesn't even look like pay-for-a-snippet (which I agree is a pretty audacious business model). They sell you a black box, which you interact with through REST.

As a firm believer that knowledge -- particularly theory -- should be free, Algorithmia is repulsive. As a developer their black-box model (and not simply giving me source) is a no-go.

[ Parent | on HN ]

Re:Announcing algorithm development support for Python 3, Ja... by anowell on Friday May 20, 2016 @01:12AM (#11735671)

While I also generally prefer to choose open source services over closed source ones, it seems to me there is a place for both....

That said, over 75% of algorithms on the site are open source. Perhaps that's not obvious enough?

[disclaimer: I'm an engineer at Algorithmia]

[ Parent | on HN ]