I had an idea.
I wrote it down. (repeat 86x – based on Evernote usage since August ’09)
I had an idea.
I wrote it down.
I talked to a lot of people about it. (repeat ~30x)
I had an idea.
I wrote it down.
I did a basic landing page, or mockups. (repeat ~20x)
I had an idea.
I wrote it down.
I did a basic landing page, or mockups.
I told everyone I know. (repeat ~3x)
I had an idea.
I wrote it down.
I built a prototype, 90% of the way there. (repeat ~20x)
I had an idea.
I wrote it down.
I built a prototype.
I finished it.
I opened it up for the world to see. (repeat ~5x)
I had an idea.
I wrote it down.
I built a landing page.
I told everyone I know.
I built a prototype.
I told everyone I know.
I pitched it all the time. (repeat ~1x)
I had an idea.
I wrote it down.
I built a prototype.
I told everyone I know.
I pitched it everywhere I could.
I assembled a team.
I wrangled advisors.
I dropped all other professional distractions.
I got us accepted into an incubator. (in progress)
When Mike IMed me a few months ago to ask if I would join him and Kunmi for StartupXLR8R, I resisted at first. I am already focusing so much time on STRUCTO, and already distracting myself with side projects like WhoMails.Me and ProudlyMadeinDC (the latter of which Mike is a co-founder), that another project would be too much to handle. But we approached it with a different spin. While we all had our own startups, and all the other teams applying were for the most part well established startups, we were going to go from zero to launch. In two days (16 hours).
We had tossed around a number ideas, but ended up choosing a smaller idea that we knew we could launch an initial version of by 4 PM Sunday afternoon, when it was time to pitch our product for “investors” (not that we had any interest in raising capital for this). We settled on an idea I had, a daily e-mail customized for you.
So Saturday morning, we headed off into our breakout room (thanks Microsoft for an absolutely gorgeous space), and just started cranking out code. Kunmi worked on the frontend display, Mike on the email and configuration backend, and myself on all the data sources. It was an exciting weekend, with sixteen straight hours of heads down, don’t-talk-to-us coding.
By 3:59 PM (we were pushing code as we were walking into the conference room), HeyAstro was born. In the few days since we’ve launched, we’ve had over 150 people sign up, an immense amount of amazing feedback, and a laundry list of new things we want to do with the product.
Sign up for HeyAstro, and let us know what you think!
Final Stack:
Ruby on Rails (#FTW)
Heroku for Hosting
Delayed Job for Queue Management
SendGrid for email delivery and analytics
Gems, Gems, Gems. There is no way we could have built this in any other language than Ruby, as so much of the backend relied on particular gems. Koala for Facebook, HttParty, Instagram, OAuth etc…
In February of this year I presented at the DCPHP meetup. Based on my experience deploying large client applications built in PHP using AWS and Rightscale, I gave a detailed walkthrough of how to set it up. Fuzzy video below:
Structo was born out of a pain I saw as a web developer, and, in speaking with many other web developers, I found I wasn’t alone. There are many components in a web application that have to be repeated every single time.
What is Structo?
Structo is a hosted RESTful database. You specify your schema in a drag and drop interface, and then have a full infrastructure for your web application. You get to focus on what’s really important in your web app – we’ll handle the rest.
I had a need for switching back and forth between two different values in a hidden input field. I couldn’t find any decent plugins to achieve this, so I wrote a little jQuery plugin to handle this.
Download here. Using it is dead simple, just pass in the two values you want to toggle between.
But first things first… in order to make it even easier for myself and for other developers, I built out a ruby gem that interacts with the Open311 and Geolocation API. I know I’m helping out my competition, but it’s all for the better good, right? Here you go:
Download and install it, and you should be set. It’s fully tested using rspec, so just open up the spec folder and you’ll be able to exavtly how it works. I’ll eventually add this to rubyforge, but for now you can take the extra step
DC OCTO has a ways to go. The API is supposed to launch on July 1st, but they certainly have a lot more to do with the API, both in terms of development and documentation. The Geolocation API seems to be pretty solid. The 311 API has unimplemented (stubbed) functions, spelling mistakes in variable names, no error handling, and no documentation (the “submit” function, the most important, is a total mystery). It’s too bad that people will find this when they head to the CodeJam over the weekend, but oh well.
Like so many things from my college experience, I never realized how amazing this was until years later.
Rather than taking a “usual” internship at a government agency or consulting company for the summer, I instead chose to stay on campus for the summer, and in addition to a lot of relaxing, worked at two amazing labs at UMD. One was the Center for Satellite and Hybrid Communication Networks, where I got to play around with awesome gadgets like motion tracking cameras and sensor networks. The other position, which I kept throughout the following school year, was as a researcher in the Human Computer Interaction Lab. I was tasked with working on an Java interface (using Swing) for the MALACH project, which, in a nutshell, was to allow people to browse and view a massive amount of transcripts of recordings of Holocaust victims. The recordings originated from the Shoah foundation, whose mission is to record and store all memories of survivors of the Holocaust (side note: my grandmother volunteered as one of the interviewers). So I spent the summer working under Ryen White, who now works at Microsoft Research.
As we were presented with developing an entirely new interface, we had a number of challenges. We came up with some pretty cool solutions. One thing I came up with was an interface element nicknamed PygmyBrowse. PygmyBrowse is a rather simple and easy method of bi-directional navigation of infinitely complex trees in a compact environment. I thought nothing of it at the time, but the professors around me disagreed, and suggested I pursue it further. After spending a good chunk of the semester holding user trials (in between running one of the largest student groups on campus, being actively involved in my computer, and oh yeah… double major), we ended up with an academic paper, with yours truly as the primary author. Which ended up getting accepted to a major human-computer interaction conference. Whiiich I didn’t attend, because there was one or another extra-curricular event going on (kicking myself 3 years later).
When I first made the jump from a big consulting company to a tiny creative agency, I was worried I would be spending most of my time doing small boring web projects. Yeah, right.
As if I the C-SPAN ConventionHubs that I wrote on earlier weren’t successful enough, we at JESS3 teamed up with C-SPAN and New Media Strategies again to launch the Debate Hub. I served as lead developer. Not only does this have the same great features as the Convention Hubs, like embeddable video, Twitter coverage, and blog content from all over the web, we added in some really killer features.
The Timeline is one of the best features. We’re pulling in transcripts as the debate progresses, and have it segmented out by speaker. Click on a piece of the timeline, and magic happens. This is built off of MIT’s SIMILE project.
The Transcript Treemap is another awesome feature. It shows the most used terms of each candidate, based on the transcripts, along with sparklines (more appropriately, the jQuery version). I even managed, thanks to some more jQuery magic, to allow you to dynamically filter what debates/candidates to show. Of particular importance to me, as the Treemap was created at my alma-mater’sHuman Computer Interaction Lab, where I was a researcher for a short period of time (more on what I did there later).
It’s received a ton of press coverage, including ZDNet, Ars Technica, TechCrunch, RedState, and a number of other sites. It’s been a wild ride.
I’m not a very politically involved person. For me, I make sure I vote when possible, care about the issues I choose to care about, but never get more involved. However, I like to think that, when working on projects like these, we’re giving more people access to more information than they previously had, and allowing them to make more educated decisions when it comes to electing their leaders. It may only make a small difference for a small group of people, but that’s all that matters.
Hard to believe I graduated college two years ago…
I'm a hacker/founder, I'm working on Contactually, an email interface to CRMs. I'm also a freelance developer, working as skeevisArts while working on Structo, a cloud service for web developers.
When 
When I first made the jump from a big consulting company to a tiny creative agency, I was worried I would be spending most of my time doing small boring web projects. Yeah, right.
