Ask HN: Share your projects and the stories behind them

https://github.com/imperialwicket/hubot-asgard

I think the chatbot mentality is a good one, as it forces you to script/automate things so that a chatbot can accomplish them. That level of automation is a good target imo. It's also a great way to be certain that you can solve many issues remotely - since you can generally provide a meaningful chatbot command via mobile device.

In terms of scaling and cloud (aws) management, Asgard is awesome (seriously - if you're on aws and haven't checked it out, go now: https://github.com/Netflix/asgard/wiki). The biggest complaint I have about Asgard as a tool, is it feels like more than most organizations need; because it is. My goal with the hubot-asgard scripts is to wrap some of the more essential features, and hide the bulk of the solution that many orgs just see as excess tooling.

There's also the Cloud Prize...

If anyone is interested and needs assistance or wants particular features that aren't wrapped yet (it's still very young) I'd love to know. And of course - all feedback is welcome and appreciated.

Show HN is the suggested way, but a monthly startup roundup would be interesting. Show HN does not bring enough attention somehow recently.

It was inspired by a trip to Japan last year[2] after having been away for a while and seeing how badly my reading comprehension had deteriorated.

[1] http://www.macaronics.com/

[2] http://denis.papathanasiou.org/2012/06/10/yokaben-read-write...

http://sideprojectors.com

I thought it would be great to not let many of the interesting side projects done by developers, entrepreneurs be abandoned like so many, but let them find home in the hands of others. It's early days and we are learning and figuring out our ways.

Its open-sourced at github.com/sdslabs/leaderboard and uses github organizations for authentication (anyone in your org can login).

http://www.placeunit.com

http://AppStore.com/placeunit

It's an app that lets anyone build a mini-responsive website entirely from their iPhone/iPad.

I knew someone who needed a new site. So instead of giving him a fish, I got him a fishing rod.. :)

[1] http://pureflickr.com

[2] https://itunes.apple.com/us/app/pure-for-flickr/id660272928?...

[3] https://itunes.apple.com/us/app/potretr-flickr-photos-browse...

http://cardflashapp.com

It started with me wanting to learn Chinese characters and being annoyed at Anki's UI/UX (Oddly I also "accidentally" learned all US state capitals and World capitals while debugging the thing).

It's Meteor + Heroku + AWS S3 (user card review data via client-side signed uploads)

I built it because I was tired of manually mixing together ambient sound to block out background noise when I work in public spaces.

I'll quote from my comment in my unsuccessful Show HN:

---

This is the first thing I've ever 'coded' (if you can call it that), so it's pretty rough around the edges. It's pretty similar to RainyMood [1] and Coffitivity [2] but with the added feature that you can play more than one type of sound at the same time and adjust the volume of each to find a mix that suits you. I was inspired to make this because that's something I ended up doing manually quite a bit.

It's HTML5 with jquery and jquery-mobile, mainly for the slider elements (which I didn't know how to make myself). I wish I could have avoided using jquery-mobile because it was quite a headache trying to deactivate a bunch of the default features and just use the slider.

It uses the standary HTML5 audio player. When the play button is pressed, all the audio streams play. When the value of each slider is changed, the volume for that player is updated with JS.

iOS devices don't support changing the volume via javascript in any fashion, so this won't work on any iOS device. I don't see a way of getting around that. It should work in most other modern browsers. Seems to work in IE10, FF, Chrome and the stock Android browser.

The only slightly tricky thing was getting seamless looping for the audio samples. The HTML5 audio player has a "loop" setting, but in every browser I tried it in there was an audible gap between the end and start of playback. Very annoying in this kind of application. To get around that, each audio track fades in and out at the beginning and end. When the player reaches the beginning of the "fade out" section of the main track, about 10 seconds before the end, it triggers a 25s "glue" track to start playing, which fades in as the main track fades out. When the glue track begins to fade out, it triggers the main track to fade in again at the beginning. The end result is something approaching a seamless loop with constant volume. I think it works quite well for ambeint noise, but it obviously wouldn't work for anything with a beat.

The sounds are all CC licensed samples from freesound.org [3][4] that I did my best to mix and clean up. In the unlikely event this ever became popular, I'd love to make some field recordings myself to get a bigger range of sounds in higher quality, but that's a bit of a pipe dream.

    [1] http://www.rainymood.com/
    [2] http://coffitivity.com/
    [3] http://www.freesound.org/
    [4] http://asoftmurmur.com/info.html

If I have time and I find anyone who wants to use it, I hope in the future to add more sounds, better initial presets, saving presets, stop-playing-after-x-minutes, fade-out-over-x-minutes and support for (10+2)*5 [5]

    [5] http://www.43folders.com/2005/10/11/procrastination-hack-1025

I had to crawl the search engine to restore some of the pages, which led to the free service www.recovermywebsite.com.

I don't have much time to develop this further, so if anyone are interested in buying it, contact me at support@recovermywebsite.com.

I tell my wife I'm obsessed with models.

A tool for seeing the relationship between chord & scales on the guitar. I built this a few years back to sharpen my Silverlight skills. (stop chuckling). Just rewrote it in HTML & JS, with help from teoria.js.

We really want this tool to make it much easier/quicker to let devs release and launch their own SAAS products, concentrating on their unique functionality and not waste their time with the boring stuff.

http://www.bugscore.com/ and let me know your feedback & suggestions.

You can use Bugscore to score a product you have used or seen, a business you have worked at or one that has served you and finally, people you have met or seen. It's a platform for you to express your opinion on almost anything on Earth! Similarly, you can see what others think about products. The same goes for businesses (including schools and colleges).

Why I built it?

My wife was struggling with pronouncing certain words while learning English. So I built a webapp for her to practise pronunciation by herself.

Later on her classmates also started using it, now it has quite a few users.

Let me know if you are trying to learn English as well, I will bump your account so you can use it indefinitely.

I built it because I was tired of getting in arguments with my family over what words were acceptable. You can use it to check a word or insert wild card characters and tile delimiters to find the best words. I was unimpressed with how unfriendly existing tools were so I built my own in 14 hours. If you like scrabble you should check it out!

I've only sold around $1,000 of licenses so far, so maybe this post will help...

I don't have plans to commercialise it, and keeping in mind the costs to run it are low (less than $30 per month) I'm happily resisting the temptation to place advertisements all over it.

It's been live for about a year and a half, in the tens of unique visitors per day (hundreds of page views, 5+ minute avg session times) and certainly keen to get it more publicised.

I thoroughly enjoyed making the application (and enjoy supporting it still), and learning about all the cool tools/tech most people on HN get to use in their jobs full time!

Feedback welcome - and if you have an aquarium be sure to keep doing you water changes, your fish will thank you for the effort ;)

EDIT: fixed url

http://github.com/jbaiter/spreads

http://spreads.readthedocs.org

With it, going from a physical book to a digitized, postprocessed and OCRed PDF takes around 30-40 minutes for an average 300-400pg book with no pictures/illustrations.

My (rather lofty, I ad mit) motivation was to make it as easy as possible for people to free printed information from their pyhsical shackles and enable them to share it (I'm currently working on a plugin that allows the user to directly upload a scanned book to the Internet Archive).

The inspiration for it came after I purchased a kit for Daniel Reeds' incredible DIYBookScanner[1] and built it on a free weekend. Upon toying with it, I realized that, while there was a lot of great software available for helping with the scanning and postprocessing, using it required a lot of often tedious manual interactions that could easily be automated.

The tool basically handles all communication with the capture devices (cameras, but the code is kept rather general as to allow for the usage of mobile phones or flatbed scanners) and calls a bunch of 3rd party applications to deal with postprocessing and output generation.

I learned a lot about multithreading/multiprocessing in Python, got to brush up my PySide knowledge and am currently learning AngularJS by developing a webinterface for it, to allow the headless control of Raspberry Pi-connected scanners.

[1] http://www.diybookscanner.org/

I mentor a lot of startups in Boston. Between the Harvard Innovation Lab, CIC, Lean Startup Machine, Youth Cities and others, I've heard a lot about what problems early tech startups face. A big one I see with a lot of startups is that they have some idea that requires a relatively simple or at least well known algorithm, but they have neither the expertise nor infrastructure to implement it themselves or even install Mahout.

In an effort to help these companies, I created a service to help startups run algorithms without having to worry about the details. You just choose the algorithm you need and build a data model appropriate to your project, and then Algorithmic.ly generates an API for you to interact with your model and continuously runs the algorithm you need on your data. Then all you need to do is query Algorithmic.ly for the algorithm results whenever you need them.

This makes everything from spatial search to netflix-style recommendations within the reach of small startups who only have web developers or iOS developers and can't afford hiring data scientists to do it for them. Right now we're in a limited beta with several startups as we figure out how to scale, but we're looking for more companies we can help, especially in the Boston startup ecosystem.

It's been getting pretty good traction thanks to decently high Google search rankings for API related queries. A decent amount of large companies have submitted their APIs or requested to claim their page.

I haven't been able to spend a lot of time on it recently to add features but I have a lot of cool stuff in mind.