BarCampBlock - Day 1 Summary
Here are my random musings about the sessions I attended today.
What Should I Read? Coping with RSS Glut
This was a session run by a guy named Jon who works at Tacit Software, which has a product called Illumio, which is a learning and filtering app for RSS feeds. It sounds like a decently cool idea, but pretty much everyone at the table moaned when John says Illumio was implemented in .NET (C# to be exact). And John does deserve some decent props for specifically designing the (desktop) application with privacy in mind.
Some of the interesting concepts of the talk:
- Email had created a culture of guilt in the past, and RSS was supposed to be a possible solution to that… Yet somehow, we seem to be falling into the guilt path again.
- There seems to be several specific archetypes of RSS consumers. (Eg. casual perusing, committed read-it-alls, show me only the "brands"/blogs/people that I specifically choose, show me related stuff that I *might* be interested in.) There were jokes about a Meyers-Briggs type of RSS-consumer labeling. Maybe there’s an opportunity here to (first off, identify and then) tailor experiences to these individual type?
- Only 3% of the population is reported to be using RSS. So, the RSS culture in the Bay Area is a statistical outlier.
- Redundant posts really frustrate blog consumers
Functional Web Testing with Selenium
This was my own session. I will be making a separate post about this. (Including my slides.)
Open Source Visualization Tools
This was mostly a demo of GraphViz by Kent Bye. I have used GraphViz before for graphing ANT build files. Kent has uploaded several examples of the power of GraphViz to Flickr. There were several other options (both open and closed source) that were discussed. (One that sticks out in my mind: IBM Many Eyes.)
Exploiting the Borg. (Unlocking the Collective Learnings of the Last 60 Years of Computer Science.)
This was presented by one of the people from Koders.com which is a Source Code Search Engine. The basic premise is that Computer Science as a discipline is unique, in that all our collective learnings are contained within source code. (I take umbrage to that assertion, because source code doesn’t contain the intelligence about HOW to develop the code, which I feel is a large discriminator between an academic and a "real" Software Engineer.)
So, the Koders.com guy was trying to data-mine the audience for info on what challenges we see with adopting a "Search Driven Development" process. One of the potential problems that came up was how to appropriately flag the code search results, identifying what areas the provided code excels in… I.e. How do you prove and/or present that version A of an algorithm excels in memory optimization versus version B of an algorithm which excels in process optimization, etc…
Overall this talk left a bad taste in my mouth. I kind of dislike the idea of blind copy/paste programming. I hate it within a codebase, and I think it would be a bad pattern between codebases as well… (Unless there can be ‘trusted’ user-generated ratings about the efficacy of a snippet.)
From 8 to 8 Million, Scaling Pandora
This presentation was given by the CTO of Pandora. It was actually pretty illuminating. My favorite concept from the whole thing was how they handle their marketing. Evidently they provide a GREAT experience for their customers, because their user base pretty much grew organically. (They don’t really pay much for marketing.) What helps to define them is how they handle customer support: All emails in to their support address are sent to everyone in the company, and any one email may be answered by multiple employees, without any filter. Purportedly, this gives the user a feeling of "ownership". I LOVE this idea, and will be sharing it at work.
Another thing that I learned was that sometimes venture capital people can provide more than just money. Pandora had an investor that provided a full-time analyst for 6 months that concluded a subscription-based model wasn’t in their best interest, and recommended an ad-based model.
And, as far as scaling is concerned, he used a term: "the hockey stick". It describes (I assume) the increased rate of adoption when your site comes into the public consciousness, and the sharp upswing of pageviews (and consequently computing needs) that happens. He said that at this point their physical scaling became an exercise in planning more than technology, as it took up to 3 weeks to purchase and install a new server.
Take 6 Months Off and Travel the World
This was a presentation by a Palo Alto local woman who has put aside her business to pursue extended travel opportunities. She is trying to help other people realize they can do this. However, her approach was kind of ham-handed in this Web 2.0 crowd, because she is trying to affect this change by selling a book and specifically avoiding blogging. She also approached the presentation as a way to showcase her specific experiences, rather than as an opportunity to provide a cookbook for inexpensive living and travel.
The best ideas out of her were to work with volunteer/humanitarian organizations in exchange for room & board. She did some healthcare-related stuff, but she said there are organizations the world over without a web presence and no idea how to tackle the web.
This idea is not new on the ‘net (in fact I think I first saw it via LifeHacker), but in reality in the technology field, if you ask your employer for a 1-month leave of absence or whatever, it is *much* more cost effective to allow you to leave, than to have to go through the trouble of hiring in and training another engineer…
Some more things to look into if you are interested in this topic: Couchsurfing.com, Hostelling International, and something with a name similar to "Hotel Club" that is really just a house-sharing connection service.
Ok. That’s my report for today… I’ve got beer to go drink!