Cloud Camp Seattle

Leave a reply

I spent most of the day on Saturday in Amazon’s HQ at “Cloudcamp,” an unconferece focused on cloud computing.  Microsoft, Amazon, RightScale and Rackspace/Mosso/Slicehost were the sponsors, but their presentations were enough to get a taste of their offerings while remaining mercifully brief.

I’d hoped to get an idea of how people were approaching scaling out websites using cloud services, how they were navigating the risk of lockin to something like SimpleDB against the difficulty of building their own scalable services and deploying them on something like EC2. I even led a well-attended session on the topic, but in the end, it was clear that cloud computing, is, at least for most of the attendees, something they are still scoping out.  A session on “success stories,” ended up being more about aspirations than actualization.

This disconnect between my expecations and the reality of the conference was itself pretty informative.

Last night I pounded out a long blog post with more of my impressions using my iPhone, but rather than posting it all in one monolithic post, I’m going to break it out into a bunch of smaller posts that will follow this one over the next few days.

Recovery.gov May Be Just The Beginning for Open Source in Federal IT

1 Reply

Techmeme wasn’t exactly buzzing with the news that the new Recovery.gov website, which provides information on federal recovery programs, was built using the open source Drupal community content management system.  It made it on to Techmeme, but just barely; only one of the linked sites noted the Drupal connection, the others, including some prominent tech news sites, were only covering the launch of the site.

On the other hand, the Drupal community, including one of the Drupal creators, thought it was pretty damn cool. I also thought it was a pretty big deal for lots of reasons, perhaps formost among them, the site was apparently created quickly by a small team, quite in contrast to other federal IT programs where a court of well connected contractors charge millions, or sometimes billions, for IT projects that seem to fail as often as they succeed.

This anonymous comment on Reddit by someone claiming to have an insiders view of the Whitehouse’s IT of Recovery.gov illustrates how big a deal this is:

i’m on an anonymous acct b/c there may be one or two things i say here that aren’t entirely ready for public disclosure.

the white house is tied to an infrastructure that the bush administration bought shortly before leaving office from a major contractor for millions of dollars (public information based on contracts)

there are high level internal discussions about actually ditching this for drupal to run WH.gov altogether although there are a lot of opponents that come in two camps:

1) technologists that are saying use rails, django, or something else as drupal won’t be secure enough or flexible enough etc etc etc

2) the old guard that is afraid of anything that doesn’t come with a $10 million dollar contract (the existing WH contract is something like 20 mil — for a f*cking website)

needless to say moving from the tools had before the campaign to the ones inside the WH is a step in the wrong direction. recovery.gov being based on drupal is a positive step (technical squabbles with drupal aside) as it is a bucking of the idea that for anything to work it has to be ‘enterprise’

i come from a tech background and despise PHP and know the baggage that drupal or any “CMS” carries, but trust me, the white house launching a .gov domain without paying Boeing (yes.. boeing and other “military” contractors win website contracts over established firms b/c they know how to play the gov’t contracting game) is huge if only because it just saved millions of dollars of taxpayer money (and from a technical angle if you think drupal is bad imagine what kind of CMS Lockheed Martin or General Dynamics could cook up)

people inside the white house are talking to other very smart people in camp 1 and know not to listen to camp 2

Go back and read that again.  Whitehouse.gov was a fricking $20M website.

I’m very cautious in my optimism about the Obama administration, but things like this make me think this really is “change I can believe in.”

Predictors of Facebook Engagement

2 Replies

This report from a panel discussion at a user group for R, an open source statistics package, shines a little light on what factors Facebook found helped predict a new users initial and ongoing engagement with their site.

Itamar conveyed how Facebook’s Data Team used R in 2007 to answer two questions about new users: (i) which data points predict whether a user will stay? and (ii) if they stay, which data points predict how active they’ll be after three months?

For the first question, Itamar’s team used recursive partitioning (via the rpart package) to infer that just two data points are significantly predictive of whether a user remains on Facebook: (i) having more than one session as a new user, and (ii) entering basic profile information.

For the second question, they fit the data to a logistic model using a least angle regression approach (via the lars package), and found that activity at three months was predicted by variables related to three classes of behavior: (i) how often a user was reached out to by others, (ii) frequency of third party application use, and (iii) what Itamar termed “receptiveness” — related to how forthcoming a user was on the site.

via How Google and Facebook are using R : Data Evolution.

Drupal: from a Dutch Dorm to Delivering Democracy in Less Than a Decade

Leave a reply

I think it would be a very interesting piece of journalism to follow Drupal’s origins in a Dutch dorm in 2000, through it’s use in the Dean and Obama campaigns to it’s use on recovery.gov.

It should provide a great lens for examining the rise of OpenSource software post the dotcom crash, the blossoming of social software, the evolution of political organizing and fundraising in the Internet-era, the role of agile development in government IT, and the role of the Internet in reshaping and opening government. Plus, it started in a dorm, so I’m sure there is some sex, drugs, and rock and roll in there too.

Webfaction beats Pair.com

Leave a reply

It’s going to take a while to cut-over, but I just flipped the switch on moving this blog from being hosted on pair.com, to being hosted on webfaction.  This comes after spending a bunch of my weekend one December trying to optimize things on my old host.

I already see a big performance improvement. Webfaction’s server finishes processing a page in about the same amount of time as pair’s server spends just loading up WordPress. Most of this is down to the fact that webfaction uses eaccelerator, a PHP opcode cache so that it doesn’t have to spend as much time parsing the wordpress code on every view.

Webfaction also has their webserver set up to automatically compress responses so they download quicker. I’d gotten this going on Pair, but it required some plugins for dynamic pages and some rewrite rules and scripts to serve precompressed javascript and stylesheets. The new arrangement is simpler. There are fewer things to break and fewer steps required when I install an update.

Pair has been really reliable for what is closing in on a decade. We’ll see how Webfaction does, but they have a good reputation, and they did well with a little experimental site I tried a year ago.

I’m going to see how this goes for a few days, if there are no red flags, I’ll move our other blogs over. Then I have to figure out what to do with Pair, they just automatically renewed my service for another year at the beginning of the month.