Pyodide: Bringing the scientific Python stack to the browser | Hacker News Reader

Pyodide: Bringing the scientific Python stack to the browser

hacks.mozilla.org

90 pointsmmohades26 comments

Open articleView on HN

Previous discussion: https://news.ycombinator.com/item?id=17830226.

Actual previous discussion from less than a day ago: https://news.ycombinator.com/item?id=19677721

I guess this slipped through the dupe detector due to https://news.ycombinator.com/item?id=19684806

Just so, thanks!

% python —version Python 2.7.15 %

I get: 3.7.0 (default, Mar 21 2019, 17:54:42) [Clang 6.0.1 ]

Does not work in Safari for me (latest Mojave version).

Happily works in Chrome and Firefox.

Can we update the link to the original technical post?

https://hacks.mozilla.org/2019/04/pyodide-bringing-the-scien...

@dang

Sure! Updated from https://venturebeat.com/2019/04/16/mozilla-details-pyodide-a....

What.... why? Why must we insist upon piling so many features on top of browsers? It's a web browser, not an application platform...

>It's a web browser, not an application platform...

I'm assuming you have just arrived from 1996. I can't wait until you find out who our President is!

"Oh God, it's Arnold Schwarzenegger, isn't it?"

I hate to break it to you, but the best interface for scientific python is Jupyter Notebook, which already runs in the browser. This just moves the execution of the code into the browser also.

It's also not adding any features that weren't already there, unless you're objecting to the entire WebAssembly project, in which case I guess that's allowed, but a bit off-topic.

Because it is. Maybe not the platform we would've designed, but it's one we have.

Welcome to the world of the Browser Application platfRrm fun. I call it B.A.R.F

You're using a web app now.

Is this bad for the long term preservation of knowledge? Sure. But it's the world we made.

How are webapps worse for the preservation if knowledge than native apps? I can't open most of the apps I used on a daily basis 10 years ago.

For reasonably complex documents PDF/A is a pretty good choice for long-term archival:

https://en.wikipedia.org/wiki/PDF/A

That said, the comparison I was trying to make was the pre webapp web to the current web. We have AOL today, just not one big company running it. It will mostly be lost because it's not legal to save it.

The future will know us through scraped videos from Youtube before it was taken down suddenly and without warning in the mid 21st century due to corporate politics and/or the takeover of the government in DC by descendants of the house of Habsburg.

It doesn't add any features to the browser. It's just a Web Assembly application.

This is not a new feature. It's a wasm library/app that you can use to code for your browser in Python.

So, it's actually pretty cool.

Also, the web browser has been an application platform for nearly 20 years now, so I'm not sure I understand your remark.

It's not included in the browser, it's just a web page.

Browsers are application platforms and have been for a while. It is the most portable and accessible method of software, if the features you need fit within its sandbox.

Dear Mozilla, I love you, but stop working on these side projects. Your work in these non-browser projects is excellent, but pointless.

It may be that I don't understand Mozilla, but I believe they should focus on Firefox, Thunderbird and MDN. Even Rust is sort of a weird CS project that you don't technically need, but fair enough, if makes Firefox development easier, go for it.

Mozilla get most of their money from search engine companies, so may they should start focusing on trimming down to the point where they could survive without that big Google check and drop some of the side projects.

And why is it that there's no "Support Us By Donating" button on www.mozilla.org? I feel there should be.

This is exactly the kind of thing Mozilla should be working on, in my opinion. It's enabling something new to work in a browser, and if Mozilla didn't do this Google would have, only it would be locked into Google Sheets or something.

Maybe, but is this honestly solving a problem for anyone but the slimmest niche group. Thimble is no more and in my mind that's a similar kind of project (I bet most of you forgot that Thimble was a thing). Firefox OS is dead, Persona is dead, Pocket and Send is still being questioned.

Why should Pyodide or the whole Iodide project be any different. Other than Firefox, Thunderbird and MDN, only Bugzilla and Rust really stands out as successful projects, and that's only because they've been developed to solve an actual problem.

I'd be surprised if the Iodide project will see any real usage 5 years from now. Are any data scientists really look to move to a browser based workflow?

This project was designed to meet the internal needs of data science at Mozilla. Them releasing and developing it openly likely has no/little impact on the amount of energy going into their core products.

I think it is good, if for no other reason, than it gives a Mozilla a large project they can use to test how well the browser, dev-tools, etc. perform with a large complex web application. I think one reason why the chrome dev-tools is better than the firefox dev-tools is that they have a lot of other projects where developers need to make heavy use of the dev-tools for large complex web appjs.

Reply on news.ycombinator.com