Alternative Super Suit

Not all of us can be Supermen. Michael for instance chose something else entirely…

Supermen

There’s nothing like the right t-shirt to boost self-confidence. It’s Mikkel in red and Toke in blue.

Actually, we are considering making Superman t-shirts mandatory throughout the office.

Helsinki

I had the pleasure to visit Helsinki this Tuesday doing a presentation about Summa. The audience was a good mix of librarians, technical people, and decision makers so I tried to strike a middle balance in the technical level of the presentation.

It was a two-hour thing so I was a bit nervous about whether the timing would match up in the end. Luckily it did that very well.

It was in relation to the Finnish National Digital Library project and I was number two (of three) presentations that day. You can find more information about the program here (sorry, Finnish only, but there is a picture of me me me!). I found this english article about the Finnish National Digital Library.

You can download an annotated PDF version of my slides on the Summa wiki. The PDF export messed up with the gradients a bit, but it is readable I will have the ODP sources and PNGs up when we have a proper place to host such things. If you want them feel free to email me.

Pink Navigation

Haven’t see the navigation by colour concept before. It works well, however at gizmine.com – apparently “the world’s largest Japanese gadgets and lifestyle design shop”.

Given the amount of pink stuff on the site, I expect their customers to be mostly gadget-crawing Japanese teenage girls.

Thx to Toke for the link.

Google makes bad usability to get user data?

Google’s search result page is now a wiki
Google has the introduced a new service that they call the Search Wiki.
The service allows a user (logged into his Google account) to:

1.  rearrange the order in which Google search result items appear,
2.  remove search result items and
3.  comment items and read other user’s comments.
4. In addition, the user can add his own items in the shape of URL’s to a given search result.

Here’s an example of a search for Barack Obama. The items can be moved up or removed. The top item has a green arrow because I moved it to the top spot. It can be moved down the list again, but for some reason it cannot be removed entirely ( a bug?). Clicking the bubble opens a comment slot.

The changes made by the user will remain and appear on future searches only if the user is logged in. Also, the changes will only be visible to the user himself and will not influence other users’ search results.

Bad usability in exchange for user data?
The remove item feature seems to be usable, but its a bit unclear what the point of being able to rearrange items is. Google does not mention whether they’ll be using the data about users’ rearranging activities to improve searchs result in general, only that the they now do offer the rearrange feature.

From my point of view, this is useless customisation. There is no real value in being able to move search result items around.

I do, however, expect Google to be smarter than this. There’s real value to be had about user preferences in the way people rearrange or remove items – and I’m pretty sure that the Google folks  are collecting such data in order to improve their general search feature.

Actually, they may also be looking to use it for their personalised search feature. Personalised search is an interesting Google feature living outside the spotlight. Occasionally – and only if your are using the English Google, I believe – you’ll notice a small message in the top right side of the screen, saying: “Personalized based on your web history”.

This is a potentially very smart feature: over time, a search engine can get collect a pretty good profile of users’ preferences through search activity analysis and use it to filter away noise and ambiguity from search results – all without direct user interaction.

But of course, its even better if you can cheat users to send more feedback by employing a Search Wiki service.

Comments as link meta data
The comments feature seems to be a bit more usable. One of the problems with links is that you don’t know whats behind them. Annotated links are conceptually good because you can use them to make an informed decision about whether to actually click or not.

But then again – what happens when you have 634 comments on a link? Or – as is the case with the top item for a Google search for Barack Obama – that you get comments like:

“BAMA MY MANNA”

“OBAMA!!”

“this function rocks!”

“Yes we did!”

“Very nice website”

“change!”

“Mr. President”

One of the comments, though, was quite usable: “modern website for an innovative leader. you’ll find all his social networks including flickr and youtube there (check out the backstage pics and videos)”.

Have a look at the comments yourself

Back to the Labs
Overall, this is an interesting Google experiment. The overall purpose is quite unclear, especially because people in general not are inclined to customise – even when the benefits are more evident than in this case – and I am sure Google knows this. There is no doubt that Search Wiki will get a lot of interest because its a Google service, but in the long run I’m quite sure that it’ll return to the Labs for dismantling. This said, I am sure Google will get heaps of good behavioral data to put back into their search engine from the experiment.

See also:

• An extensive guide to Google Search Wiki
• A critical view.

Summa@Oslo 2008-11-20

I am writing this blog post while waiting for my flight at Oslo airport. Unfortunately the airport doesn’t seem to have wifi so the actual posting will have to wait until later. I would have thought that all airports would have wifi by now, but I thought wrong.

So why am I in Oslo? I was invited to give a small talk about Summa/Search at Biblioteklaboratoriet. In particular I was asked to talk about the various services we have integrated in the web interface.

The talk itself went quite well I think (slides available here for the interested).

The main presentation was about a project called Pode, which is a library project working on mashups and alternative presentations of library information. They weren’t that far along in regards to actual implementations – however they appeared to have some good ideas, and I hope the meeting as a whole was able to provide some inspiration for them as well.

biblioteket.se also presented what they were doing. They have a nice website mashing up data from various sources, but what I actually found most interesting were some of their statistics. More than half of their users come to them through Google – ie. they completely bypass the site’s search interface and any information that might be available on the front page, and instead go directly to the full record view.

Also they have (among other things) user created tagging, ratings, and reviews. Their conclusions were pretty clear: Practically nobody writes reviews, tagging does a bit better, and ratings is the thing that actually gets used.

Another project they were involved in is Öppna Bibliotek. This is an attempt a creating an open service were libraries can store additional metadata about items (for instance reviews, discussion, etc.), and all the data is under a Creative Commons license.

After the presentations there were some general discussions about mashups and their potential – but the focus rather quickly shifted to that of ownership of the library records. It was seen as a real hindrance to creating open communities that the libraries were unable to give interested users a programmatic access to their records. And looking at things like the situation with OCLC I can only agree.

All in all it was a good trip, and I got to talk to some interesting people working on interesting projects.

A few random pictures from the trip:

Poltorak’s law of blog activity (PLBA)

Its relatively quiet here at the office right now. Summa 1.0 and even Summa 1.1 are out the door and the new version of Statsbiblioteket’s search engine has been released too.

At the same time, I’ve noticed, the activity on this blog has gone down quite considerably. Actually, I would expect the opposite to happen: more time, more time to think big thoughts and blog about them. But not so, apparently.

Observing this, I propose that: Blog activity is proportionally related to general work activity.

I call this relationship Poltorak’s law of blog activity (PLBA).

The general idea is that its easier to blog when being busy because you are forced to think less about what you blog about and in what style. When there’s much time available, however, own expectations about the outcome and quality of the individual blog post raise and pressure to deliver also goes up. As a result less blog posts are produced when time is an abundant resource.

I hope to be able to elaborate on this relationship in an upcoming article, but right now I have the time for it. Keep you posted.

The Summa Sandbox

For a while we have been talking about making a Summa Sandbox where we could show how Summa was being used. Both some of the real life production stuff but also the more experimental stuff we sometimes make.

Something we implemented a long time ago is OpenSearch. Using OpenSearch it is possible to add the search from the State and University Library to the Firefox search bar.

Because we have an OpenSearch interface it is also possible to add the result of a search to an rss reader and that way get automatically notified of any new books in the result set.

We have also successfully used our OpenSearch interface to quickly try some things out – the rss format is extremely easy to parse and use in pretty much any programming language.

One of the things we have made is a gadget for iGoogle.

It allows you to search the materials at the State and University Library and shows the first 5 hits in a compact format. Clicking on a hit will take you to the full record at our regular search website.

To try it out go to your iGoogle page, and click “Add stuff…” then click on “Add feed or gadget” in the menu on the left. In the input field enter the following url:

http://www.statsbiblioteket.dk/search/summagadget.xml

Keep reading this blog to stay updated on other things we might try out.

Summa 1.1 Released

I believe that a resounding “w00t” is in order now that Summa 1.1 has hit the block. Another small open source project tried to steal our thunder by releasing the day before SUmma 1.1, but I don’t think they got away with it

The development phase leading up to 1.1 was really nice. We had everything estimated and written down before we started and our estimates held right up to the last day.

I can’t help but feel that it was kind of like playing chicken with time itself. Personally I had some tricky stuff to write and a lot of stuff to consolidate before 1.1 and there where moments where I had my doubts that I would be able to make it. But even though the roadmap was pretty packed  I could see that we had consistently been meeting our deadlines so i just repeated to my self “Trust the plan… Trust the plan” and stuck to the to my designated tasks. – And you know what? Yes you do, we did make it on time

So here’s to the team!