The internet is quickly becoming a taken-for-granted medium for archaeologists to communicate with their many stakeholders and constituencies. A vast array of archaeological information is available online, including access to primary excavation data, academic articles, personal blogs, news from around the world, virtual reconstructions, photographs, and movies, and it increases in number and variety every day. Since Carol McDavid’s (1999; 2004) seminal research regarding the Levi Jordan Plantation website, in which she quotes there being between ‘1,060,001 and 1,870,000 Internet web sites that either deal with or are about archaeology’ (a number that she produced by running an online search) there are now over 300,000,000 hits on Google for a simple search for ‘archaeology’.

While McDavid also questioned the utility of a website for public outreach in archaeology, many have embraced an online presence as a major component of their archaeological outreach. Julie Schablitsky maintains that, for archaeological projects, the ‘establishment of a Web site where the public can learn more about the site is a prerequisite before agreeing to share research data with the media’ (Schablitsky and Hetherington 2012: 151). With the increased attention toward Web 2.0 and attendant social media, barriers to participation on the internet have been greatly decreased; whereas early participants on the Internet had to know HTML and other programming languages to make a webpage it is now relatively easy to set up a dedicated blog, image archive, and Facebook page for archaeological projects with a minimum of technological savvy. We are increasingly being inundated with requests to ‘like’ projects on Facebook, follow archaeological conference Twitter feeds, and fund new projects on Kickstarter. At this volume of noise it seems impossible to imagine how ephemeral the archaeological presence is on the internet.

It is a rude surprise when it happens; John Hawks (2012) tweeted, ‘The real tragedy: a clearly well-funded and rich (online) resource now is represented only by an archive page with hit counter locked at 7451’ after finding out about the closure of an online atlas of embryonic brain models (Figure 1). Even ‘well-funded’ websites fall prey to budget shortfalls, changes in technology, and changing priorities within archaeological projects. A return to Carol McDavid’s Levi Jordan Plantation website reveals that it has not fared well in the years since it has been built. While the content is still all in place, the website is not being updated and some of the biographical information is consequently incorrect. Julian Richards (2002) found a similar case with archived digital data, that 3D models, databases, and CAD files created by digital archaeological practice must be actively maintained to avoid obsolescence. Cornelius Holtorf’s ground-breaking thesis written in HTML (2000-2008) contained a number of hyperlinks that have maintained their internal integrity as they are entirely self-referential, yet the subsequent Internet Archaeology (1999) article links to outside references that have long since disappeared.

The most illustrative case of the fragility of online archaeology came in April 2009, when the company Yahoo! announced they would be shutting down their United States-based Geocities webpages. Geocities was once the most common hosting service for low-cost personal webpages, including hundreds of public outreach sites about archaeology. In this paper we consider the fate of these archaeology-based Geocities websites to illuminate the benefits and risks of using commercial infrastructure for online archaeological outreach. In doing so we first provide the background of Geocities and what led to its demise, then we discuss our research methodology, consider the results of our research, then finally assess the current state of archaeology on the internet.

The context surrounding the wide adoption of Geocities for archaeology public outreach sites merits investigation. Geocities was first launched by David Behnett and John Rezner as BHI (Beverley Hills Internet) in 1994, changing its name shortly afterwards in late 1995. It allowed the creation of a free personal online homepage without the need for knowledge of a programming language such as HTML (Gill 2004). The service was enthusiastically embraced by novice web authors, although, as Gill notes, the websites Geocities hosted were often abandoned when the novelty faded. Although the contrast has been drawn between the static websites of the 1990s and the community-building internet of the Web 2.0 era (e.g. Richardson 2013: 4), this is an over-simplification, as through message boards, guestbooks and webrings, many of these websites presented an attempt at fostering an online community. Indeed Papacharissi (2002, 644) recognised homepages as setting up ‘a virtual meeting point… [they] also provide a virtual home base for online communities and help connect their members’.

Geocities was a ‘mega-community’ (Dahlberg 2001: 617), with sites (called ‘homesteads’) arranged into clusters modelled on city streets, with broad thematic similarities (sites about entertainment, sites about technology, etc.) grouped together in ‘neighbourhoods’. Upon reaching a neighbourhood, visitors would be faced with a page containing icons of numbered blocks containing small images of houses. Clicking on a block would reveal a page with actual addresses and icons separated into two columns by a winding road (Papacharissi 2002: 648). Figure 2 shows the Geocities front page in 1996, highlighting a ‘featured homestead’. In practice, however, the hosted sites were often self-contained, or part of ‘webrings’ extending outside of the Geocities domain, rather than being oriented towards an intra-Geocities community. In a primer to the concept of ‘Web 2.0’, posted in 2006, Madden and Fox (2006: 6) contrast the metaphor of a place used by (then waning in popularity) Geocities, with the metaphor of a person deployed by the then-ascendant MySpace. Geocities was, however, already shedding the geographical metaphors: soon after Yahoo! took control of the site in 1999, users were offered personal vanity URLs in addition to their ‘street’ addresses. Yet the metaphor of place and ‘community’ may have influenced subsequent interaction and self-presentation on Geocities. Papacharissi (2002: 655) found that Geocities members tended to display much more of a sense of community than those whose sites were hosted with Earthlink (an internet service provider that makes free web space available to members).

Although it remained popular, Geocities had reported losses since the Yahoo! takeover, and in April 2009, Yahoo! announced that it was closing the service. With the exclusion of Japan, (the Japanese-language geocities.jp still offers free homepages) all other webpages hosted by Geocities went offline on 27 October, 2009. At the time of the initial announcement regarding the closure, a Google search for ‘archaeology’ within the domain geocities.com yielded more than 10,000 results. As with the numbers cited in our introduction, in practice these numbers were inflated; many of the hits for archaeology were different pages within a significantly lower number of cohesive sites about archaeology. The first 88 individual sites encountered within the search results were catalogued as part of this survey, with brief notes made on whether or not the site was judged to be ‘active’ (updated since 1 January, 2008); whether it was a personal or corporate homepage; the kind of content it contained, for example whether it was a thematic compendium of links to other sites, or a collection of essays or photographs, and whether it had a region or period-specific focus; as well as the country in which the site was created. Of these 88 websites, 14 were still active, and two redirected to other websites. The content of the 88 sites visited ranged from personal homepages of academic researchers, to local and student archaeological societies; an apparently long-defunct British campaign group (‘Archaeologists and Development’, see Figure 3), to personal collections of essays on more esoteric and contentious topics like King Arthur, aligned sites in Glasgow, and Black Athena. Brief notes about the 88 websites are presented as Appendix 1.

One of the sites related to a specific event, the 7^th Gender and Archaeology Conference held at Sonoma State University in October 2002, and may never have been intended to form a lasting archive. Two others were the websites of student archaeological societies that may only have been planned to last for a single iteration of the committee (although that of Oxford University Archaeological Society was still active at the time of the survey- it had only been established in 2008). The majority, however, contained essays, photographs and other interpretative or reference materials. It is possible that no thought was ever given to the longevity of the websites by their creators.

After consideration of these websites, a short survey was devised for further information about the websites. In April 2009, a link to this survey was sent to the webmasters of the 58 sites for which contact details could be found. Of these 58 emails to webmasters, 18 of the e-mails were returned undelivered, nine of the site owners completed the survey, and one further webmaster sent a personal response politely declining to answer the questions. The questions are reproduced in Appendix 2.

While just 10 per cent of the webmasters queried responded to the survey, their answers were still informative regarding what the closure of Geocities meant to their websites. All of the respondents identified themselves as sole authors of their sites, although one did say he was ‘sole maintainer on behalf of a group’, and another the sole caretaker of a site whose author had passed away. Six of the sites were described as providing information and sharing knowledge, two provided contact details and membership communications for archaeological societies, and one was a directory of web links. Four of the sites were based in the USA (34 of the 88 sites visited were American), two in the UK, one in Germany, and one in Georgia. Three of the webmasters described their sites as no longer updated. Two described their content as primary data that was not available elsewhere, and two replied that their sites contained interpretations not available elsewhere. After the closure of Geocities, six of the respondents said that they would be moving their sites elsewhere, the remaining three were unsure. Seven respondents said that they were aware of published references to their Geocities site, and six said that they were aware of incoming links to their sites online.

There was also a free text portion of the survey wherein webmasters could ‘make comments about the closure of Geocities, your experience of running an archaeology-related site on Geocities, or your thoughts about archaeology’s place on the internet in general.’ The testimonials that the website owners left told the histories of many of these websites, and related the marginal space that they suddenly found themselves occupying. For example, one of the archaeology societies were hosting a website on Geocities as their local authority stopped hosting the page on the governmental website, and after changing all of the links in other websites and publications from the former URL, would have to change them once again on finding another host. The website owners expressed sadness and frustration and lamented the time it would take to build the websites elsewhere. Interestingly, while there is this aforementioned loss of information and wasted time and effort after using Geocities, several of the respondents commented on the utility of the internet in general to disseminate research results to the public and to the academic community. Further, ‘it is a good way to enhance communication between scholars in archaeology and cognate disciplines as well as between archaeologists and the general public.’ Two respondents specifically wished for permanent repositories for archaeological information and for personal webpages for archaeologists.

Three years after the initial closure and survey data were collected, we are happy to report that many of these websites have been rebuilt elsewhere on the internet. One of the responding sites still exists at the same URL, albeit hosted elsewhere. Another has been transferred to Google Pages, another free hosting service. The author of another site still maintains an active online presence, embracing Twitter and using Academia.edu to self-archive his work. Of the sites that did not respond to the survey, 14 have moved to new hosts (in fact, eight had already moved at the time of the survey), while a further six of the authors still maintain an online presence.

There have been several initiatives to preserve these webpages by mirroring the Geocities content. These include geocities.ws, oocities, geociti.es, and a few others. The Internet Archive, established in 1996, is a non-profit digital library that offers permanent storage and free public access to its archives; it has a service called the ‘Wayback Machine’ that provides snapshots of internet content sorted by year. The webpages that were not rebuilt can be accessed (albeit often without the original media or links). Incidentally, at the time of this initial review, the Internet Archive was not available online (25 October, 2012).

Sorting through the wreckage of the closure of Geocities is reminiscent of the current mode of curation of old computers, data formats, and software, confusingly also called ‘digital archaeology’. We encourage this confusion; a contemporary archaeology of the recent digital past is welcome, especially as curated data formats can help us retrieve information that would be otherwise out-dated and locked away. In an ideal world, this other ‘digital archaeology’ would be informed by archaeology’s engagement with materiality and our profession’s attention to standardized recording. This survey of Geocities archaeology websites is an illustration of the affordances of digital media as outlined in Manovich’s principles of New Media (2001). The websites are represented numerically, composed of digital code and are subject to algorithmic manipulation, thus allowing automated archival and redisplay on other hosts. These websites have thus been reproduced and exist as several different versions, showing their lack of fixity. That the websites have not entirely disappeared can be considered a result of strong conservation strategies that rely on the affordances of digital media. Developing a stronger sense of the context and ‘depositional processes’ that digital media are subject to would be a productive venue for further interdisciplinary research. Beyond this potential collaboration with archivists and new media theorists, there is also an opportunity to engage with a larger audience; as the museum exhibits dedicated to older computers have shown, people are open to the idea of a ‘digital archaeology’ that reminds them of the computers of their younger years.

Our examination of the fate of archaeological websites on Geocities has been instructive regarding the risks of hosting crucial archaeological data online. In the words of one of the reviewers of this article, it is akin to building critical infrastructure on a fault line; there should be some degree of earthquake preparedness for a lasting digital presence. Investing a large amount of time and money in interpretive projects online may not ultimately be the best investment in outreach time and money. Websites that are dedicated to one topic or one excavation are reduced to isolated nodes; as shown by the subsequent profusion of Geocities sites to other hosts, to preserve archaeological information it is imperative to distribute data as widely as possible. Using the previously mentioned affordances of digital media including modularity, variability, and the ability for algorithmic manipulation (Manovich 2001) enable a better approach: an archaeological website used as an organizing structure to gather information coming from other feeds, to capture tweets, images, and blog posts in a single place, rather than attempt to host all such content at a single point of potential failure. If the website must host all content in a single place, then creating a PDF from a series of screenshots of your website might be one of the single best ways to insure the longevity of your web presence; while this circumvents the utility of hyperlinked content, the website structure and cohesiveness will be apparent, even after another Firefox, Google Chrome, Internet Explorer or Safari upgrade breaks your entire website for a large percentage of the population. Abundant advice now exists for digital preservation [see for example the Preservation Handbook (Digital Preservation Commission 2002-2009)]. Further, entrusting private companies to host vital information is risky; many archaeological sites now have a presence on Facebook and Twitter. While the demise of these particular social media sites may not be imminent, access to the account associated with the archaeological presence is not perpetually ensured. The closure of Geocities in 2009 was followed in 2010 by the decision of Ning, another site with a burgeoning archaeological presence, to discontinue free hosting, and the ensuing closure of a modest number of archaeology-related social networks (Law 2011: 13).

Yet all is not dark, as Julian Richards has noted, the digital age ‘provides archaeology with both a crisis and an opportunity’ (2002: 33). Many of the sites that were formerly hosted on Geocities were preserved through one of the various previously mentioned schemes. The webmasters who were losing their sites still remained positive about the potential for online outreach. While we certainly should not rely on the archivist vagaries of internet citizens to protect our data, we should encourage the idea that archaeology is important and accessible by continuing to engage with the online audience.

This article provides a case study to better understand the long-term presence of archaeological research online and the benefits and risks associated with using ‘free’ services hosted by corporations. As websites such as McDavid’s near the end of their second decade online, we can begin to take a longer view of the survivability of the archaeological web presence. Drawing from this study of Geocities sites we can come to a few conclusions and recommendations for future directions. The lessons learned for archaeologists interested in online public outreach from the demise of Geocities are: diversify your content by hosting in more than one place, do not rely on public or private companies for hosting in perpetuity, and most importantly, to think of alternative, creative places to perform outreach online.

This last point, to think of alternative places to perform outreach online, deserves elaboration. Contributing to the ongoing effort put into Wikipedia by updating, editing, or adding information about your site, region, time period, or artefacts can ultimately be a greater contribution to knowledge about the past than a stand-alone website. Being available to large online communities such as Reddit, where experts answer questions that people have about history and archaeology by hosting an ‘Ask me Anything’ session allows a direct connection to an interested online public. During the Morning Star excavation in 2013, John Hawkes related updates through images and video shared on Twitter, and would answer questions about the process, providing unprecedented access to a paleoanthropological excavation. This kind of archaeological outreach provides an excellent supplement to more traditional websites that are often built and then abandoned. Creatively diversifying online public outreach by distributing quality archaeological content does not necessarily ensure the longevity of online content, but can address a larger audience without the intensive investment in a large, elaborate, purpose-built website.