I was asked to talk about ‘The guerilla approach to aggregating online collections’, after my ‘
Exploring Museum Collections On-line: The Quantitative Method’ paper given at the international Museums and the Web 2008 conference in Montreal in April.

In my talk last week, I tried to give an overview of the ‘guerilla’ methods of aggregating museum collections, including my Freedom of Information (FOI) requests project and Mike and Dan’s screenscraping project, ‘hoard.it’. I also talked about the Object Wiki project I’ve been leading on at work, which is kind of a guerilla method of collections data creation, and the WhatDoTheyKnow website from MySociety, who I’ve been working with to add museums to (culminating in a spreadsheet of the size of museum’s collections).

Towards the end of my talk though, I focused on the need to build cross-museum collections website (which was what some of the morning talks were about). I often feel that there’s not enough thought put into the purpose and use-case for these websites, which are often driven by politics and funding requirements rather than a good idea. To try and counter this, I proposed 5 ideas for cross-site museum collections websites, each of which hopefully has a clear purpose and angle. For each of the ideas I also thought up a name and registered a domain name (as that’s now the first step after coming up with any idea).

Idea number #1 comes from some observations with my Object Wiki project. The observation is that people really relate to museum objects that they own, or have owned in the past. Nostalgia is a great pull, and I think in museums we’ve known this for a while. So I thought, we’ve already got the fairly well-known e-commerce site iwantoneofthose.com - how about having a site called ivegotoneofthose.com? The key idea here is to be able to browse or search objects from a variety of museums, and then, if you spot one you recognise, hit the ‘I’ve got one of those’ button. This could then spark a conversation - and would also allow you to create lists of ‘most recognised’ objects.

Idea number #2 is based on the idea that curator are the key people who get to decide what museums acquire. I’m not all that familiar with the process of object acquisitions - I understand that it involves numerous people, legal checks, asbestos checks and a few committees - but that there’s usually one curator who’s the champion of the object. The name is a riff off a MySociety project called TheyWorkForYou, which lets you see when MPs make speeches, and how they vote on bills. I thought it’d be good if you had a site called TheyCollectForYou.com. You could perhaps scan and OCR acquisition forms to generate the data, and then explore and examine the interests and prejudices of individual curators.

Idea number #3 derives from another observation from the Object Wiki project. There, one of the sections we’ve created for numerous objects is called ‘how it works’. This could be turned into a key way of viewing objects. There’s already a cool website called howstuffworks.com. How about creating a past tense version - howstuffworked.com? This could usefully be cross-museum, as explaining how one past technology worked might involve a component or comparison with an object from another museum.

Idea number #4 is another acquisitions based idea - but forward thinking. Perhaps we could engage people in the decisions about what should be collected and displayed in museums? Perhaps we could encourage people to donate objects to museums more? The name for this idea is taken from something Indiana Jones says - ‘that belongs in a museum‘. People could suggest stuff that really belongs in a museum, such as an antique, or a first-generation iPod. Others could vote on whether it really does belong in a museum, and then perhaps even museums could bid on whether it should be their museum that it should belong in - a kind of reverse eBay.

Idea number #5 involves something think people are always interested, the value of museum objects. Museums tend to shy away from talking too much about monetary values of objects, but this could become a key way of viewing online objects. My proposed name for this project is pricelessartefact.com - and would let users see the creme-de-la-creme of museum objects, and would ask them which they consider the most valuable - however you measure ‘value’.

So, five ideas. I’m happy to give any of these away, for free, if you’re interested in following them up. You can even have the domain name for free (registered for a year).

Otherwise, I’m interested in your feedback. Would these website projects interest you. How could they be funded? Are they the type of projects that museums themselves should be initiating, or should they be produced by the commercial sector (or perhaps a non-profit like MySociety)?

The day was run by Mike Ellis, who this year decided to provide some more collections data to play with by working with Dan Zambonini to screen-scrape the data (outlined in his blog post, Bootstrapping the Naw).

I decided to tack away from object data altogether, and instead play with exhibitions data. This was partly because I’ve realised recently that exhibitions data (the dates, names and descriptions of museum exhibitions) is even more invisible than object data, and yet potentially more useful, and more socially interesting. For this reason, I put together an exhibitions information API for the Science Museum a few weeks ago, after manually collecting the data from trawling through old websites (some of which are no longer even publicly accessible).

I spent the first half of the Mashed Museum day exploring Freebase, which I’d heard about a few times but never properly investigated. After a bit of reading through the help pages, I created an Exhibition ‘type’, and then started to import the Science Museum’s exhibitions through a semi-automated import process.

I spent the afternoon learning how to use the MIT Similie timeline to plot the exhibitions on a timeline, using the API from Freebase. I came to the conclusion that the API is actually pretty powerful, even more so because it has a useful Query Editor tool where you can run test queries and instantly see the results.

The finished demo is here: Exhibition Timeline.

It starts with a javascript prompt asking you to enter a museum name. I suggest entering ‘Science Museum’ if you want to see it with some decent data. However, you can also refresh and try a few other museums, and if there’s no data, then you can enter it in Freebase and instantly see the results.

I may try and refine it a bit over the coming weeks, but as a quick demo I was quite pleased - not particularly with the end results, but that the process of using Freebase and Similie can work so easily.

Have a play. And if you’re interested in investigating Freebase, sign up and leave a message on my profile.

There’s no set theme, but as I and a couple of the other organisers come from a museum background, I expect there’ll be quite a lot of talk about geek-led innovation in the ‘cultural sector’.

If you’re interested in coming along, add yourself to the Upcoming event and the mailing list.

What I haven’t written about yet is the overall ideas, thoughts and mental ‘to-dos’ that I came away with. So, whilst the conference is more or less still fresh in my mind, I thought I’d do that now, in a quick, succinct, list form.

All lists need a swanky name, so this one is called ‘Frankie’s top 10 take-home messages from the MW2008 conference‘.

1. Participate in Flickr Commons. George Oates (from Flickr) gave a great presentation about the benefits of the project, which are many, and the Q&A discussion at the end expressed the negatives, which are very few. Amongst the delegates, there was a great deal of positive buzz about the project, helped by the well-timed announcement that the Powerhouse Museum had already become the first museum to take part. So, all museums with a historic photographic collection should jump onboard. I’m doing my bit by persuading the National Media Museum to sign up.

2. Investigate faceted navigation. Faceted browsing isn’t hugely new, but it came up several times in the conference in several different sessions. Put simply, it allows you to navigate content (or data) along lots of different ‘facets’. For instance, on an e-commerce site these might be cost, colour, size, category, brand, style and so on. Usually, the more facets the better. Faceted navigation also applies to search, where the facets act as a filter that can be applied after receiving too many results (compare this with ‘advanced search’, where you have to select the facets in advance). Definitely something for museums to investigate and explore more, particularly for object websites.

3. Implement OpenSearch. This is a relatively simple framework for allowing third-party services to search your site and interpret the results. Terry Makewell from the V&A explored this a bit in his talk, and the National Maritime Museum announced that they’d implemented it whilst the conference was still taking place.

4. Lightweight and simple beats complex and hard. And I think that’s all that needs to be said about this one.

5. Blogs are still a pretty powerful engagement format. Okay, so they’re no longer new or innovative, but the form has settled, and they can build an ongoing relationship with readers in a way that’s hard to beat. Museums might still be worried about ‘barriers to blogging’, but these are getting easier to overcome, and it’s now becoming more about ‘how’ than ‘if’.

6. There’s a community of museum website people out there who care about openness and interoperability and standards and doing stuff. This realisation is one of those intangible benefits that a conference can have, coming not from the formal sessions, but the discussions and networking and debates that happen in the breaks and the lunchtimes and the evenings. There was a definite sense of frustration but yet optimism and action from a subset of the delegates, who I suspect will start to form a stronger community, and perhaps events and meetings of our own.

7. Successful web departments require a team of dedicated developers (or ‘geeks’). This came out in Aaron Cope’s talk, and also from the work of Seb Chan’s work at the Powerhouse Museum, and numerous other examples. They all suggested that technical development should be done mostly in-house, with a team who work together on the same code base, building experience, with control over their websites, and the freedom to try things out. Geeks drive things forward.

8. Publish exhibition data. This didn’t come directly from the conference, but was an idea I had whilst I was there. Museum people have talked for ages about opening up object data (including me, who gave a presentation on the very subject), but we’ve overlooked exhibition information. Exhibitions have a lot more social currency than objects (think how many exhibitions you can remember going to, versus how many museum objects you can remember). We tend to promote them whilst they’re on, then forget about them when they’re over, possibly even deleting all trace of them from our websites. I think there’s a definite role for an exhibitions information aggregator (basic things like opening and closing dates, title and location), and possibly even a social service. I’m going to investigate this by seeing whether the Science Museum even has a list of every exhibition ever put on. I invite other museums to do the same.

9. The OAI-PMH protocol (I’m not going to bother trying to explain it) is still yet to prove itself. It was rejected by the NMOL project and has been retired by Google as a sitemap-style tool. Maybe it’s not dead yet, and it still has some defenders, but it’s yet to prove itself as a must-have for every museum.

10. Air Canada food sucks, but food in Montreal is good, although some of it is best eaten drunk. (Okay, so there had to be at least one facetious comment).

So why have I got one, and how did I get it? Well, I bought mine through eBay, which some people have said is bad, but is recommended by the organisation itself as a quick way for developers to get hold of an xo. There was an official way to buy an xo last December, when OLPC ran a ‘give one, get one’ offer where you’d buy two xos, but one would go to a child in a poorer country. I would have loved to have gotten one this way, but the offer was only open to those in the USA and Canada. So I had to turn to eBay instead, paying an inflated price (and then having to pay customs import charges).

Why have I done this? Partly it’s to show support in the One Laptop Per Child program, which I think is a wonderful, idealistic and forward-thinking idea. Partly also it’s to explore the cutting-edge technology and thinking which has gone into the laptop’s design (including a low-power screen that you can read in direct sunlight and effortless wireless networking and collaboration). As well as this though, I’d like to try and help the mission directly, by helping to develop and produce some of the programs (called ‘activities’) which sit on the laptop. In particular, I think some of the ideas and concepts that I’ve worked on at the Science Museum, including the Launchball game which has been so successful, could easily be applied to the xo. I doubt Launchball could be fully re-created on the xo (it’s pretty complicated and processor-intensive), but something similar adopting the same science-based principles could be achieved.

I’m no way near being able to produce or code any of this yet, however, which is why I’m interested in hooking up with other London-based xo users and developers, or even just anyone who wants to learn more about the project. To this end, I’ve set up an OLPC London mailing list, and will look to host a group meeting once I’ve found enough members. So, if you live in London (or nearby), and have an xo, please join the group. Alternatively you can contact me at frankie@frankieroberto.com.

I’ll keep this blog updated with progress into the future.

(Photo by me, licensed under CC-BY)

Generally, I really enjoyed the conference, and had a great time, so I’m not going to bother going over everything that I liked about it. However a there are some notable elements that are worth mentioning, for the sake of other conference organisers if nothing else:

• The ‘Crit Room’ and ‘Usability Lab’ sessions. These are pretty unique, as they give selected website owners some immediate on-the-spot feedback as to how they can improve their site. The first has a panel, who look at the websites in advance and prepare a short response (as well as inviting comments from the floor). The second uses a live guinea-pig, picked from the audience, who is asked to complete a ‘task’ on the website (eg find out some information), watched over by a usability consultant. Both pretty useful, and concrete, even if you don’t own the particular website.

• The organised museum tours. I went on the day trip of museums in Montreal. Not only do you get to see a bunch of museums, but you get to see them with other museum folk too, and with someone from each of the museums present to explain their thinking. This was pretty fun, and it was interesting to see how museum people visit museums - unlike the regular visitors, we were taking photos of the interfaces, the signage, and were trying to figure out how it all worked.

• The evening dos (receptions). These were awesome - ever-flowing wine, tons of good food, and a relaxed atmosphere where it was easy to mingle and meet new people.

Okay, so time for some negative feedback (and these are mostly more specific to the particular conference):

• The wi-fi went down a few times. Okay, so this happens at nearly every conference and it seems churlish to complain, but hey, it was annoying at the time.

• The ‘Exhibit Hall’ (where commercial sponsors promote their products) didn’t seem to fit in well with the rest of the conference. Sure, the companies are helping to fund the conference, but there could have been a more interesting way of incorporating this (we discussed some ideas for a Dragon’s Den style pitching session, with voting from the audience).

• The workshop I went to (Everything RSS) seemed a little heavy on the presentation and light on the ‘work’ (which was just reviewing some web services). It might have been more engaging to have led people through some real-life activities.

• The closing plenary, by Clifford Lynch, was pretty hard-going, dragging on for ages, with no slides, and not much relationship to what had happened at the conference (he couldn’t actually attend most of it). It would have been better to have someone draw out some general themes and conclusions (and there were plenty of those, which I’ll save for a later post), with reference to some of the presentations, discussions, papers and so on. Even a slideshow of photos taken from the conference Flickr pool would have livened things up.

• The ‘roll-your-own’ session space was a good idea, but didn’t quite work on the day (it was difficult to publicise the sessions). It might have been better to have a dedicated time in the schedule when nothing else was on where people could propose last-minute (on the day, or perhaps in the week before) sessions, chalked up in a prominent place or quickly printed out and photocopied.

I hope these points are useful, and I invite other participants to leave a comment if they agree or disagree (or have any other bits of feedback they want to make public).

The only other minor quibble I have is that the style guidelines for the written papers should be updated so that “Web site” can be spelled “website” and “on-line” can be spelled “online”. :-)

Update: Mia has blogged her feedback too.