Why do We Need an Updating System?

You, the Community, have proposed a number of arguments for introducing an updating scheme.

• We want people to trust the material on h2g2, and that trust will decrease markedly if people decide that entries in the Edited Guide are out of date. This applies across a range of topics. Restaurants, computer games, singers who release new material - if we're aiming for a dynamic, living guide, then these are areas that are really going to show us up.

• Many of the older entries simply aren't up to our current standards. Liver for example. Anyone who put that in Peer Review wouldn't stand a chance these days. We could appeal for new material for these entries in either a Flea Market style forum, or via an 'Update of the week'.

• Many h2g2 users are fixated on one topic or another, and are understandibly disappointed if the entry on their 'thing' doesn't include all the info they think it should. It must be disheartening for these Researchers to hear that their new and better entry on a subject won't be included in the Edited Guide, because we've already got an Edited Entry on that topic. We should be benefiting from their knowledge by giving them the facility to add their bit.

• Many of the early entries covered very generic topics Lycra for example, and a number of others (Newspapers, Mathematics, Judo) which should be definitive. We could put these entries out to tender - recycle them as 'Topic of the Week' spots, to glean the Community's collective wisdom on them.

• There are many topics on which Researchers either aren't writing or aren't submitting to PR, because they think the entry would become outdated too quickly in the Edited Guide.

• Better links within older entries encourage people (and especially new users) to explore h2g2 more, especially as the older entries tend to be on topics that will be among the first a newbie would enter in the search. For example, Australia was first stamped official on 17 April, 2000 - the latest thread on it is dated 30 January 2002. It has the air of tumbleweed about it - the author is long gone, as is the Editor, and the links are similarly elderly.

• There is also a historical reason for introducing an updating scheme. Updating relates closely to Douglas Adams' dream of being anywhere in the world, sitting down at a café, getting out a handheld device and submitting copy to The Hitchhiker's Guide to the Galaxy:

I did have the inkling of an idea that a collaborative guide, one that was written and kept up to date by the people who used it, in real time, might be a neat idea.

Spring Clean

The bottom line of introducing an updating scheme is that it would revive and spruce up older areas of the Guide making entries more useful, practical and reliable. We will be reviving interest of a number of absent Researchers by informing them, via email, that their entry has been updated. And by putting 'Recently updated entries' in a section on the front page, we'd be giving the Community another area of h2g2 to look at. Specialists will also be relieved to find that there is an outlet for them to get creative on already-established topics.

The Current Situation

At the moment there is an informal updating system in place. Currently we update entries in a number of ways:

• Threads are checked the day after an Entry has been up on the front page and if any inaccuracies crop up in there, we'll correct them usually within 24 hours.

• Broken links, typos and inaccuracies are pointed out in Editorial Feedback. All members of the Editorial team are subscribed to this page, so the turnaround time on these quick changes is short, within 24 hours on a working day.

• Authors or Researchers with an interest in a topic contact Subs on their Personal Space with updates and inaccuracies. The Sub then contacts a member of the Editorial team either by email or their Personal Space. In an ideal world Subs should subscribe to all entries they have edited, so that any conversations on a subject are immediately brought to their attention on their personal conversation lists. However, Subs do not remain Subs forever, and even those who are in the job long term may not get round to updating, as it's not the primary role of the job.

• Two Researchers, Jimi X and Mikey the Humming Mouse, are volunteer Sub-editors and have taken it upon themselves to update entries. Jimi and Mikey approach updating in two ways: Researchers can either propose candidates for updating at Update Headquarters which Jimi deals with, while Mikey is working through the entries in chronological order.

That's four different methods of getting changes made, which is clearly more confusing than it needs to be, as well as being an incredibly lengthy process (especially in the case of substantial changes). In the Edited section of the Guide, we've only got just over 4000 items to keep 'alive' which is more attainable than the millions of pages associated with Google, say. Giving Researchers one accessible area to tell us about changes that should be made would solve this problem. An Update Nominations page would achieve this, similar in look and feel to Peer Review.

Update Nominations Page

Originally we came up with the idea that volunteers could go through all the entries from the earliest to the most recent looking at attached threads, adding information from them. This ensures that all entries are covered chronologically, it's democratic and fair. However, the older an entry is, the more likely it is that the threads are potentially out of date, too. In addition the author is likely to be long gone, and it would be preferable to serve the interests of our current audience.

An Update Nominations page would allow the Community to come up with the best candidates for updating. The scheme would be administered by a group of volunteers, who, like Scouts, would 'recommend' entries to be updated depending upon the feedback in the threads. The criteria for recommending an entry would be:

• Are facts incorrect in an entry?
• Is the entry out of date?
• Is there newer, better information in the threads that should be included?

That way the Community and the volunteers can prioritise which entries need updating the most, and certain entries can be modified again when needed (which wouldn't happen if we took a strictly chronological approach). The Community could inform us of very simple changes such as a new album released by a band since the entry was originally edited. The Community would also have the opportunity to contribute more facts to the entry via the threads as long as the entry was active on the Update Nominations page. The details could be thrashed out among the Community, which benefits the Guide since the more accurate an entry is second time around, the more polished the Guide appears to be.

In addition, it would be possible to do some clever calculations to rate the urgency of updating, to aid the volunteers in picking entries. For example, we could take into consideration:

• The number of Postings added to its Forum since the last updated date.
• The number of new Conversations created since the last updated date.
• How long ago it was last updated.

From this information it would be reasonably straightforward to create a guideline statistic that would determine how much an entry would need to be updated, which could form other indicators, helping volunteers in picking entries from the Update Nominations page. Though if the number of postings posted to an entry is going to indicate the urgency of the update, it does open us up to spamming in the threads.

The role of volunteer Updaters would be two-fold: as well as picking entries to be updated, they would, like Subs, be given a batch of entries to work on. The 'Updaters' job would entail adding in new Researchers, adding in new and relevant links, and trawling through the threads and adding in any salient facts.

How Entries be Displayed on the Update Nominations Page

As with PR, the whole Community would contribute to entries on the Update Nominations Page. In fact, once they've been posted up there, it puts the subject into the public arena, and it would elicit more material to go into the entry anew.

Each entry would have the following information attached to it:

• Entry ID, Subject, Date Entered and Last Posted, as for PR
• Date Created
• Last Updated
• Proposed By (as this person could have posted some substantive new information in there)

We would only allow entries onto the Update Nomination page that had had postings made to them since the last updated date.

Volunteer Updaters

In summary, the system would work like this:

• The Update Nominations page works like Peer Review.
• The Community puts Edited Entries into the Update Nominations page, where they are discussed.
• The Updaters pick from the UN page, like Scouts picking from PR.
• These entries go back out to the Updaters for updating, like entries go to Subs for editing.

If we were to have, say, five updates a week going up on the front page, that's 20 entries every four weeks. Allowing one entry in every 21 as natural slippage, we would need a group of seven volunteers picking/processing three entries every four weeks.

Updaters could have had experience of being a Scout or a Sub, but not necessarily. The job could very well stand alone, independent of other schemes.

• We would only ask volunteers to update if they really wanted to do it. Some Researchers have a particular passion for updating, and if that's the case, they are welcome become an Updater and use their enthusiasm.

• It's another badge and widens the choice of the sort of official activity the Community can become involved with.

Problems

There are however some inherent problems with updating. Simply harvesting through the threads could be difficult. The threads do contain some excellent information, but much of it needs clarification and with the author long gone, this could be problematic. Hopefully though points would be clarified in discussions on the Update Nominations page. Also, sometimes it's just not obvious where a new snippet of information should be added. Someone could have posted 'In Holland Calvin and Hobbes are called Casper en Hobbes'. It's an interesting, relevant fact, but it needs some context. To address this, we could create a new heading at the end entitled 'Other facts about Calvin and Hobbes', which would be easy for the Updater to do.

How the Project Could be Handled Inhouse

Administering a new volunteer scheme would involve more work for the inhouse team. Once entries come back from volunteer Updaters, changes would have to be checked by an inhouse member of staff. This is a requirement of all Edited Guide entries. This would be a very short job if the Updater had simply included a new album release or new book, but the way the tools are set up, the staff member would still have to read the whole entry of up to 5,000 words.

Inhouse members of staff would also flag recently updated entries on the front page. There would be a new section of, say, the latest five entries updated that week. This would recycle the content, drawing attention to the great work of the volunteers and Researchers who've worked on reviving an entry. It would also engender a sense of nostalgia, with long-time Researchers reminiscing about when they first saw an entry go up on the front page. An odd concept, perhaps, for the Internet which is still considered to be something so new, but giving h2g2 a sense of history, builds upon the 'sociology' of the Community and reinforces its identity.

Tools Requirements

A series of new tools is at the core of making an Updating scheme viable. Currently, we don't have specialist tools and consequently large-scale overhaul updating that we do (as opposed to quick tweaks) is time-consuming and informal.

Essentially tools should enable Entries to be:

• Proposed on an Update Nominations page
• Discussed by the Community
• Picked by an Updater
• Sent out to an Updater
• Looked at in house efficiently

Considerations

We've already got the architecture for an Update Nominations page and updating system. Included in this scheme should be a mechanism where the original author(s) would be notified via email that someone has submitted their Edited Entry for an Update, thus reviving their interest in h2g2.

As with getting Edited Entries into the Guide, it would be important to retain copies to refer back to:

• Of the original entry
• The original Edited entry
• The new Edited entry

As Edited Guide entries become older, as each is modified a number of times, more and more Researchers will have contributed to it. For example Subs who have previously worked on an entry can be added to the Researcher list, but they didn't write the piece and didn't research it. So, strictly speaking adding them to the Researcher list is inaccurate.

To get around these problems, it is important to implement a proper superseding system. This would work by creating an edit 'crumb trail' for each Edited Entry, so that if entry A70091 has been updated and a new version found at A200034, any incidences of A70091 in the Guide would automatically take you to the latest update at A200034. In turn, when A200034 is updated, any links to A200034 would take you to the latest version. The most recently modified version would always come up higher in the search engine. You could mine into older versions, too, or link directly to specific versions. Appropriate subs and authors would also be credited, creating a fair credit list.

To expand upon this, we could add a 'Previously edited by' box, under 'Written and Researched by' and 'Edited by', but adding in extra lists in the margin will begin to make the page look cluttered. Instead a simple 'View History' link would show:

• Date entry was created
• Scout who picked the entry
• Sub who subbed the entry the first time
• Nominated for updating by
• Updated by
• Date modified officially again

Given the 'crumb trail' superseding system two Entries would be possible, which would get around the problem of inhouse staff having to read the whole entry. When an Updater has finished with an Entry, new text, or text that has changed from the original could be highlighted in a different colour in the Entry Editor to make it easy for the inhouse editor to check changes. Time-wise this would be a much smaller/shorter task than re-checking the entire entry. The highlighted changes could also be visible to the Community which would show how organic the Guide can be.

It's also worth mentioning here that, potentially, creating a number of entries on (say) Australia would throw up loads of entries on Australia in the search. If we are going to hang onto all the old versions, then we should give them a URL 'dna/h2g2/archive' so the entries would still be in the Guide, but wouldn't show in a normal search. A search archives link could be put onto the search page if Researchers did want to trawl through older versions. The crumb trail should also cover entries that have been put in the categorisation system, so that we don't have to re-categorise any entries that are updated.

For thoughts and ideas that have gone into this scheme, special thanks go to
Jimi X,
Mikey the Humming Mouse,
Pegasus,
John-the-Gardener,
Peregrin,
Trillian's Child,
LÒÒnytunes,
MaW,
Anonymouse and
Axe.