The Gong Show

Playing with Turntable.fm has taught me one thing for certain: our imagination about the range of interfaces for web services has been thus far quite limited.[1] Turntable’s a glimpse of going a step further.

90% of the web services I have played around with involve an iteration of the following three design templates:

1) Element pages (on Vimeo, this is a video page… on HackerNews, this is a news items permalink page and comment thread).

2) Profile pages (the permalinks for the users on the service).

3) List pages, that are comprised of some combination of elements and profiles (on Google, this is a SERP with 10 blue links, speckled with the occasional avatars of your friends…  On Twitter, this is the timeline, with tweets and their originating tweeters).

Turntable.fm by contrast tosses out a lot of these design metaphors in exchange for something different (not necessarily better or worse, but deliberately *different*). 

If a developer (in this case, Billy Chasen) built Turntable in Django[2], I think it likely it would have come out looking more like a basic CRUD app revolving around these 3 core pages. Instead, Turntable borrows design metaphors from gaming, chat, virtual worlds, and some design elements that are entirely their own (where else have you seen a web service that changes your avatar based on what operating system you use?).

Turntable is addicting and great, but I could apply those adjectives to most games I play regularly, and that’s not what I’m writing about in this post… I like turntable more just for being different and expanding our repertoire of design metaphors. Kudos.

Footnotes:

[1] I’m pretty sure Zach Klein my a similar statement quite broadly (not about turntable) some time ago… I can’t find the quote, but I agreed with him then, and I agree with him even more now.

[2] That’s not a knock on Django… I think Django is amazing, the people that have built it are even more amazing, and I’m incredibly grateful for all the hardwork that has gone into every open source web framework. More importantly, perhaps Billy *DID* build the service on Django for all I know… I’m just trying to say that frameworks often drive us towards default design decisions that best fit the framework’s construction, and that’s why if you squint you eyes at two apps built with the same framework, they often start to blur together aesthetically.

This post is not about a 55 MPH bus or Keanu Reeves (…Bodie!!). Sorry to disappoint.

I used to subscribe to the idea that “speed is a feature.” That aphorism is used to imply that making your web app load and run faster is a competitive advantage, akin to having better features. Now I find that aphorism to be a gross understatement. Speed is far more important than tag clouds, printable formatting, syntax highlighting, or any combination of features one might see on a product manager’s build schedule.

I often use a plain text editor (Notepad++) over a fancier word processor because it’s faster. I’ll give up the myriad of features Word offers in order to load the app 2 seconds faster.

Gmail is in the doghouse for me right now because of the countless loading bars I spend all day watching crawl. If I could pay Google $20/mo ($40? $100? What will it take?) for a guaranteed SLA on inbox load speed, I’d do it without blinking. Odd that Gmail is so slow considering how blazingly fast google.com is. Google has proven repeatedly that there’s an inverse correlation between page load time and usage, I’ll cite their data when I’m back on a computer tomorrow (thumbing out this post on an iPhone).

Web frameworks have made it incredibly easy for developers to build new apps, but it usually takes startups years to truly optimize those frameworks (or often completely rewrite many of the backend layers of abstraction) to make them run as fast as Google or Facebook.

I’ve been learning more about tools developers can use to speed up their apps. There are a number of diagnostic tools that help developers figure out what parts of their app are slow, but very few tools I’ve seen actually automate the process of fixing speed issues. Im not expecting some magic solution that turns clunky backend Ruby into highly optimized C code, but does anyone have any recommendations of apps that automate the process of optimizing a web app for speed?

Over the past few months we have been hiring at Spark Capital.  We announced the process with a blog post that friends around the internet kindly reposted, and we got a bunch of inbound interest.

The quality of the talent we saw was incredible.  After making more than a few tough decisions, we managed to narrow down the pool to a single offer.  I’m delighted to announce that this week David Haber has joined us at Spark Capital.

David has a interesting background, which included graduating from Harvard with a BioChem degree. Not a typical degree for a VC firm but we appreciated a different perspective and background than others at Spark.

He also is the raw clay that we think will work really well in an Analyst position: intelligent, net-native, passionate about startups, a self-starter, quantitative, and knows how to tell a good story. 

And most important of all, as we had one-on-one conversations with David, it just clicked.  VC firms are partnerships (not kingdoms) for good reason. The organizations are very flat, and the people that thrive in these environments know how to take a stance in a discussion and reason well from that stance in an intellectually honest way.  Every voice around the table is essential, and we know David’s going to fit in great around our table.

So, please help us in welcoming David.  You can reach him on: Twitter, Tumblr, and LinkedIn.