PassingNotes.com

“A child of five would understand this. Send someone to fetch a child of five.”
-Groucho Marx

A while back, I wrote this article about a web based bookmarking utility called Furl (originally published for SCIP ). As the weeks passed, curious emails started flowing in from readers regarding a site with a shockingly similar name and seemingly like features called Spurl. Given my unabashed dedication to equal-cynicism-for-all, it seemed to make a lot of sense to approach the founder of Spurl to discuss what the site is all about, and perhaps razz him…but he turned out to be a genuinely nice, well-intentioned guy. Upon much closer inspection, it is fair to say that these two sites started from two different ends of the see-saw…and so I present this detailed (and kinda lengthy) interview with Spurl.net founder Hjalmar Gislason as a view from the comparative fulcrum.

To preface this: if one says that Furl is about building highly personalized resource collections to potentially share with others, one might say just as simply that Spurl (www.spurl.net) is about building highly collaborative resource collections with others to perhaps personally keep. Like Picasso, Chaplin and Hitchcock (oh, and Cartman), Furl and Spurl are maddeningly simple for viewers to comprehend…but this simplicity belies a much more sophisticated goal for the audience. Each offers a unique feature set, certain distinct advantages, and quite different long terms plans as ‘going concerns.’

For readers who have absolutely no idea what Spurl.net is all about, could you offer an explanation that might make sense to my parents?

HG: Spurl.net helps users remember! It stores the useful and interesting things they come across while browsing. Once stored, these quality information sources are easily accessible again through full-text search. As thousands of users are using the same tool, Spurl can recommend things that you will be interested in and enable you to find related information that other Spurl users have come across. You can also easily share your findings with others or even work together in groups, collecting information on a given topic.

Can you tell me a little bit more about who you are, where you come from and what you do to pay rent? Is Spurl full time for you now?

HG: Spurl.net is my full-time job and is funded by a small group of private investors. I am the only full time employee at the moment, but have a good group of people around me, helping out on various fronts, ranging from programming and IT to business development and planning…I started thinking about a Spurl-like product a little more than a year ago. The original idea was a recommendations engine - somewhere that I and my friends could send in links that we found interesting and then at a convenient time check what was there and take a look - rather than bombing each other with emails and Instant Messages every time we came across something.

Then I started thinking about how this would scale and how one would - from a large number of users - make sure that you’d see the things you were most likely to find interesting. I didn’t have the time to do anything about it at the time - but wrote something about it on my blog in December last year. Around Christmas I found some free time and started coding what was released as “I like!” in January and became Spurl.net in early February.

You’re from Iceland, the most sparsely populated country in Europe. Could you tell readers exactly how to pronounce your name, “Hjalmar Gislason”?

HG: Not an uncommon question, so I have a standard reply for it: Hjalmar is like “hallmark” without the “k”. Gislason is “geese”(plural “goose”)-”la”-”son”. If you want to go into nitty-gritty details: 1) The “hj” sound is a phoneme that doesn’t exist in any other language 2) Icelanders have a unique naming convention where every child’s surname comes from the first name of their father. My father is “Gisli”, so “Gislason” literally means “son of Gisli”. His daughter (if he had one) would be “Gisladottir” (”dottir” = “daughter”)

Is “Spurling” something that you believe should exist at the major search engines? It would seem that ‘divine human screening’ is paramount, and this stands to set the search engines apart in an entirely new way.

HG: Not as much “believe should” as “know will”. I’m convinced that all of the big search engine / portal players will offer Spurl-like functionality within 12-18 months. There are two reasons 1) It’s great service for their users and more importantly 2) It’s probably as direct a source of human information about the Web as it can get.

I’ll get a little deeper into 2..This human information can be used to improve the search engines’ web site indexes, consequently improving the search results they return (for all users, not only the users of the Spurl-like features)…Consider the paid editors at Yahoo! and the human intelligence used to calculate PageRank…there are also other clear indications that the search industry is looking in this direction. Here’s a quote from a recent interview with Craig Silverstein, Technology Director at Google (and employee no. 1):

“In broad terms, techniques we use fall into, like, two or three categories, and one is we try to understand and leverage human intelligence. We look for signals that people put in to indicate intelligence, like deciding to link from one page to another or annotating text with the description of what the text is about.”

Some analysts claim that Google’s acquisition of Blogger was a way to tap into a huge source of human web site information, as blogging is - in a sense - all about linking heavily to other sources and related information. The “Advanced” features in the Google toolbar - when turned on (Displaying PageRank is one of them) - also enables them to tap into human usage information for all websites visited with the Google Toolbar installed. That is however only of limited value as it does not say anything about whether the user found the info on that page useful or why he or she was looking at it.

Spurling a page - on the other hand - is a sure indication that the page was worth the visit. In addition the user categorizes the page, writes a description, a few keywords or highlights an important snip from the text on the page. All of this is highly valuable, humanly written, meta information about the page.

Despite all the talk about Google, I still believe that of the three big ones, Yahoo! is in the best position to tap into this kind of source. They have a much tighter relationship with their users and are providing much more than search. Providing a Spurl-like service would be a logical service addition for their users, while at the same time contributing to their all-but-dead search engine. It could be a way for them to start fighting back as they desperately need innovation to stand a chance against Google (and soon MSN) in the search space.

Who do you think offers a relevant competitive service? Furl jumps to mind, but what about other online bookmarking sites?

HG: There is a lot of good stuff out there. I try not to spend too much time looking at the competition, but rather continue to develop Spurl.net, listening to my users’ feedback and following my own vision for the best bookmarking system. I also try to keep an elaborate balance between introducing new features and making what is already there as user-friendly and straight-forward as possible. The vision is that when it comes to maintaining a collection of online resources, whether it is for yourself only or to publish for others on the Web, Spurl.net will be the tool of choice.

One question that jumps to mind relates to the accusatory emails I’ve received in response to my Furl.net review - suggesting that Spurl has copied Furl. Could you comment on this?

HG: I don’t want this to be a “who was there first?” argument. There are similarities in the products, no doubt - some of them were introduced first in Spurl.net others in Furl, but most of them are simply logical for a product of this nature and many have been floating around for years in different products, research projects and online discussions. As for the unfortunate similarities in the name, “Spurl” originally comes from a friend’s idea: “Special URL(s)”, but is not supposed to stand for that - it’s just a name. I don’t remember if I had learned of Furl at the time of deciding on the name, but if I had had it in mind at the time, I would probably rather have refrained from using it.

The main differences between Spurl.net and Furl are in my mind: 1) the sidebar that quickly becomes a favorite tool of the users that give the service a spin, 2) the directories that allow groups to work together on a subject and 3) the complete anonymity even to the Spurl.net staff (all you have to do is choose a username and password).

Furl is a great tool and Mike is a very clever guy (we’ve spoken several times). I don’t know his exact plans with Furl, but for Spurl.net it is pretty clear. We’re going to be the best bookmark manager available. Using it is going to be free of charge, anonymous and without any ads. Our business is analyzing the web site information in our database and selling that unique source of human intelligence about the Web to search engines, web filtering companies, web traffic analysts and research firms.

How will you make money with Spurl, or will you? Is there some kind of plan beyond offering yet another cool, free service?

HG: It was a pet project…and I didn’t really see a business model in there at all, but the amount of feedback I got kept me up at night, so I continued to improve it. Many users insisted that I should turn the recommendations engine into an online bookmarking service…taking a look at what was available, I saw that not much was needed to be fully comparable to the best available, so I continued, and as I saw the loads of data that was coming in I spotted a business model in it:

Making use of this data to create a unique source of human intelligence about the Web and sell that information in aggregated form to search engines, filtering companies and others that benefit from such information. This is explained in our Terms of Use and Help section .

To explain, let’s use the search engine example. They will use this data to improve their web page index databases…One could say that while Google and other search engines have robot “spiders” harvesting information about the Web, Spurl.net’s thousands of users are doing the same by hand. Such human information (based on a human taking the time to actually read and review a page) is much more valuable than any automatically gathered data.

To emphasize that, two major stepping stones in the history of web search are based on humanly gathered information: 1) Yahoo’s directories, that for a long time had paid human editors and were at the time the go to place when searching the Web, and 2) Google’s PageRank, largely based on counting links to a page as a sign of human intelligence that the page had something to do with the words in the link and surrounding text.

Please note that the information Spurl.net gathers and sells is anonymous information about the websites - not about individual users. Beyond choosing a username and password, Spurl.net users do not have to provide any personal info and Spurl.net will not share any such information with a third party (more info on Spurl.net privacy).

To make this a reality I decided to go ahead and turn Spurl.net into the best online bookmarking service out there. The full-text search, the “Spurl bar” and cross-browser and OS compatibility did a lot to help with that. The social aspects and the syndication options were already there.

Who do you believe is offering the most innovative advancements for user in the world of search and meta-search? This could include university research, if something there jumps to mind.

HG: Google. But recently there are others doing very interesting things: Blog and feed searching with Technorati, Daypop, Feedster and the like….A9’s search query backlog is an example of something very simple that is a big improvement…Personalized search is interesting, but extremely tricky…Eurekster has some of the more innovative things coming out of that space…Clustered search results (a la Vivisimo) will be more common but needs to improve first.

If you could order the major search engines to change a thing or two, like a ‘boss for a day’ gig, what might your dictum sound like?

HG: Spurl.net related issues set aside I think some of the big opportunities are in tapping into the way users are using the search engines already - and analyze that to improve the search results.

I have for example never understood why the search engines are not counting the clicks that a link gets in the returned search results. This applies especially for very broad term searches. Example: If one of the result items in a search for a term like “music” is very seldom clicked, it is probably not a good result and should be replaced with another better one. In a similar way if a result that appears - say - number 15 in a result list is clicked more often than the law of averages says it should be, it is probably a good one and should be moved up.

The competition is all about the first 10 items in the search results. Most users never browse past that and will give up or turn to another source (or try another search) if the first 10 don’t turn up anything useful.

Another thing of a similar nature (and something that the engines might already be doing), is “listening in” on how the best searchers dig up information. You can learn a lot about how to improve your search engine if you can take a look at and analyze a whole 5-10 minute searching session of an advanced user, using maybe a dozen queries and carefully selected search conjunctions and modifiers.

What’s next for you, beyond Spurl? Are there other major ideas that you’ve been kicking around with real (or imaginary) friends?

HG: Spurl.net is my thing for the time being. But my notebook, that I keep with me at all times, is constantly filled with new ideas - most of them totally wild and wacky ones, others something that could be turned into value if one had the time and opportunity. I do share these with a close circle of friends that are in a similar position for feedback and improvisation, but as I said - my mind is 150% on Spurl.net now.

What’s on the near term horizon for Spurl? Any major features or changes readers and users should be looking for?

HG: We’re constantly improving and many of the latest improvements involve group-spurling or what we call “Spurl directories” . They allow users to work together on research on a specific subject and share their findings with one another or publish it on the Web (or just as a convenient form for an individual to publish his links). You can for example see my science and technology directory on the site. An improved version of the directories is in the making and we have high hopes for that part of the service. Example users of such features are obviously school-room classes, but also corporate departments (think e.g. programming groups), friend groups, organizations, group bloggers, etc. benefiting from each other’s online findings.

There is a range of other improvements on the way, mostly driven by user feedback - both from the Spurl.net forums and in emails from users.

Could you list a few of the most recent interesting Spurl sites for readers to check out? These don’t have to be your own, nor your favorites, just a few that you see as highly representational of what Spurl can do.

HG: The current implementation of the “group” feature is rather basic, but does its job. Still, that is the reason for the directory feature being kind of hidden, although it is open for everybody to use. An improved version of that is expected in the next ca. 2 weeks. For some quite active Directories you can for example take a look at:

http://dir.spurl.net/iteachered/
http://dir.spurl.net/Hjallis_sci_and_tech/
http://dir.spurl.net/creation/

If you have other questions, please feel free send a friendly email to Hjalmar Gislason, founder of Spurl.net at hjalmar@spurl.net - that’s what I did ;)

Share This

Some similar nonsense, if you like that kind of thing:

None Found