HNHacker News
TopNewBestAskShowJobs

pbiggar

7,962 karma · joined January 26, 2009

Cofounder of Tech for Palestine (https://techforpalestine.org)

Previously founder/CTO of Dark (http://darklang.com), and founder/CEO of CircleCI (https://circleci.com). PhD in CS.

https://hachyderm.io/@paulbiggar

https://bsky.app/profile/paulbiggar.com

https://twitter.com/paulbiggar

https://blog.paulbiggar.com

Contact: paul@techforpalestine.org or paul.biggar@gmail.com

submissionscomments
pbiggar··on Safe dead code removal in a pure functional language
In Dark, we label (either manually for stdlib, or automatically for other functions) whether they are pure or not. Pure functions (most of them) do not need their results stored in traces (Dark has an always-on-debugger that uses "traces" saved from real inputs). Making a (mostly) pure functional language enables so much!
pbiggar··on Robinhood apparently selling people's shares ”for their own good“
If you read the subsequent messages, there's a strong suggestion that these are shares bough on margin call.
pbiggar··on Show HN: StaticBackend – A web and mobile back end without lock-in
Looks really cool, nice job!
pbiggar··on Show HN: StaticBackend – A web and mobile back end without lock-in
Disclaimer: I'm founder of DarkLang, a sorta competitor.

While I personally agree that self-hosting is a nightmare, the biggest complaint we've heard about Dark has been related to vendor lock-in. I think this is a pretty good selling point for StaticBackend, for a few reasons: 1) what if you get too big for the hosting? 2) what if the company goes go out of business! 3) what if the company determines that hosting you is against the T&C (see the Parler thing this week, for example, although I completely agree with AWS' decision there).

I don't think self-hosting is really a good idea if you can avoid it, but a lot of people are soothed by the idea that they could, if they really needed to.

pbiggar··on Ask HN: Co-founder wants me to leave but won't entertain a buy out offer
This is a delicate and nuanced matter that's hard to know the right answer without all the details. I've gone through this 4 times, and also recently talked to 20 people who went through it, specifically about this topic. Feel free to email me and we can have a call and talk you through the options (paul.biggar@gmail.com)

First thought: You don't have 40%, you have at most (11 months / 48 months vesting) * 40% = 9%. But, at 11 months you haven't hit your cliff, so you actually have 0%. That is, if they can fire you. It sounds like they're the CEO so they probably can, but if not the board can.

pbiggar··on A tiny island nation just launched major effort to win back control of its ccTLD
Good, I hope they get it. The fact that anyone other than the Niue people get to decide what to do with their domain is ludicrous. The claim of "neo-colonialism" is exactly right.
pbiggar··on ReScript 8.4
It has different goals from them: https://github.com/darklang/tablecloth
pbiggar··on Advanced Compilers: Self-Guided Online Course
Even if you don't do the course, the first paper they list is a fascinating must-read:

Producing Wrong Data Without Doing Anything Obviously Wrong! Todd Mytkowicz, Amer Diwan, Matthias Hauswirth, and Peter F. Sweeney. ASPLOS 2009.

https://dl.acm.org/doi/10.1145/2528521.1508275

I happened to be at that ASPLOS and I thought it was fascinating. Good to see it recognized as one of the most important papers in compilers.

pbiggar··on Docker's Second Death
Thanks!
pbiggar··on Docker's Second Death
Fully agree. I'm going to keep your post in mind when I write about the business model of darklang.
pbiggar··on We Rewrote Everything in $hotlang, and Our Startup Still Failed
Hey, thanks for the mention! Happy to explain my thinking here.

Regarding marketing and selling what we have: we tried that and it's pretty clear we're missing important bits, in particular packages for common vendors (stripe, slack, etc), and a package for user management, as well as common things like validation. People love the concept (as do you, thanks!), but were struggling to build things because of what was missing.

As we went to build these, the tech/product debt that we had built up during our experimental phase really started to become apparent. We were missing common and expected things like user-defined types (maybe not important for dynamic langs, but we're a statically typed functional language, so being able to create types is kinda important), configurable (or at least changable) HTTP middleware, and a few other things.

The reason for the rewrite is that had been taking long detours around our tech stack for a while. Now that there's limited resources (as in, just me), the cost of taking the long way to everything we build is just a little too high, so it feels like a rewrite (note: just of the backend, the majority of the code is in the frontend and doesn't need to change!) is necessary to get to product/market fit.

I also hope it won't be the death of us. Fortunately, F# isn't a valid value for $hotlang! (Good thing I didn't pick Rust!)

pbiggar··on Most-mentioned links in Hacker News comments, 2006–2015
Was kinda surprised to see https://circleci.com at #37. When I clicked through, turns out it was me repeatedly reminding anyone who'd listen that we existed.
pbiggar··on Why Dark didn't choose Rust
Right, it's to make it more scalable. Since our users can write arbitrary code that runs on our servers, and can call 3rdparty servers, it's really not that hard to DOS us accidentally. Switching to async means that we're not just sitting there wasting resources when users do common stuff.
pbiggar··on Why Dark didn't choose Rust
The premise of Dark is that we can remove a huge amount of accidental complexity by having an integrated toolkit of language, editor, and infrastructure. This allows for three major steps forward:

- we're able to deploy (safely!) in 50ms rather than minutes because of it. [1]

- we completely host the infra and you don't need to think about it (including DBs and queues, which are nicely integrated into the language) [2]

- trace driven development, where we use real values from production and show you them in your editor [2]

    [1] https://blog.darklang.com/how-dark-deploys-code-in-50ms/
    [2] https://darklang.com/launch/demo-video
pbiggar··on Why Dark didn't choose Rust
Thanks. Just updated the post to be a bit clearer on this.
pbiggar··on Why Dark didn't choose Rust
Yeah, that's sorta right. I didn't want to evaluate every language under the sun. Rust I assumed had similar semantics to what I had been using before, which is why it was so heavily considered.
pbiggar··on Why Dark didn't choose Rust
Yes, that's right.
pbiggar··on Why Dark didn't choose Rust
> If it's GC'd, how is that implemented?

The language in which the Dark interpreter is built has a GC, and we just use that.

A GC is a huge undertaking, and we already have to build an editor, a language, a stdlib, and infrastructure.

pbiggar··on Dark’s new backend will be in F#
I wrote that in latex a decade ago, and don't know how to update it anymore!
pbiggar··on Dark’s new backend will be in F#
No, feature flags are used to wrap an expression and conditionally choose another expression. Its a fancy if expression. In Dark you add one by selecting `add-feature-flag` from the command palette.
pbiggar··on Dark’s new backend will be in F#
We're trying to do a new thing. If people don't like it, that's fine. So far, people love the concept (no infrastructure, trace-driven development), but it hasn't been complete enough, so that's what I'm working on.
pbiggar··on Dark’s new backend will be in F#
That's right, we version all stdlib functions. I talk about it a bit here: https://blog.darklang.com/how-dark-deploys-code-in-50ms/

When we add a new version of a function, all the calls to the old function are unchanged. The autocomplete shows new users the new version (and not the old version), while people who are calling the old version get shown both.

pbiggar··on Dark’s new backend will be in F#
I had a similar experience. The syntax formatter is https://github.com/fsprojects/fantomas, and there are some controls, though not as many as I'd like.
pbiggar··on Dark’s new backend will be in F#
No one was more surprised than me!
pbiggar··on Dark’s new backend will be in F#
HN needs a like button!
pbiggar··on Dark’s new backend will be in F#
A little over a million. While that should last 5 years, I expect to hire again in the new year, once I have an actual plan figured out.
pbiggar··on Dark’s new backend will be in F#
Ah yeah, that confused a lot of people. It's just me at the moment. Planning to scale up a little bit next year once I have a solid plan of how to get to product-market fit.
pbiggar··on Dark’s new backend will be in F#
Thanks! I think?
pbiggar··on Dark’s new backend will be in F#
Really wasn't that funny
pbiggar··on Dark’s new backend will be in F#
Good point! I quickly updated the blog post with a blurb. Was not expecting this to be #1 on HN!

> Welcome HN! If this is your first time hearing about Dark, check out the website[1], our What is Dark [2] post, and How Dark deploys in 50ms [3] to understand what we're about. Thanks for checking us out!

[1] https://darklang.com [2] https://blog.darklang.com/what-is-dark/ [3] https://blog.darklang.com/how-dark-deploys-code-in-50ms/

← PreviousPage 13 of 34Next →