The Most or The Best
Open Source, Teamin' Up, The Stack, and Looking Forward
Shirts and Shipping
Upcoming.org: The Shirts
Final 24 Hours!
For backers only. If you're a backer of this project, please log in to read this post.
In case you missed it, the Upcoming.org homepage was updated seconds after the Kickstarter project ended on Friday:
For those who didn't recognize it, or clicked through before the music started playing, the original music on the page for the duration of the Kickstarter project was a MIDI version of Elton John's "Don't Let the Sun Go Down on Me." Naturally, I had to change it once the project ended.
And don't worry, the new Upcoming won't autoplay music. (Probably.)
Did you know modern browsers don't support playing MIDI files? With all the advancements in HTML 5 Audio, I still had to convert the MIDI tracks to MP3 to get them to work, ballooning their filesize to over 6000%. So inefficient, but those are the sacrifices you make for that old-school sound.
Okay, the shirt designs are finalized and ready to print. All of you who backed the project for $75 or more get your choice of two designs:
1. The Retro Jersey. Printed on Champion raglan baseball jerseys with three-quarter sleeves, these are perfect reproductions of the first Upcoming.org t-shirt from 2005, minus the Yahoo logos. These are unisex sizes, available from small to 3XL. A sizing chart, with body length/width and sleeve length, is available here.
2. The Modern Minimalist. I commissioned a new hand-lettered version of the Upcoming logo from Clarke Harris, a promising young illustrator in the process of moving to Portland. These will be printed grey on black, on American Apparel's ridiculously soft tri-blend, the same fabric we use for all our XOXO shirts. They're available in men's sizes from small to 2XL and women's sizes from small to XL. American Apparel tends to run small, so please check your size.
Both of these shirts are only available to backers, putting you in an elite group of certified hardcore badasses. If you see anyone else wearing these shirts, you'll know you share something in common, and you've just found a new best friend. Maybe even marriage material.
By the end of this week, you'll receive a Kickstarter survey asking for your shirt pick, size, and shipping information. These shirts are being printed for each one of you, so please double-check your requested shirt size. If the size is wrong, I likely won't have any additional extras to replace it.
You'll be able to change your shipping information on Kickstarter at any point until I send the final confirmation, which will be later this month once the shirts have been printed.
IT'S ALL HAPPENING, PEOPLE
We made it to the last 24 hours, within spitting distance of $100,000 raised! (We went over the $100k briefly on Tuesday night, but dipped back down when a $5,000 backer pulled their pledge. They wanted to advertise their marketing analytics company to you, and I said it wasn't a great fit.)
The Kickoff Party
First things first—if you're in the Portland, Oregon area, please join us for drinks tomorrow as we count down the project's end, and celebrate the kickoff for Upcoming's relaunch. The Kickstarter project ends at 3pm PST, and we'll be on the large outdoor patio at Green Dragon for the countdown, arriving around 2:30pm. I expect we'll be there for several hours, so feel free to drop by after work.
As predicted, there's been a small flurry of media about the project over the last two days. The Verge gave Upcoming the feature treatment, a wonderful profile by Casey Newton about "the brilliant life, stupid death, and improbable return of Upcoming.org." It hit the coveted Top Story slot on The Verge this morning, and has stayed at the top of their homepage all day.
Moments ago, Engadget published their own feature by Nicole Lee, with some stats from former Upcoming PM Vince Maniago and thoughts from Laughing Squid's Scott Beale about how Facebook Events' semi-private state never made it a viable Upcoming replacement.
For the few of you who aren't sick of hearing my voice yet, I did an hour-long interview with Myke Hurley for the Cmd-Space podcast on 5by5.
Finally, I wrote a long, personal story on Medium about "selling out" online communities, and mending past mistakes. It was difficult to write, but I'm happy I did. It includes this sweet GIF I created from Wayne's World. Party on.
Restoring the Archives
I've been working with members of Archive Team to restore the original Upcoming data from its current state—a staggering 3.5 terabytes of compressed HTML, CSS, JSON, and images scraped from the site shortly before it was shut down.
For those who've never heard of them, Archive Team is a group of volunteers working to save websites and other digital material in danger of disappearing from the Web entirely. They were responsible for saving most of Geocities before it was shut down, later releasing it as a massive 640GB torrent, and many other prominent sites that have since been deleted.
They led an effort to back up Upcoming with a decentralized tool called Warrior, a virtual machine that anyone can install and run to help preserve dying sites. It takes tasks from a central server, retrieves a set of files from a web server, and sends them back for collection.
In three days, everyone running the Warrior application collected every event from Upcoming and Archive Team uploaded everything to the Internet Archive.
The data is preserved in a format called WARC, a format designed by the Internet Archive specifically for storing large-scale data from web crawls, supporting any number of binary types and all relevant metadata.
From these WARC files, and the right tools, a site can be reconstructed to its original state. There are a number of tools built upon the standard, like a one-click GUI for preserving and playing back archived sites, a Chrome plugin for creating WARCs from any page, and proxies for viewing their contents.
To reconstruct the archives, I'm using an incredibly useful tool called Warcat, built by Christopher Foo. Warcat expands and parses WARC metadata to reconstruct every saved URL as files on disk, sorted into directories based on the original archived URL. http://example.com/path/test.html will be saved on disk as test.html in a new subdirectory called "path" in a directory called "example.com".
Warcat overwrites the file every time, leaving a single copy of every file referenced across all of Upcoming.
The drawback is that this process takes time, which is why I started automating it now. Currently, it takes around 40 straight hours to process the 10 million records found in each 25GB piece of the Upcoming backup. Two archives have been processed so far, with 140 to go. At this rate, it would take seven straight months to extract the full 3.5TB archives from a single server.
Christopher's looking at ways of optimizing warcat's performance, and worst case, I could split up the tasks among multiple servers, though moving that much data isn't fun.
I knew this would take time, which is why I started automating the process now. While it's lower priority than working on the new Upcoming, I want the archival process to run in tandem while I do design and development on the new site.
So far, 211,090 legacy events have been extracted using Warcat. The oldest is event #200, a Goldfrapp show in Seattle from October 2003. The most recent is #9,995,523, an October 2012 event offering free haircuts for breast cancer survivors at a Tucson, Arizona JC Penney's. (What a difference a decade makes.)
If you're curious about what the process looks like, this is a constant stream on my desktop.
See you on the other side, everyone. And thanks again!