This is the first time I am mentioning this publicly, but I have been working on Educator News. It's a clone of HN for the education profession.
It's been attempted before, and was well received. The people who made it previously were not programmers, so when it immediately outgrew the platform on which it was built, it disappeared. I think I can do it right. :)
I am planning a soft launch this weekend. After working out any immediate bugs, I will do a Show HN.
No, I'm building it in Django. I have been looking for a well-defined project to build, and this ended up being a perfect project. It's inspired by HN, and to everyone here it's a clone of HN, but most educators who end up using the site will probably never know HN exists.
It's a completely open project, and if things go well I'm going to write a Django tutorial based on building EN. I'm going to write the tutorial as part of another project I started this year, http://introtopython.org. It's been a busy year. :)
Automated NBA game recaps with more context than your typical boxscore.
Note: I am not a web developer, so this project has been mainly to teach myself a new technology (Rails, primarily). It is very much a work in progress. Also, clearly, I am not a designer and am not really sweating that for now.
Thanks for the feedback! It's running on a single droplet on DigitalOcean, so my guess is it can't handle even the slight amount of traffic I'm getting from this HN post. Beyond server-side caching, I think my code is fairly non-optimal.
If anyone has any tips for serving up a site like this with limited resources, feel free to teach me.
I don't know if you're aware, but the community section of Digital Ocean has some informative articles on stuff like this, there was actually one posted today about scaling Ruby on Rails applications.
Thanks for the link, I will definitely take a look at that. I have found some of those articles to be useful, although external blog posts and gists have proved just as valuable. One recent accomplishment was automating deployment to a single command with Capistrano. Took me a lot longer to set that up than it probably should have, but feels great now :)
Scraping. There are a ton of sources for NBA, be it nba.com/stats, ESPN, CNNSI, etc.
Many of these sites serve up their play-by-play on a JSON feed, which makes it even easier. However, I must warn you, the data is not as clean and robust as you may like. A large part of my work was implementing a cleaning algorithm to infer missing data.
I made a 3D flappy-bird inspired game: "Droppyship".
https://itunes.apple.com/app/droppyship/id820172550
Finishing up a quick update this weekend to add gamecenter and ios7 controller support, but I don't know if I'll take it much further.
LottoLane is a lottery pool management tool. It will handle email invites/reminders, payment tracking, sharing pics of tickets, and more. It will take care of everything involved in a lottery pool except for buying tickets.
that's funny 'cause i would actually play the lotto (esp. those scratch off games) if buying a ticket was as easy as clicking a button, or sending an email. A daily email with the results of the lotto would be great too.
Legality of doing something like this is questionable. Plus, there's the issue of trusting the site if the ticket hits the jackpot. I believe they have people that physically go to convenience stores to buy the tickets for you.
I do have some ideas of how to facilitate playing in pools without having to be geographically close to the pool manager. But, overall, the pool manager should be someone you know.
I'm downloading all of Project Gutenberg and gonna run a markov chain text generator on it and see what works of genius it produces (inspired by the King James Programming post yesterday).
I think it'd be cool to have a site that lets people input sources and then it displays output from the generator. Need to do some more reading on markov models and see what improvements I can make to Barrucadu's code
A twitter clone with a fully nonblocking Play framework/Redis backend and a angularjs frontend that communicate (almost) exclusively by websocket. It's nothing that hasn't been done before, but it's a chance to play around with some new tools.
I am working on a (to be, hopefully) open source support ticket system based on Laravel and Bootstrap (So easily themeable). The primary feature is that it works with, not against other ticketing systems. For example I can 'push' a ticket to github or bitbucket. I also want to try and integrate email, so that sending an email to a support mailbox will automatically create a ticket. It would be nice to also have an API for web sites to use, so support tickets will also be created on errors on the production servers.
Just in the initial planning stages but it would work well in my workflow.
- video overview: http://www.youtube.com/watch?v=yvjeIZgykiA
- early web UI: http://www.youtube.com/watch?v=yvjeIZgykiA#t=24m5s
It's written in a combination of React and Closure. It's a single page app that doesn't care about IE<9. I get to use workers, blobs, transforms, svg, flexbox, typed arrays...
Oh, and everything in the UI live-updates automatically as a side-effect of using React + Web Sockets.
the current task: refactoring my code to incorporate new things I've learned, namely: TDD and knockoutjs + json API. After 2 years of self learning python, I'm looking to transition from teaching into development and am hoping to use this project to showcase what I've learned!
I'm playing around with creating feeds from metadata - specifically, building feeds with 'expanding' views which build trees from the outbound links on pages. The reason the url is appended there is the thing doesn't even have any caching - but just add a url to the form at the top or tack it on like so and it should build the feed out of whatever.
The [+] links should expand the feed. Unless they don't. Because it sometimes doesn't. When it does though it's nice.
Also trying to get dynamic rss to work but unless the dynos at pagodabox are really spinning this can be insanely slow:
Designing and prototyping an elegant bookmark manager. I find existing solutions clumsy in their functionality and ugly on top of it. The inspiration comes from the design world. Think about the look of table of contents in a beautiful art book or signage and navigation in a modern library building. I would like to have my hundreds of links presented with minimal elegance and modern information design functionality...
Building a media management and streaming server for my local network, in javascript with node.js, along with an android app, so I can organise my music and listen to it at home from my phone :)
All these components exist, but its an exercise in software design, as well as learning both about Android and Javascript.
Well, I'd done a small kernel for a class. Later, I wrote a lockfree malloc and realized that large lockfree programs are actually pretty managable, and that was the start. It's generally just a matter of designing around simple data structures (which are the only ones you can really do lockfree) and a repetition of some key tricks like refcounting, generation counting, and type-stable memory. Writing the data structures and primitives was tough, but otherwise you don't have to worry about locking and things come out quite neatly as a result.
It's been attempted before, and was well received. The people who made it previously were not programmers, so when it immediately outgrew the platform on which it was built, it disappeared. I think I can do it right. :)
I am planning a soft launch this weekend. After working out any immediate bugs, I will do a Show HN.