Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: What are you working on this week?
8 points by zatkin on Sept 28, 2015 | hide | past | favorite | 30 comments
I haven't seen a thread posted like this before, so I thought I'd ask.



Building Space Invaders in C++/SDL[0] and the entity-component framework it uses. This week I'm working on GUI basics and event handling, so hopefully it starts looking like an actual game.

Also continuing the basic Game Development (Unity3D) course at Coursera[1]

[0]https://bitbucket.org/kennethrapp/spaceinvaders

[1]https://www.coursera.org/learn/game-development


I have a service which allows git to be used to manage DNS. The way this used to work was that you'd host your DNS records in a github/etc repositories, configure a webhook to point to my server, and then your updates would trigger DNS-updates.

Rather than fighting to understand webhooks from the various git-hosting services I've decided that volume is low enough that I should just host the git repositories myself.

I've spend a few evenings preparing for this transition, updating documentation, and reworking my codebase to avoid all use of webhooks.

It's a fun update, which will simplify the service for new users, but keeping both systems in use at the same time makes it a little more complex than I'd like. I guess I need a flag-day in a few months where I drop the old webhook support.


That is very cool. Any chance a company could run that tool on premises? I think there is a market for that.


Given that it ultimately pokes around with Amazon's route53 infrastructure I'm not too sure that there'd be a gain if the site/processor were available in-house.

Even if the manipulation were local, the Amazon API would be remote, and not within a company's control.

Interesting idea though. Most of the backend is portable perl and easy to pull out..


Lots of companies already trust AWS and host various services there. Many of those companies might not trust your git servers to be up and running 24/7 with Amazon-like SLAs.


Thanks for the interesting perspective. I guess when I strip it down I've written a tiny layer of magic to convert a (bastardized)TinyDNS zone-file, or Bind zonefile, into Amazon Route53 update commands.

Selling that as a one-off utility would be hard, but it does currently work well as a hosted service via the webhook integration-layer, and a small amount of git-magic.

I will have a think to see if people would prefer it self-hosted, though nobody has expressed any interest in that previously.


I'm working on solving a problem I've been playing with for a while.

Fundamentally it involves picking a feasible solution from a solution space of (225 choose 45) possible solutions.

That number is obviously too big to brute force randomly, but I was able to get very close to a solution by hand, and so it feels like there ought to be an algorithm to generate a solution. (It doesn't feel like a problem where the solution wouldn't be polynomial time.)

It feels like some kind of branching with aggressive pruning would deliver a solution but I have much I need to learn first to actually develop it.


There was a thread posted 2 hours before you posted this :)

https://news.ycombinator.com/item?id=10291148


Just launched https://crypt0.space/, a new bitcoin & litecoin mining pool. Working on adding MixPanel analytics, & advertising with the right orgs to try and get my first users to use it. Have a 0.027 BTC bounty to get someone to connect a couple miners to it & test it so I know it works.


Opening the url on my phone I get a security warning. Kind of bad marketing...


Ooh, I might check that out! Unfortunately any miners I'd attach wouldn't be anything permanent, or even good as far as hashes/sec.


Finishing Part I of The RSpec Book. The hands-on tutorial is giving me a feel for why BDD's advocates advocate BDD. It also giving me a feel for the cost of another layer of tooling on top of the "running code".

Continuing: Coursera Modeling Discreet Optimization: I wish MiniZinc had better tooling and documentation.


Just getting out the first HVSImage 2016 behavioral tracking system for Mercer University. The original machine was slightly slow in tracking and real time analysing due to the i5 chip being a little below par - now I'm slightly slow in testing due to being coshed in Geary prior to flying back on Saturday. . .


I'm drifting, dunno if that really counts. Trying to get rid of some bugs of WTF level that I only see in my own machine and that I can't really use to justify my paid time, so I'm getting stressed out by the minute (after a few days already) while I try to figure it out and go back to regular work.


Very vaguely, what kinds of bugs? I'm curious.

Can you whole-world clone the entire stack you're working on/with?

1. Verify the bug exists in your whole-world clone

2. Chop bits out of one of the copies until the bug goes away, and/or see if you can copy the whole world to another machine (which you've probably already tried)

Alternatively, perhaps let fixing this bug go, and as was suggested, use another machine (or a VM, if that's fast enough and it "fixes" it) for now. You might figure it out while you work on something else.

This position is horrid to be in :P but letting go right now in this situation is sometimes the least long-term damaging solution.


Maybe work on a different machine for now? May be a lot cheaper.


School! And robotics team stuff.

I'm currently polishing a Slack chatbot for engineering-notebook-chatops, working on some boards in Eagle for a universal, multicell NiMH charger, and also hacking on my fork of ReplicatorG (I'm trying to clean the code up, it's a bit WTF-y right now).


I've been getting some landing pages and ad copy ready for when the Instagram ads go public, which I was told is tomorrow.

Also, I am working on adding a test suite to one of my previous applications that I made. I finished up the model/controller tests, so am on to the integration tests.


Manually wrangling recent nuclear eng research docs to get and homogenise data for my own research protocol. Hopefully the protocol will become the template for the work of a remote network / open community of independent nuclear researchers in the months to come.


I'd love to build a SAAS so now I'm trying to figure out what type I'd like to create. Also doing research on everything involved from what tools to use to lessons learned from those before.


I'm trying to learn Rust. The Getting Started documentation on https://rust-lang.org has been pretty interested so far.


Coursera cryptocurrency tech course, https://www.coursera.org/course/bitcointech



Building www.heynay.com and you should definetly check it out and critique it if you have time :)


Heynay looks very interesting and promising. Very attractive design.

In terms of critiques, I know this seems like a petty critique, but it's important: make sure to check your spelling and/or improve your spelling in English. The word is spelled "definitely", for example. Since you're representing Heynay, it's important to give a good impression, and I can tell you that major spelling errors will give a poor impression to the majority of the population, particularly among exactly the people you want to impress.

Again, it's a petty criticism, but the small details like spelling are well worth paying attention to, as they serve as useful signals to many people.


Trying to map census data for my city using Python to retrieve the data and R to make the map.


Applying for TC Disrupt London and trying to plan next steps for a travel app I am working on


fixing up the javascript to my mobile web app that focuses on food. I have to try to find beta users who like to eat out as my next task.


Working on a personal project built with rails and reactjs :-)


Learning Elm!




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: