The 24 Hour Museum Tunable RSS News Feed

http://www.24hourmuseum.org.uk/home.rss

Abstract

This paper follows on from an earlier offering, 'RSS - the 24 Hour Museum News feed' (Pratty, 2004) presented to Museums and the Web in 2004. That paper outlined some of the background to RSS culture and demonstrated the successful build and deployment of the UK's first Museum and Gallery news feed service by the 24 Hour Museum (http://www.24hourmuseum.org.uk) and System Simulation Ltd of London (http://www.ssl.co.uk). Version 1.0 of the RSS feed (http://www.24hourmuseum.org.uk/home.rss) is working, and in January 2005, technical partners SSL developed the feed into a  multi-region, multi-city, tunable content feed, capable of offering tailored output to meet specific editorial needs of the 24 HM Web site, as well as of partners in the museum sector and beyond in the UK. The paper examines the technical challenges such expansion involves. Questions about RSS version choice, scoping, determining the logistics problems and operational decisions are explored. Since this paper proposal was submitted in September 2004, there's been a gratifying leap in the popularity of RSS as a mass publishing mechanism. The writer feels the growing momentum behind the technology grandly justifies the 24 Hour Museum's initial investment in the technology in 2003 and our continued commitment to the format. A final section of the paper deals with this gathering enthusiasm within the Web world and looks at brief case studies of other museums starting to use RSS.

Keywords: RSS, News feed, RDF, XML, Museum, Virtual Museum, Weblog

Fig 1:  display box showing the new tunable RSS feed

Building The 24 Hour Museum RSS Multi-Region Multi-Subject Feed

A Brief Resume Of Version 1.0

In 2003, System Simulation Ltd built an RSS 2.0 output for the main 24 Hour Museum Index+ plus Content Management System (http://www.ssl.co.uk/content/indexplus.html). The feed broadcast a small range of content types associated with the whole geographical coverage zone of the Web site via a single RSS address: http://www.24hourmuseum.org.uk/home.rss.

Version 1.0 was an unqualified success, boosting visitor sessions, improving visibility on search engines and reaching audiences the conventional Web site couldn't reach. By January 2004, the RSS feed was the third most popular point of entry to the site.

Why Build The New Feed?

In the light of this experience, it became clear to editorial staff that a subsequent major project, City Heritage Guides (http://www.24hourmuseum.org.uk/cityheritage/index.html) funded by the UK Department of Culture Media and Sport's Culture Online team (http://www.cultureonline.gov.uk/) would also benefit from multiple feeds.

This project saw the successful launch in September 2004 of museum, gallery and heritage guides to 10 popular English cities. The extra Web visibility, plus the possible partnership gains from being able to offer a free city culture news feed, meant that an RSS set-up was a must for the project.

As an extension to the 10 City Heritage Guide feeds, it was also realized that regional feeds plus outputs for our children's site www.show.me.uk would be key to site development.    

Getting Started With Scoping

Scoping the tunable feeds was even more difficult than the first feed. That first feed outputted certain article types from the homepage of the main site. It was difficult to choose content types, geographical feed areas and methods of display for the new feed. Another problem was deciding how much scope there should be for giving outside agencies like partner museum organizations a Web interface where they could choose feed parameters, effectively building their own feeds. There's still little knowledge of RSS in the museum sector, so surveying the sector's needs would be hard.

In the end, pressure of work building, launching and sustaining the City Heritage Guides meant that we opted for a simple version 2.0 tunable feed with no 'customer' interface and a rudimentary display box offered as JavaScript that can be used by Webmasters on partner sites.

Fig 2:  RSS display box

Resourcing difficulties also curtailed plans to make the feed tuner capable of offering subject-based feeds, at least in this incarnation of the project. To do so would need the database to be batch sorted with subject fields to allow themed feeds to be created.

Another resource drain would have been that the likely subject areas would need to be researched. This would need analysis of audience searches on the 24 HM site, an appreciation of popular UK National Curriculum subjects, and some things not covered by the curriculum, like archaeology and art collections on-line.

Since this phase of the project was being self-financed, it was decided that subject-specific output and themes could wait until 24 HM had further developed the main Web site to accommodate these feeds, when there might be more resources for proper research and development. At this stage, 24 HM will not offer a resource discovery facility, but will publicize the existence of a feed in the normal way. At all times, the technical team from SSL led the build and offered advice about scaling the idea, possible technology problems and so on.

RSS Version Choice

When the first feed was built, it was clear that version choice was possibly a difficulty. In the UK, the academic community, the further education sector and UKOLN (http://www.ukoln.ac.uk) recommended that RSS 1.0 should become a standard for the sector. But publishers in other sectors disagreed, going for RSS 2.0.

In version 1.0, RSS stands for RDF Site Summary. It uses the Resource Description Framework based on the W3C recommended framework for metadata. Find out more on the W3C Website at: (http://www.w3.org/RDF/) and see the 1.0 specification at: (http://Web.resource.org/rss/1.0/)

Version 2.0 is not a development of version 1.0.  It is simpler and it was developed from versions of RSS from the 0.9 series. It stands for Really Simple Syndication. Details of version histories can be found here: (http://blogs.law.harvard.edu/tech/rss)

More about the benefits of version 1.0 can be found on the Web: Paul Miller, now Director of the Common Information Environment project in the UK (http://www.common-info.org.uk/), wrote some excellent guidelines for RSS use in an Ariadne article, published 2003, (http://www.ariadne.ac.uk/issue35/miller/.) He also contributed some concise guidelines for RSS version choice to an EEVL (http://www.eevl.ac.uk) document about RSS – (http://www.eevl.ac.uk/rss_primer/#ref4.) This is a reliable and updated history and primer written by M. Moffat.

The decision to make the 24 Hour Museum feeds in RSS 2.0 was not taken lightly, but was a judgment based on research into the popularity of the various version types in the many sectors of publishing that the site crosses into.

Fig 3:  RSS versions as seen on Syndic8.com

Look at this page (http://www.syndic8.com/stats.php?Section=rss) on (http://www.syndic8.com) and see how many feeds using Really Simple Syndication (V2.0) there are. (Syndic8.com can be slow to access.) Versions of RSS 0.9 and 2.0 completely dominate the displayed pie chart. The cliché that ubiquity is all is a powerful signifier here – RSS 1.0 may be superior, but 2.0 is everywhere.

As of February 2005, the momentum now in most mass publishing outlets is behind RSS 2.0. The 2003 decision by 24 HM over version choice is vindicated by the further spread of 'Blogging' software completely driven by RSS version 2.0.

Fig 4: Tunable feed configuration record

Prototype Working

As of January 30, 2005, a prototype of the feed tuner has been working, organized by a configuration database record contained within the Index+ database we normally update the Web site with.

Regional data fields, city-based data fields and other content zones of the site such as institutional records (there are 3300 of them on 24 HM) and the children's section were switched through the database so that RSS output was possible in a number of different permutations.

For example, a Yorkshire museums feed is possible, including cities like Leeds where we have a City Heritage Guide. Institution records describing museums and galleries can be selected on the basis of their city or administrative region. These are both controlled values that identify information of relevance to particular constituents. 24 Hour Museum is currently cleaning up and validating the addresses so that more detailed selections, based on a geo code or postcode, can be made.

Other content types do not have an obvious address; however, any of them may be associated with one or more institutions. If this is the case, the content item is selected on the basis of the location information from the associated institution record. Where a content item is associated with multiple institution records, values will be inherited from all associated museums. For example, a trail that references a museum in Brighton and one in London would be served in both the Brighton and London feeds.

Fig 5: Config record initialized, ready to input

Setting Up A Feed Step-By-Step

It's a reasonably simple process, carried out within the Index+ editing client, which looks like this, above. This is a WIX GUI that interacts with the database of the site directly at Rutherford Appleton Laboratory in Oxford.

Once logged in to the database, a new configuration record is created for a news feed record. This creates a record initialized with empty news feed parameters. The Advanced tab takes us to the Name box, and a short descriptive news feed name is entered.  This name forms part of the URL for the new feed and so is made up only of lower-case letters, numbers and underscores. For example: london_events. Then the Params tab is clicked and feed parameter values are entered, such as a brief title for the feed; a description; the museum region whose info is being outputted; cities covered.

At this stage, types of content can be chosen. Content types that can be fed out in this version of the feed are museums (institutional records), events (listings), educational resources, articles (news, reviews or features), trails (long narrative features), Web sites (site of the week stories) and children's stories.

Some of these are quite simple in their structure; for instance, a Web site of the week  record is little more than a title, a description and a URL, whereas an Institution record describes a museum, contact details, collections, opening hours, facilities, etc in over 30 structured fields. Any content to be supplied in an RSS feed has to be mapped on to the fields supported by RSS 2.0. This mapping is done when the specific feed is configured.

Fig 6:  Setting up content types in the config record

Content channels can also be set at this stage. This is because the 24 Hour Museum database supports a number of different Web sites. Content can be channeled to specific Websites or pages within a Web site. These channels can also be used to select content for the RSS feeds.  Adding to the list of channel names and then associating the channel with the relevant content items can set up new channels for selection.

Once all these parameters have been set, the record can be saved and the feed can be examined, either within a browser as source code, or through a feed reader or a test display box. The feed is viewable for authorised users at this stage on the 24 Hour Museum update Web site

http://www.24hourmuseum.net/production/redesign/rss/<feed_name>.rss

[readers of this paper won't be able to examine that - it's a password protected site.] The feed reader URL would be this:

http://www.24hourmuseum.net/production/redesign/rss/<feed_name>.rss

All the feeds when published are public in that there is no formal access control such passwords or IP-based restriction. 24 Hour Museum staff can set up a feed with an obscure name if they want to hide the existence of a feed.

Fig 7:  Looking at the feed in a reader

Presentation

Most feed users will access the feeds using the RSS feed client of their choice. The 24 Hour Museum facility also allows users to create a Web page with an embedded panel displaying a feed by including the following in their Web page:

    <script language="JavaScript" type="text/css"

       src="http://www.24hourmuseum.org.uk/rss/display.js?feed=<name of feed>">

    </script>

The JavaScript supports version 4 and later browsers. The default CSS classes provided give the 24 Hour Museum branding. This can be overridden to customize it for a particular Web site.

Glitches

When specifying the feed, 24 HM staff didn't realize that certain fields in the database needed updating. Fields for the regional museum agencies needed renaming and the physical boundaries, checking. In most cases this is not a problem, but for certain regions, namely the South East, this is urgently being rectified.

This means there's currently a problem when trying to output a feed matched to certain regional museum boundaries, but that will be fixed a few weeks.

There are also some issues about displaying the feed. When scoping the tunable feed, staff requested the development of a natty 24 HM-branded display box. This would show whatever record types are being fed out to a region or city. But how do you distinguish graphically between news links in a feed box and some listings?

Fig 8:  Display box showing record type signifiers

At the moment we are trying to preface each link with a 'signifier’ like this - ‘Article: Churchill Museum Opens In London' It's not perfect, but we need to be able to mix record types in one display box - or we think we do. This is uncharted territory for us, and we need to get the project working in public and user-test the various box design options, including single subject boxes, multi-subject boxes and all the variations in between.

More issues discovered in February include display box length - on an external site, does the text extend the length of the box? Does it just reach a length limit and then stop dead?  Do the old story links always show? What happens if Web masters cache the site (or 'scrape' it), and what if Web users 'flush their cache' regularly, as some do? 

Logistical Aspects

Feeds need content. The content could be listings, virtual objects, calendars, features, institutional information, slow changing content, quick changing content or even multimedia files or pictures.

In terms of content logistics, the tunable feed is closely linked with a network of 12 student journalist writers in nine UK regions set up by 24 Hour Museum December 2004/January 2005. See this article about the network - http://www.24hourmuseum.org.uk/nwh_gfx_en/ART26114.html

Students on the MLA (www.mla.gov.uk)-funded network have competed for bursary awards (£500 UKP) and are 'embedded' (Gulf War-style!) within museum or gallery marketing departments, with the idea of covering stories as they pass through the museum. Special emphasis is placed on covering the results of MLA's Renaissance in the Regions regional funding program, as it was Renaissance funding that actually supported the bursaries and pays for supervision from our Brighton office. More about Renaissance here: http://www.mla.gov.uk/action/regional/00renaiss.asp  

As the students get working, more and more of their regional stories will power the new tunable feeds, and their content will be shared with museum, gallery and heritage sector partners in the area, as well as anyone else in the city or region.

This copyright-free situation places emphasis on the major quality of XML - if you can't share it for free via XML, don't bother with RSS!

Validation

One troubling aspect of RSS feed publication is the constant worry about validation of the feed. The reality of all this super-efficient streaming of content is that an awkward diacritic character can slip through checks into the live feed.

The way our feed works means that one odd bad character can jam the feed to many syndicators and make the feed look odd to home users, although feed reader programs are generally very tolerant of bad character reference entities.

Editorial staff had to learn to check and treble check diacritics, especially in abstracts or titles. There's no current way worked out to preview the validity of our pages as we use an update workshop site, which would not validate anyway via conventional means. The first version of our feed does not publish until the site is sent live, which happens overnight using a chron job on the Unix/Solaris host system.

The new feeds promise more flexibility of content types, but there is also the danger that feeding out previously unseen parts of our database will mean more diacritics to fix, and the prospect of unexpected glitches arising as a result.

It's clear we need to develop a more sophisticated means to instantly validate new feeds and feed fields. Running a validator locally would be the answer to our problems, and there are a number of open source programs available. For more about validation see Syndic8.com or look at a validator that can also be run locally: (http://www.feedvalidator.org.) 

Try The Feeds!

A page on the 24 Hour Museum Web site (http://www.24hourmuseum.org.uk) will be live by the end of February with information about the new feeds, including URLs of city and regional feeds that are working and available for free for Webmasters to use and for individuals to access using feed readers.

RSS And Museums - Is It The Way Forward?
Publishing Challenges, New Users And What Comes Next

The first version of the 24 HM news feed was quietly 'switched on' in September 2003. At first the facility was just submitted to one 'syndicator' (www.newsnow.com); in fact, what we now term an 'aggregator.' This careful unveiling was done so that some degree of evaluative work could be done examining which RSS functions worked best: attracting readers via feed readers, Google, or adoption of the feed by other Websites.

Analysis of 24 Hour Museum Web statistics in 2003  (http://www.24hourmuseum.org.uk/etc/formuseums/TXT25733_gfx_en.html) and 2004 (http://www.24hourmuseum.org.uk/etc/formuseums/TXT25735_gfx_en.html) after the deployment of the first feed shows interesting trends. Take-up of the feed was effective but quite slow - indicating that there's lots more to syndicating content via RSS than just turning on a feed and expecting the masses to visit.

When the time came to publicize the feed, interest was generated by press releases e-mailed out to the tech sector, museum press, arts and museum newspapers, and the general UK and world media. The feed carried messages out to the wider Google sphere very successfully. Stories would appear in Web searches very quickly indeed (a matter of hours) after the feed was switched on. Once Google introduced their Google News facility, we submitted our feed as a source and it was just a matter of minutes before newly published 24 HM stories appeared on the news.

There's a keen interest in XML and RSS issues worldwide, and the tech community that listens out for such subjects is eclectic and open-minded about new uses for technology. We found lots of Web sites about RSS carrying our press release about the feed.

As stated, just turning on a new output of content like this does not automatically plug you into a fantastic stream of new and happy site users. Submitting the feed for syndication with organizations such as http://www.Syndic8.com, http://www.Moreover.com or Google News is absolutely vital to making sure potential readers can find the feed in lists of feeds available.

At the moment (January 2005), there's still not that many RSS news feeds from the museum and gallery sector worldwide. A quick search on the major syndicator sites will turn up ten or twenty US museums, some in Europe and one or two enterprising institutions in the UK. This subject rarity makes distinctive feeds such as Museums and Galleries stand out even more.

User Patterns And Some Surprising Figures

As summer ended in 2004 and the winter set in, odd things started happening with visit figures to our site. Not only that, but as December turned into January, site slowdowns began to occur. It could only be bandwidth issues, where too many people - or robots - start to request pages from the site server. 24 Hour Museum was switched on to a faster server earlier in 2004 by our hosts, Rutherford Appleton Laboratory (RAL), so it was decided that the problem wasn't simply a slowdown server-side.

Our investigations turned to who might be visiting the Web site. One or two new browsers were logged entering the site through the RSS feed. This coincided with the site slowdowns. The slowdowns happened with clockwork regularity twice a day, particularly around 5.30 pm UK time. 

On the face of it, stats analysis showed visitors to the site were rocketing –resulting in 24 HM publishing gushing stories about record visitor numbers, nudging the half million mark in December. The truth, arrived at after some frantic enquiries around the Web sector in London, was interesting, though it demolished our record viewing figures story.  

Around November 2004, Pluck.com (www.pluck.com) launched their Plucker feed reader to the world, a free program that blitzed 24 Hour Museum servers with frequent 'polling' events, checking for newly published RSS stories. Plucker appeared from nowhere in our stats as being responsible for a massive 44% of visits in December and even more in January 2005. 

We realized that perhaps only 1 in 10 or less of the new visits were human - the rest were robots.

Fig 8a:  Graph of visitor sessions

With Pluck in mind, here's a quick look at visitor sessions recorded recently by the 24 Hour Museum's Summary statistical measuring software. 24 Hour Museum marketing manager Ruth Cobb has been looking into the way Summary records all visitor to the site in total, visits via RSS, without robot visits, and with robot noted.

First I calculated the percentage of entries to the RSS news feed page - which up until November were around 5%. I then took that percentage off the total number of visitor figures. When we have the huge jump in Dec and Jan - doing this actually puts our visitor sessions on a downward trend.

However, the second thing I did was to add up the number of both known and unknown robots each month. Then I took this figure away from the overall visitor sessions. Again with these there was a huge jump in Dec and Jan. However the interesting thing about doing this is that this puts our figures on a healthy upward rise.

I'm not sure which is a more accurate method yet - but it is important as the difference in overall trends is quite stark.     (Ruth Cobb)

Fig 9:  RSS feed stats measured in different ways

It's too early to make precise conclusions about this episode, but what is certainly needed are guidelines from public sector cultural agencies in the UK and worldwide about how to record Web stats in the XML and RSS era.  When more museum sector collections databases are joined up and collectively searchable by individuals and robots alike, it'll be really hard to gauge usage, relative success and measures of value for public money. 

Measuring RSS Stats:  The UK's Culture On-line Project

To find out a bit more about what a museum sector agency approach to measuring RSS stats could be, I asked Chris Barrett, specialist in Web accessibility and Web standards at the UK's Culture Online project (www.cultureonline.gov.uk). Barrett's statement is not an expression of official standards for RSS measurement, but an idea of what future standards might involve.

In brief, the traffic generated by delivering an RSS feed itself should not really be considered as a page impression if it does not involve a human visit.

Typically, RSS feeds are aggregated on a regular schedule by automated tools (either installed on a users desktop, or via a Web service of some kind). This may occur very frequently (per user), depending on the tool / or schedule set up. It is not until a user follows a link from these aggregated lists that a true page impression occurs.

This is not to say that a high number of RSS feed hits is not a good thing. It's just hard to claim from this alone that lots of traffic is being generated this way.

In this case of a user following a link from a feed, the page impression will appear in the server log files as a deep link (either with a referrer if the aggregated link were published on a Website, or possibly with no referrer if the link were followed from a desktop tool.) To make it easier to positively identify links from RSS feeds, you may wish to place an identifier in the URLs distributed in the RSS.

The difficulty in establishing user reach, of course, is when the syndicated content itself stands alone. Here, a single RSS item contains its full content - not just an introduction enticing the reader in for more. In this case, for each request of the RSS feed, you may potentially be receiving a true page impression in terms of human eyeballs.

For this reason, it is important to ensure that the http caching pragmas such as last-modified are set, and that your Web server responds to a head request. This way, a properly configured user agent (supporting these features) will not request the whole RSS feed until the content has changed (and at such point could more legitimately be counted as a true page impression). 

(Chris Barrett, access and standards advisor, Culture Online)

Weblogs Affect User Figures, Copyright Issues Too

One and a half years in from the initial project idea a whole new group of Web users have made RSS a key part of their ways of working. Webloggers or bloggers use systems like Movable Type (http://www.sixapart.com/movabletype/) or Blogger (http://www.blogger.com/start) to post on-line diaries, automatically updating content using a simple content management system. Most Web logs are available as RSS or Atom feeds.

Many bloggers make content using quite restrictive technical facilities - some blog programs don't allow direct uploading and publishing of photos. It's common practice for bloggers to do Google image searches for pictures to illustrate new stories, then deep link to pictures on the Web sites of others - such as the 24 Hour Museum!

This is another unpredicted source of traffic to our site. A cursory viewing of the 'most popular page element' in our Summary stats package often shows Web visitors are just accessing parts of a page - usually a picture. It perhaps won't surprise readers to know one of the most popular images served up by the 24 HM is one of Marilyn Monroe. Naturally, of course, we are concerned about the copyright legality of these partial page views. Wholesale adoption of RSS Web casting in my opinion will inevitably lead to complicated IPR situations.

If Web sites such as the 24 HM are to develop and evolve into widely Web-searched resource banks or portals to on-line collections, solutions to copyright challenges offered by projects such as Creative Commons (http://creativecommons.org/) need to be adopted creatively. CC is already able to offer layers of contextually appropriate licenses embedded within a digitally published RSS object.

Further developments like MLA's Knowledge Web project (http://www.mla.gov.uk/documents/iik_kw.pdf) could make intelligent use of RSS and XML content sharing mechanisms - and these developments will need to acknowledge the fragmentation of museum content and collection objects on the Web and the need to control copyright, or develop new less stringent attitudes within museums to copyright issues. 

 Fig 10:  National Museums Liverpool homepage

Brief Case Study – Good Practice And Initiative At National Museums Liverpool. Why Make The Feed?

 As you have read in the introduction to this paper, the 24 Hour Museum was the first UK museum to develop an RSS Feed in 2003. Other UK museums have been slow to adopt the technology for their use. National Museums Liverpool stand out as pathfinders for the museum sector, however, having self-developed in-house a range of useful news feeds from their Web site. The feeds, at http://www.liverpoolmuseums.org.uk/about/developersguide.asp, went live September 17, 2004.

Dawn Capper is Britain's first museum 'e-PR officer', based at National Museums Liverpool. NML also have just appointed a Web Strategy Coordinator, Karen Miller, and there are regular meetings of the Web Communications Group.  

"We were aware that a lot of people are looking for content for their sites and as a national museum with changing exhibits, events and lots of things going on in general, we felt that a feed of stories about what was going on would be a really good idea," said Dawn, interviewed for this paper. 

At the moment the feeds syndicate content that falls somewhere between being press info and content aimed squarely at the public, something that Dawn is happy to admit. So there's still some editorial tuning to get done.

As yet, the output has not been submitted to any syndicators or aggregators.

Although the user figures aren't earth shattering, I think you'll agree they are a good starting point and that the project is in its infant stage. I would certainly like to publicize the feeds and make sure people know they are there, but haven't had the opportunity as yet. We do want to be innovators though, not emulators, so this project fits in well with that strategy. (Dawn Capper)

For NML, seeking inspiration from the cultural sector for innovative new ideas is not that simple. "It's very difficult when thinking of the Web to find a benchmark for new technologies within the museum world. You always should be looking outside the sector to see what new approaches people are planning," said Dawn, explaining that mass publishers like The BBC and The Guardian newspaper were good examples of this. 

One really good Liverpool idea in particular is the 'Picture of the Month' feed, which broadcasts a .jpg and some text to publicize a collection item. Here's the feed address – copy and paste it into your feed reader and check it out! http://www.liverpoolmuseums.org.uk/whatsonnet/pom_feed.aspx?venue=7

Fig 11: The Liverpool feeds

How Is It Done? 

Neil Harding is the Web expert at National Museums Liverpool who took the plunge and put the feed together.

To be honest it's just something that I thought we should have. I'd read some tutorials on how to put feeds together and just followed them through! But for a large, regularly updating content provider like us it's a natural facility to have. The BBC, The Guardian, et al have had them for a while.

How long did the feeds take to put together? Not long at all, according to Neil:

approx. half a day for the first feed and maybe 30 minutes for each subsequent feed. An RSS feed is a simple XML document, which is native feature of the .NET framework that we use for all server side programming across the site. So from a coding perspective it is quite straightforward.

What's your CMS?

We don't have one! Well, not one to drive the entire site. Most of the data-driven features currently use an MS Access back-end with an ASP .NET front-end. We are currently developing our own CMS, initially for our Intranet but ultimately it's likely to drive the Website as well. That uses a MS SQL Server 2000 back-end.

Would you consider making other parts of the db available to the feed, such as on-line collections or listings?

Yes: we already provide listings feeds but we'd like to more of our content to be aggregated.

We can't provide collections feeds just yet as that part of the site isn't data-driven. When it is, it would also be possible to provide customized feeds so that content aggregators could choose which areas of the collection they were interested in and a feed be generated to suit they're specific needs. Ultimately I'd like to see all of our content available as feeds but this will only be possible once the site is run via a robust CMS.

How popular are the feeds? 

National Museums Liverpool quite rightly admits that feed figures are hard to gauge. Neil Harding concurs:

We cannot really give you an exact figure, but doing our best to exclude search spiders and robots, our January 2005 WebTrends (http://www.Webtrends.com/) reports are telling us this:

 1036 Visits

34 Unique visitors

7752 Views

(Period of time measured, one month.)

Neil again rightly points out the problems of measuring RSS output and usage.

RSS feeds, by their very nature, are there to be used in various ways: the stats are not a definitive impression.

So if someone consumes our feed as content on their Website, they could, if they want, cache it so it only fetches the latest version once a day. Therefore we'd only see one view per day from them, but that site itself could be generating lots of traffic for us. It's just that when a user visits that site they are only seeing a local version of the feed.

It's the same with desktop news aggregators (feed readers). The one I've got on my machine can update at intervals ranging from every minute to every week, so the feed can appear more popular if someone is checking regularly.

And that's exactly the situation that users of Plucker from Pluck.com could be creating if they set their popular new feed reader download to check for new content every few minutes by default! 

Would National Museums Liverpool do it again?  

"It's a starting point for us," said Dawn Capper. NML don't currently have easy access to their collections on-line, so making more content types available via XML and RSS is something that's possibly out of reach at the moment. "It's certainly something we'd be interested in working towards, though. The feeds maximize the chances we have to reach diverse audiences with our collections."

Where Now for Museums and RSS?

Mainstream publishers are beginning to harness the technology to their own ends. Spring 2005 is an exciting time for followers of technologies like RSS, XML and SOAP. There's a sudden momentum behind developments that was absent before the end of 2004. In the US and in Europe, the bigger technology portals and other mass Web publishers like the BBC, MSN and The Guardian newspaper are developing their own bespoke feed readers that will keep readers within Web site hierarchies, rather than 'doing their own thing' on their own computers with open source feed readers.

Doing this gives the big publishers a chance to retain advertising within the feeds - important at the moment, as there's currently a big rise in revenues from on-line advertising.

Fig 12: The Guardian (uk) bespoke feed reader

So the chances are we'll see a rise in the number of built-in feed readers such as that on CNET.com and the new beta version of Newspoint on The Guardian (UK). Simon Waldman is the Director of Digital Publishing of The Guardian, one of the world's most Web savvy newspapers. Speaking to the Editor's Weblog (http://www.editorsWeblog.org/2005/02/emthe_guardiane.html) he explained where he thinks the rationale for adding customized RSS access comes from.

Newspoint (The Guardian's bespoke, browser-dwelling reader) is just one of a number of things we're testing and examining at the moment in the RSS area. Actually, our thinking is not simply driven by RSS, but about the whole principle of the distributed/ decentralized Internet which we think is going to be one of the main developments and challenges that face publishers over the next three years . . .

The online world is changing. We are entering a world where people are increasingly going to access our content through RSS, aggregators, bloggers and search engines. It is the distributed Internet and it is a real challenge for us. The more we learn about it; and the more we engage with it now, the healthier we will be in the medium and long term future. 

(Simon Waldman, Director of Digital Publishing, The Guardian)

By the end of 2004, blogs had established themselves as key parts of on-line culture. Two surveys by the Pew Internet & American Life Project (http://www.pewinternet.org/PPF/r/148/report_display.asp) established new contours for the blogosphere: 8 million American adults say they have created blogs; blog readership jumped 58% in 2004 and now stands at 27% of Internet users; 5% of Internet users say they use RSS aggregators or XML readers to get the news and other information delivered from blogs and content-rich Web sites as it is posted on-line; and 12% of Internet users have posted comments or other material on blogs.

What Does This Progress Mean For Museums? How MLA May Use RSS

If RSS and XML technology become more widely adopted as a means of distributing content and sharing databases in the cultural sector, we can expect to see more initiatives from central coordinating organizations such as MLA (The Museums, Libraries and Archives Council) in the UK.

MLA (www.mla.gov.uk) could adopt RSS or SOAP (Simple Object Access Protocol – see this: (http://www.w3.org/TR/soap/) in a new project, the Case Studies Database, which was the subject of an invitation to tender at the end of 2004. The ITT asked interested developers to provide alternative means of output for the project, via RSS or SOAP.

Lots of people want THEIR view of case studies.  Just their subject, geographic area, etc. RSS would do news alerts on the latest case study for third party Web sites.

SOAP etc. will mean if you search the collections advisory site then you will find relevant case studies too. It will be all in the same system but could be integrated into other portals. So the Scottish Museums Council Website could show just Scottish case studies, or those also from England.  

(David Dawson. Senior Advisor in ICT at MLA.)

On-line information services consultancy Simulacra (http://www.simulacra.com/) has won the contract to develop the project for MLA.  "The Case Studies Service will be an online knowledge base - a collection of the experiences that museums, libraries and archives have gone through in their respective fields," according to Jeremy Tarling of Simulacra.

A Web front end is being developed to allow any museum, library or archive to be able to log in and upload their case studies. The nine regional MLACs (Museum Library and Archive Councils) will all be able to review and publish these case studies, either with their own regional branding or with central MLA branding. MLA will manage the users, workflows, securities, etc. The project will make use of an on-line publishing solution called Harmonise, the code of which was recently released by Simulacra under an open source license.

"MLA specifically asked for RSS output in their ITT, but Harmonise is equally able to provide SOAP input/output should that be required," said Jeremy Tarling.  According to Chris Turner, ICT advisor from MLA, the Case Studies project was conceived as a 'service' and as such the system is intended to be able to deliver content through a variety of mechanisms. He qualifies the author's implication that RSS will be central to the project in this way:

The obvious mechanism for delivery is the basic Web search interface, which will be part of the MLA Web site. However the system is also specified to serve data through SOAP, OAI and RSS so that third parties can access the data from their own applications and present that as an integrated part of their own content.  RSS is simply seen as one of the mechanisms to achieve this and I'm not aware of any specific delivery option envisaged. 

(Chris Turner)

Find out more about Simulacra's Harmonise here: http://openharmonise.sourceforge.net

More Museum Sector Interest

Paul Miller, Director of the Common Information Environment project based in Hull, UK, suggested wider uses of XML and RSS to the paper author. The CIE Web site (http://www.common-info.org.uk/) is RSS enabled - it broadcasts two sections of the site as Web logs for other site users to display or for individuals to read through feed reader software.

Writing to the author about the 24 HM tunable feed, Miller asks:

What scope might there be for sucking content in from other sources: so that a regional museum agency (known in UK museum sector-speak as an 'MLAC') could display on its Website, archive and library content for their region alongside the museum content?

Beyond promoting new ways of accessing material for organizing bodies, there may be problems ahead for smaller organizations wishing to adopt XML output, thinks Miller.

I guess one issue for smaller institutions might be bandwidth; especially if they get to be popular. There may be scope for 24 Hour Museum or mda (http://www.mda.org.uk) or whoever, providing a Bloglines-like (http://www.bloglines.com) capability for these institutions, hosted on YOUR server, using YOUR bandwidth, but carrying THEIR branding...

Research Shows Museums Are Trusted On-line: We Need To Seize The Moment

Recent research (http://www.common-info.org.uk/docs/mori-report.pdf) by Mori, commissioned by CIE, hints that there is a need for the museum sector to adopt RSS-based technology more fully than at present. In the CIE report, published February 2, 2005, 85 % of current UK Internet users said their favored method of searching the Internet was by search engine. This means that for many, the first contact with a Web-borne resource will be out of the searched-for digital object's original published context.

All the more reason to consider carefully design and navigation issues when feeding out content via RSS. But also good reason to get the content out into the digital world using XML or RSS, where people are searching via Google and other means.

The CIE report also highlights good news for museums as information providers. Good layout and professional design values are key to many users, but for a massive 92% of users, reliability of content was highly important. Respondents to the survey also cited quality and frequency of updating as being important.

Trust is key to the way people relate to on-line information sources, according to the report. Some types of organizations are trusted more than others - museums, libraries and archives were regarded as engendering a great deal of trust from many Web users, particularly compared to commercial Web info or retail sites.

The CIE Report in its conclusions hits the nail on the head with a key recommendation - which has clear implications when museums or galleries are considering using new means to broadcast content such as RSS. Bearing in mind the way most Web users use search engines, museums 'need to work to ensure their Websites come up first in searches, have intuitive Website addresses and are generally 'top-of-mind', according to the report.

Beyond the CIE report - DigiCult's Thematic Issue 3, from May 2003, (http://www.digicult.info/downloads/thematic_issue_3_low.pdf) makes fascinating conclusions about the likely growth of the semantic internet and how it relates to the cultural Web sector. Could the current chain reaction of RSS popularity be the embodiment of Tim Berners Lees' ideas about the semantic Internet? (Berners Lee, 2000)

The DigiCult report concludes:

Ultimately the same factors that constrain the heritage sector's ability to take full advantage of the Web will constrain the penetration and pervasiveness of the semantic Web in the heritage sector. The success of the semantic Web in the heritage sector depends upon it's adopting an XML-based approach and a significant experiment that demonstrates it's benefits to the wider community. Even for all it's weaknesses the semantic Web offers a tantalizing solution to the problem of information overload created by the Web and the heritage sector needs to address how it can take advantage of the opportunities it offers.

(DigiCult, Thematic Issue 3, May 2003.

24 Hour Museum - Developing RSS Further

Taking cues from the ideas of Simon Waldman and considering the likely moves of major Web broadcasters towards browser integrated feed readers, it seems sensible to consider developing some dedicated feed reading technology, aimed at specific audience strands such as (UK National Curriculum) Key Stage 2 or 3 kids. Imagine a free and funky dedicated reader, popping up on your kid's screen at home or at school, displaying museum content from a limited number of safety-checked, well-designed advert-free digital sources.

You'd click on a button on the 24 HM Web site or the kids site www.show.me.uk to download the free reader which could be based on an open source program such as Feedreader from Sourceforge.com. (http://www.sourceforge.com)

The reader would be colorful and capable of searching for and book marking a number of curriculum or fun-related subject feeds like dinosaurs or archaeology. Taking the idea further, 24 HM could offer a series of facilities for the museum sector, such as the ability to generate and populate your own museum-specific listings and events feed. Using the 24 HM's existing Direct Data Entry system (http://www.24hourmuseum.org.uk/etc/formuseums/TXT17733_gfx_en.html) as an entry point, event, exhibition and educational resources could be inputted. Then a JavaScript code could be automatically generated, giving the Web master of the museum site some code to paste into an html page, or into a database-driven page. This would then display the museum's own news feed of listings or exhibitions, also available to others to add to their sites, or to individuals to  use locally on their machines.

Since the medium term strategy of the 24 Hour Museum is to develop subject specific areas on the site such as science and technology in museums, archaeology and art collections on-line, it would be natural for each section of the site to be available as a feed. It would be difficult logistically to make enough content to fill say 10 sections of the site, but it would be possible to make partnerships with the major national museums in each subject area to share content development costs, or simply share content to be re-purposed, then fed out using the subject feed.

These are exciting times for museums to be on-line, and RSS and XML are the most exciting of all the new methods of making content and collections available on-line. I hope my paper has shown that mainstream publishers are jumping ahead of the cultural sector in the use of widely-broadcasting Web methods: the challenge now is for the museum and gallery sector to realize the enormous value of these new technologies to truly develop the educative and communicative potential of the Web for museums.        

Acknowledgements

Thanks to Jonathan Drori, Director, Culture Online for kindly agreeing to contribute towards the cost of Jon Pratty's attendance at Museums and the Web 2005. Thanks also to Dr. Andrew Sawyer of MWR Ltd, who also contributed to costs of the paper and conference attendance. 

Particular thanks to Jane Finnis, Director, 24 Hour Museum and Ruth Cobb, Marketing Manager, 24 Hour Museum. Thanks also,  Sacha Varma, SSL Ltd, Mike Stapleton, SSL Ltd (who contributed vital technical info and will also be co-presenting the paper in Vancouver), Ross Parry, University of Leicester, Department of Museums Studies.

References

Anon. Editor's Weblog, 2005. The Guardian tests the distributed / decentralized Internet, says Simon Waldman. (EditorsWeblog.org, pub 2005).  http://www.editorsWeblog.org/2005/02/emthe_guardiane.html

Anon. FaganFinder.com. Explanation of RSS, How you can use and finding RSS Feeds. (Fagan Finder.com, pub. 2003).  http://www.faganfinder.com/search/rss.shtml

Anon. The Knowledge Web, MLA, UK. (.pdf, pub 2004).  http://www.mla.gov.uk/documents/iik_kw.pdf

Barratt, Chris.   Access and standards officer, Culture Online, UK. Email correspondence with J.Pratty, 2005.

Berners-Lee, Tim (2000). Weaving the Web: the original design and ultimate destiny of the World Wide Web. Harper Business

Bowen, Houghton and Bernier. Online museum discussion forums: what do we have? What do we need? In D.Bearman and S.Trant (Eds.) Museums and the Web 2003.  http://www.archimuse.com/mw2003/papers/bowen/bowen.html

Carroll, Dawn.  E-PR officer, National Museums Liverpool. Email correspondence with J.Pratty, 2005.

Cliff, Pete. Systems developer, the Resource Discovery Network. RSS - Sharing Online Content Metadata. (2002) http://www.cultivate-int.org/issue7/rss/

Dawson, David. Senior advisor in ICT, MLA, UK. Email correspondence with J.Pratty, 2005.

Harding, Neil. Webmaster, National Museums Liverpool, UK. Email correspondence with J.Pratty, 2005. 

Johnson, O'Doherty. Vines, Eds. Mori. Understanding The Audience, CIE, 2005. (CIE, pub 2005). http://www.common-info.org.uk/docs/mori-report.pdf

MacLeod, Roddy.  RSS and EEVL: Syndicating industry news and announcements. (Ariadne, pub. 2003) http://www.ariadne.ac.uk/issue37/eevl/

Miller, Paul. Director, Common Information Environment, JISC. Syndicated Content: It's more than just some file formats  (Ariadne, pub. 2003) Last updated, Wednesday, 14-May-2003. http://www.ariadne.ac.uk/issue35/miller/

Moffat, M.  Development Officer, EEVL. RSS.  A primer for content producers and publishers. (EEVL, pub. 2003)  http://www.eevl.ac.uk/rss_primer/#ref4

Moffat, M. EEVL Development Officer. RSS - A Primer for Publishers & Content Providers (EEVL, pub. 2003) http://www.eevl.ac.uk/rss_primer/#top

Moffat, M. EEVL Development Officer. RSS - A Primer for Publishers & Content Providers (EEVL, pub. 2003)  http://www.eevl.ac.uk/rss_primer/#top

Morley, James. Web site manager, Kew Gardens, UK. Email correspondence with J.Pratty, 2005.

Nottingham, M. RSS Tutorial For Content Publishers and Webmasters. Nottingham, M. (2002) Last updated Feb 16, 2004. http://www.mnot.net/rss/tutorial/

Pratty, J. (2004). The 24 Hour Museum RSS Newsfeed. in D. Bearman and J. Trant (eds.) Museums and the Web 2004: Proceedings. Toronto: Archives & Museum Informatics. http://www.archimuse.com/mw2004/papers/pratty/pratty.html

Prudames, D. Deputy Editor, 24 Hour Museum. Student Journalist Network To Showcase Regional Museums. (24 Hour Museum, pub. 2005). http://www.24hourmuseum.org.uk/nwh_gfx_en/ART26114.html

Rainie, Horrigan, 2005. A decade of adoption – how the Internet has woven itself into American Life, Pew Internet & American Life Project (Pew Internet & American Life Project, pub 2005). http://www.pewinternet.org/PPF/r/148/report_display.asp

Tarling, Jeremy.  Simulacra Media, UK. Email correspondence with J.Pratty, 2005.

Turner, Chris.  ICT officer, MLA, UK. Email correspondence with J. Pratty, 2005.

Cite as:

Pratty, J., The 24 Hour Museum Tunable RSS News Feed, in J. Trant and D. Bearman (eds.). Museums and the Web 2005: Proceedings, Toronto: Archives & Museum Informatics, published March 31, 2005 at http://www.archimuse.com/mw2005/papers/pratty/pratty.html