I'll be live blogging some of the event here if folks have questions or want to follow along.
The hashtag for the event is #codingforward so if you search that on Twitter you will find what folks are tweeting.
Folks just streaming in. Not sure if there is a live video feed, but there is a camera.
: Their challenge was to combine a number of disparate tools.
: They built an API like set of services to help with that but there were certain other challenges in terms of training users etc…
: They said they made a user cry… That was a low point…
: Determination was the key when overcoming the iteration cycle challenges. The demands were tough and yet they started with minimum viable product
: At first there was some outcry for the system to have x or y feature right up front but as they iterated, some of that died off as users realized there maybe wasn't as pressing of a need for the features they demanded…
: The users' demands shifted over time…
: I finding a team they wanted a proven record of shipping code…
: In terms of diversity, it was a lot of what you see on stage, guys with beards. The challenge is creating a good environment – they hired more product people that were more diverse… Good question, it was a hard problem
: How do you deal with 20 hours a day of hard work?
: Keep the end in sight.
: How do you ship code everyday?
: They practice with technology agnosticism.
: Work in what you are productive in.
: Distinctive libraries were kept via the API
joined the Chat
: Regular code freeze and weekly deployments after testing on a staging server…
: There were several surprises in performance hits (unique challenges for their system) you never know what may hit ou fast and must plan for that as much as possible.
: 670million online financial transactions after convention was 2million dollars per hour at small increments.
: In terms of load balancing there were modular components spread over geographically distant servers
: Be smart about design decisions.
: The features could be turned off so that user experience would not be terribly interrupted.
: Forms would submit to the greater API for processing.
: This is a rare set of circumstances in that Election Day is a MUST WORK day and everything else is in prep but also needing to be functionally active.
: Toward the end they curbed new feature development and worked on fixing bugs and tested many failure scenarios. They tried to make it fail to work on problem solving.
: They used a LOT of Amazon space.
AWS now is the way to go
: You can have the best app in the world and if it isn't there the day you need it, it doesn't matter. Good lesson here for every app/website CMS
: Inside tip on AWS – migration from east to west if eastern servers go down gets hard, but if you already own resources on both they don't take it away… So replicate.
: Internal team is what helped them. Vs. outsourced.
: Having the services layer helped to build great tools no new web frameworks just solid proven tech using solid fundamentals.
: Ops and org efficiency can be enormous advantage. Keeping everyone physically close together allowed them to respond effectively.
: Tech savvy needs to be more extant all over – tech literacy…
: In all orgs educate while building…
: Want to do it? Networking orgs. Follow via social media. Ask how you can help. Establish trust.
: They inherited a whole lot of random stuff and part of the hard thing was deeding what to reuse and what to rebuild.
: They used git hub a bunch…
joined the Chat
: Detailed retrospectives at the end of the campaign. What worked. Lessons learned.
: Rogue apps are showing a need and expose how an internal team is not serving well
: Over communicate with your stakeholders.
: About half the engineering team was from Chicago.
: Ok guys. I'm headed offline. New relic says they are hiring.