Panlibus Blog

Archive for the 'Five Questions' Category

Five Questions With… Otis Gospodnetić

In this posting, the first of an occasional series, I pose five questions to Otis Gospodnetić, founder of Simpy, a social bookmarking service. I first came across Simpy several months ago and my first thoughts were of how deep the search aspect ran in the service. It underlies all kinds of important features including the highly useful Topics and Topic Filters. A little Googling soon revealed the reason: Otis is a long time member of the Apache Lucene team. As luck would have it our online paths crossed recently and I took the opportunity to interview him on the philosophy behind Simpy, the rise of social bookmarking systems and the future for independent software developers:

What was your motivation for starting Simpy and how did your background in full text search influence its design?

I’ve always been interested in “information”. Yes, “information” is a pretty
broad term and quite vague; for reasons that I don’t fully understand myself,
I find pleasure in gathering information (think web crawling), processing it
in various ways (think classification), and retrieving it (think full-text
search). Perhaps this attraction to information comes from my thirst for
knowledge and the notion that a large and searchable information repository
represents access to that knowledge.

I have also always been intrigued by software that has a human-like
collaborative side to it (think intelligent agents that work together toward
accomplishing a common goal).

I started Simpy back in 2001. It started as a
personal side project, as the result of my frustration with the fact that I
had this nice collection of bookmarks in my browser, but didn’t have access
to the knowledge hidden in it, because I couldn’t search my bookmarks. Sure,
you can find a bookmark by title, if you remember it, but the real knowledge
is in the page itself, and I wanted to have the ability to find it.
Similarly, I’ve always (maybe not always, but for a decade at least)
understood that hierarchies are limited, and that search is king (see
Categories vs. Keywords vs. Labels vs. Tags
). Being one of Lucene’s developers, I built
search into Simpy from the very beginning.

Why are social bookmarking services so interesting right now? After all, backflip has been providing shared bookmarks since 1999.

Sharing:
Backflip and other bookmark services from that era were/are primarily
web-based bookmark managers. Sharing was always implemented in a way that
required a lot of extra work and not built into the flow. It was something
that required an afterthought. The current generation of social bookmarks
makes sharing flawless, and this gives us easier access to more information.

Tags:
Tags are, essentially, user-provided _search_ entry points. Traditional
bookmark managers didn’t provide that – they were busy implementing tools to
let people manage their bookmark hierarchies. Tags give us the ability to
drill deep into the large corpus of knowledge along any one of the N axis.

Feeds:
In the world of traditional bookmark managers you didn’t need feeds – sharing
was an infrequently used feature, and why would you need a feed of your own
bookmarks? Now that more information is being shared through bookmarking
services, feeds let us passively suck in more information.

Searching:
5+ years ago, relatively few people saw that search is a more powerful way to
access a large body of data. Even though we had several Web search engines
before Google, it took Google to educate the masses to the point where nearly
everyone knows how to use search as a way to get to information.

What is the secret to building successful social systems?

Imitate nature. Understand core, raw human behaviour.

We’re seeing islands of data forming across the Web held in services like
Simpy, del.icio.us, Flickr, Blogger and Backpack. All these services have
public APIs and yet none of them intercommunicate. Why do you think this is
and what needs to be done to join all these islands up?

This is nothing specific to the (types of) services you mentioned. You’ll
see the same in any other field. Until somebody doesn’t see an opportunity
in connecting the isolated islands of information, the problem will remain
unsolved. I am sure, though, there are several people thinking and working
hard on this problem at this very moment.

I have passion for (human) knowledge, and a humanistic streak in me. I think
it’s hereditary, because I observed this pattern in my ancestors, too. That
part of me would love to build bridges between those islands of knowledge,
for the benefit of the society at large. Unfortunately, there are only 24
hours in a day.

With the giants of the Web such as Amazon, Google and Yahoo releasing so
many free services, where do you see the future for independent software
developers?

There have always been giants like the ones you mention. When Google ran on
a couple of servers there was this big and powerful search engine called
AltaVista. There was Inktomi. Before that there was WebCrawler. They were
all giants in their own time. Amazon, Google, and Yahoo are simply the
giants of this slice of our time. 100 years from now there will be new
giants. Some of them will rise from those independent software developers.

Thanks to Otis for taking the time to answer my questions. If you have a suggestion for someone you’d like me interview in the same style, drop me a line at ian.davis@talis.com. I know who I’d like to interview, but who would you like to hear air their views on Web 2.0, participation and the future of online services and data.