Urllib3, Stripe, and Open Source Grants | Hacker News Reader

Urllib3, Stripe, and Open Source Grants

medium.com

205 pointsmalditojavi20 comments

I'm working at a tech company that is leaning towards such support. Have any suggestion on how would we go about doing such grants well?

We are in touch with the community around our projects, and they are doing awesome stuff. I'd love to be able to enable them more. Been looking into Bountysource a bit, but still figuring out the workable model.

I'd advise you reach out to the maintainers of some of your favourite projects and ask for their help in finding good candidates for grants.

Or do what Stripe did and put out a call for applications/proposals.

Another thing you could try is giving your engineers a quarterly/yearly budget to allocate to OSS projects, then they could do things like Gittip.

Either way, we need more companies trying these kinds of things so we have a better idea of what works. :)

> Another thing you could try is giving your engineers a quarterly/yearly budget to allocate to OSS projects

I'd expand on that by pointing out the following:

Your company's engineers are almost certainly using open-source projects on a daily basis, and finding bugs or missing features in those projects. Often in such a situation the engineer has no time to code the missing feature, much less see it through an upstream pull and release, so they will hack up a quick internal fork or custom rewrite.

Ask your team that if they find themselves in such a situation, to bring the project and feature as a candidate for your grant. For example, as Andrey mentioned in the blog post, he recently pulled code from Kevin Burke to add customizable retry support to urllib3. That's something my employer could use, but we'd need to add support for the 503 Retry-After header (which we added in-house). I'll be bringing that up as a possible grant with my employer, to incorporate that feature upstream so we can just use it through urllib3.

I applied with http://www.use-the-tree.com. It's okay with me to not have been interviewed via Skype and not receive funding.

Also the eMail from Greg (on June 5th) was nice and well written:

Hey,

Thanks a lot for applying to the Open Source Retreat. Unfortunately, we aren't able to accept your project. We ended up with about 120 applications, and ultimately only had 2 available slots, which meant we had to turn away a lot of projects that we would have loved to sponsor. We'll be announcing our selectees soon; watch our Twitter feed (https://twitter.com/stripe) for details.

We're treating the retreat as an experiment, but the strong interest we've seen makes it likely we'll do something like it again (and we hope you'll apply if/when we do). In any case, we expect we'll learn a lot from this run and hopefully have lessons to share back with the community. We've also had a number of companies express interest in doing something similar, so hopefully other similar opportunities will arise soon.

- gdb

Did Stripe require that some specific feature was added or fixed? Especially if they did, that $3750 is a small amount of money to pay get some functionality to work. Not to mention, that they get some nice publicity from this.

Regardless, kudos for Stripe for sponsoring open source projects.

I think all parties win in this, not just Stripe. Urrlib3 developer and users (new features and enhancments), and the whole FOSS community (if more and more companies start doing this). Also if the competition is fierce for the top talent, doing these kind of "support the community" acts, surely won't be seen as negative things for the company that does these things.

So kudos indeed for Stripe! Let's hope this comes "a thing"

What does he normally do full-time?

Apparently a lot — http://shazow.net/

I know, right?

Ahem. All kinds of things, maintaining a bunch of open source libraries and tools (https://github.com/shazow), bootstrapping a service that emails you analytics summaries (https://briefmetrics.com/), building random apps that I happen to need (a partial list on http://shazow.net/).

I suspect your underlying question was whether I am employed full-time—I am not.

I wish I could sponsor your next two weeks. Let's hope that Stripe's initiative catches on with others.

Thanks! :) Until you have corporate funds to back a bigger project, little ad-hoc donations are also appreciated:

https://urllib3.readthedocs.org/en/latest/#sponsorship

Well, I was wondering how you make a living if you don't get contracts/contributions like this. Not to pry too much but I guess the services you set up are actually profitable or do you have to support yourself by contracting?

I have begun freelancing myself and find that I have a lot more time for my own projects and open-source contributions. Still, sometimes I wish I could drop the contracting all together.

I'd love to one day have something that I created or contributed to, in my own time, be popular enough to get hired to work on it.

For much of the time I've been employed full-time (you can find my resume in the footer of http://shazow.net/), which made maintaining open source stuff that much harder. I've saved up money and reduced expenses to allow myself to take time to work on other things.

Some of my projects are bringing in money, but not enough to cover my living expenses. Occasionally I'll take on a contract.

Awesome to see Stripe supporting open source. One of my gov clients does this as well ... I worked with the broadcasting board of governors and we worked together on enhancing Mozilla's popcorn maker, but tailoring it to journalists. In general I think they're trying to work on open source as much as possible which is cool

I'm so glad Stripe paid you to work on urllib3. I'm building a project that uses your lib, if became something big, i'll pay you too! In the meanwhile i hope some other companies finance your work.

"51 files changed, 1809 insertions(+), 633 deletions(-)." - am i missing something, or is amount of code changed a good thing? Like was there a lot of bad code, and now it's changed to be good...or is this just a measure of improvement like "look at how much work we did". I dunno, i'm out of the loop on this stuff i guess.

All code metrics are vanity metrics, but they're fun to look at. If you want a deep sense of what was changed, take a look at https://github.com/shazow/urllib3/blob/master/CHANGES.rst and the individual issues referenced.

Edit: In this particular case, it's simply the output of `git diff --stat`

>am i missing something, or is amount of code changed a good thing? Like was there a lot of bad code, and now it's changed to be good...

That's the idea when you're doing a rewrite, especially after years of experience with the pain points, and wanting to address specific issues/bugs.

That the old code was bad, and you add new code that fixes some stuf.

It might not be perfect, but then nothing is.

I think you're trying too hard to find badness. It's just a simple few data points of what the author was able to achieve in the two week sprint. It's reasonable to assume it was predominantly improvements.

not really trying, saw it and didn't get it - figured the forum would have answers, downvotes seem to disagree, but whatever

Reply on news.ycombinator.com