Archival management systems offer a number of advantages, particularly to archives that do not already have large quantities of EAD finding aids or are dissatisfied with current workflows. A primary advantage of archival management systems is the ability to enter data once and generate multiple outputs. Rather than being isolated in separate systems, data can be brought together through a single interface, reducing redundancy and making it easier to find and manage information. Instead of having to understand the intricacies of EAD and XML markup, archivists, paraprofessionals, and student workers can create a valid EAD finding aid by entering information through a series of Web- or desktop-based forms, saving time and producing more consistent finding aids. Some archival management systems also enable organizations to publish their finding aids on the Web, thus making archival information more widely available.

However, archival management systems can be difficult to implement in some organizations and may not provide the flexibility that archivists require. Several archivists reported difficulty importing existing EAD data into systems such as Archon and AT, a problem due in part to the flexibility of EAD and the resulting variability of finding aids. Although archival management systems typically can be customized and feature user-defined fields, they do enforce a certain consistency and workflow, which frustrates archivists who have an established way of working. As one archivist stated, "Archon and Archivists' Toolkit are great, but it means that someone else has done the thinking for you about the workflow." Homegrown approaches may be more flexible and may better reflect the archive's own workflow. Furthermore, some archivists argue that putting archival description into a database structure is reductive and oversimplifies the process of producing a finding aid. In the process of encoding a finding aid, archivists better understand the texture, structure, and contents of the document. Also, XML and word processing editors provide greater flexibility than databases. As an archivist noted, "If we are doing rearranging while we're going along, we can't shift things around very easily if we're using a database. We have parts of finding aids that we can shift around in Word. ...The tool has to combine flexibility with rigor."

Other archivists emphasize the importance of adhering to standards to facilitate exchange of information and consistency. As one user of an archival management system noted, "We could have customized things to meet past practice, but we also decided to move away from old practices. We don't want to be too flexible any more." Katherine Stefko (2007) acknowledges the trade-offs in sacrificing flexibility for consistency: "To use the AT effectively implies a commitment to using current professional standards, and while it would be hard to argue anything other than this being a good thing, it undeniably raises the bar in terms of the time, training, and expertise an archivist needs in order to use it. ... Accordingly, we've redirected staff time and modified our workflow so that more time is now spent accessioning material, with the understanding that retrieval and reporting will [be] easier and reference and administrative work less later on." Indeed, one interviewee argued that the rigor and inflexibility of archival management systems are actually strengths, since by using such software, archives will ultimately produce more consistent data and facilitate the exchange and federation of archival information. If each archive, or even each collection, took its own approach to archival description, creating a federated finding aids repository would be difficult.

Perhaps the most fundamental choice that archives will make is whether to select an open source or a commercial system. Increasingly, governmental and educational organizations are embracing open source software. For instance, the European Commission has endorsed open source software because it offers a greater diversity of solutions, improves the development process through community input, offers faster deployment through customizability, and leads to enhanced technical skills of IT staff (OSOR.EU 2008). According to OSS Watch, a service funded by JISC, open source offers many advantages: it facilitates rapid bug fixing, is typically more secure, enables customization, supports internationalization, and protects against vendor lock-in or the collapse of the vendors (Wilson 2007). In addition, open source software is typically free, flexible, and continually evolving—assuming an active development community (Lakhan and Jhunjhunwala 2008). Open source software is often supported on or portable to a number of platforms (Office of Government Commerce 2002, 3). Although some worry about the sustainability of open source projects, other developers can maintain and enhance the code should the original developer abandon the project; indeed, as Stuart Yeates from the JISC's OSS Watch argues, "Sustainability is an issue for proprietary software as much as for open source software" (Smart 2005). Many believe that open source software is actually more secure than proprietary software, since open source applications can be scrutinized and verified by "many eyes" and security issues can be resolved quickly (Whitlock 2001).

Some institutions, however, lack the technical staff to implement open source software. Others may oppose it because of they fear security risks or high maintenance costs. Implementing open source software can be challenging, particularly if no support is available or if support structures vanish. With commercial software, customers can contact the vendor for training, assistance in importing data, or other services; with open source software, archives often rely on the community for help. Sometimes open source projects are abandoned before reaching fruition (Lakhan and Jhunjhunwala 2008). Documentation of open source applications can be weak (Office of Government Commerce 2002, 4). Although open source software typically is available without licensing fees, significant costs can result from implementing and customizing it at a local institution. Studies comparing the total cost of ownership of open source versus proprietary software have produced conflicting findings. Each organization should consider what it costs to switch software and what the total cost of adopting the software, including staffing and hardware, will be (Ven, Verelst, and Mannaert 2008, 55-56). Organizations should also consider the maturity of the software, including its functionality as well as support, training, and documentation (Wilson 2006).

Some institutions lack the technical infrastructure to install and maintain an archival management system themselves. Many companies will host software for organizations, enabling archives to focus on their core work. In addition to hosting, many companies will assist customers in importing legacy data into the software. Generally, customers who pay a company to host their data reported that there were few technical problems and that the company's servers rarely went down. One archivist felt relieved that a company in another part of the country was hosting and backing up her data, since her institution is in an area vulnerable to hurricanes.

Although hosted solutions offer noteworthy conveniences and efficiencies, one archivist voiced her frustration that she felt that she was in less control of her data and the way they were presented. If the data were hosted locally, she could play around with the user interface rather than having to rely on the company to make requested changes. Indeed, some institutions feel uncomfortable relying on anyone but themselves to curate their data. What will happen to an archive's data if the company fails? How will the archive retrieve that data, and in what format? Archives should also consider the annual costs of a hosted solution, although hosting data locally also entails costs in hardware, technical support, licensing fees, etc. Commercial vendors typically provide hosting services, although some service bureaus will also host open source software (for instance, hosting is being planned for ICA-AToM). If organizations are considering a hosted solution because they fear the complexity of installing and maintaining software, they should note that most archival management systems are designed to be easy to install and maintain.

For many institutions, cost is a key factor in determining what software to select. The purchase cost for archival management software can range from free (for open source) to hundreds of thousands of dollars (for commercial products with all the bells and whistles and licenses for many clients). Even open source software entails significant costs, including hardware, technical support, and customization—costs that also apply to commercial projects. Along with the cost of the license, archivists should factor in recurring costs, such as maintenance fees, user support, training, hardware, technical support, and customization. Several interviewees noted that companies were willing to "work with us" to find an appropriate cost and that smaller institutions often benefited from a price break. As one might expect, more-expensive products often come with more features. Archives must decide which features are essential.

Software comes and goes, and archivists are rightly concerned about their data being locked into a closed system. If a company collapses or ends support for a product, how will that affect archives who rely on it? Open source projects seem to offer some advantages for sustainability, since other programmers can continue to maintain and develop open source software should the original developer abandon it. However, some open source projects fade away after an initial burst of development activity, and archives, already stretched thin, may not have the technical resources to pick up development work. Nevertheless, open source projects such as AT and ICA-AToM are developing detailed business plans to ensure sustainability, looking at ways to charge fees for training and other services, offer membership, and affiliate with stable organizations that can offer support for the software. Adapting the open source model, some companies allow customers to buy in to escrow plans that will provide them with the code should the company end its support of a product. In any case, to make sure that their data can be used for the long term, archives should make sure that they can easily batch export the data in standard formats.

Inevitably, archivists will run into problems using archival management software, whether because of bugs, difficulty importing data, the need to customize certain features, confusion over how to use the software, or technical problems. Thus, they rely on good customer support from vendors or, in the case of open source software, the developers and user community. Many interviewees mentioned user support as a key factor in their satisfaction with a particular software package. Vendors typically provide assistance via phone or e-mail, user forums, frequently asked questions, and user training. In some cases, help is included in annual maintenance fees, but in others it entails additional costs. Open source projects may seem to be weaker than commercial projects with regard to user support. As one archivist using an open source system commented, "There's no help desk." However, lively communities often form around open source projects and provide support to new users or those experiencing problems. With Archon, CollectiveAccess, and Archivists' Toolkit, archivists noted how responsive the developers are to questions. In addition, support for open source software may be available from consultancies or even the developers themselves. For example, the business plan for ICA AToM includes a provision for "charging a commission for brokering ICA-AtoM technical services between recommended third-party contractors and institutions seeking assistance with ICA-AtoM installation, hosting, customization, new feature development, etc." To evaluate user support, talk to users of different software packages.

To facilitate interoperability and adherence to best practices, archives will want to select software that meets archival standards such as EAD, DACS, and MARC, as well as emerging standards such as EAC. Some archival systems, such as ICA-AToM, focus more on international (ICA) standards rather than on U.S. standards. In the case of archival software developed in Europe, Prom et al. warn that "such tools use a much more rigorous system of classification and provenance than do US repositories" (Prom et al. 2007, 159). However, even many non-U.S. applications support crosswalking between standards and include EAD support.

Some archival management software (such as Archon, CollectiveAccess, and ICA-AToM) is entirely Web based, while other such software requires a desktop client (typically a PC) and connect to a database backend. Web-based software can be more intuitive for some users and enables distributed cataloging, since anyone with Web access can contribute records. With systems such as Archon, information can be published to the Web as soon as it is entered. However, some archives worry about the security and reliability of an entirely Web-based system; one archivist noted her colleagues' reluctance to "put all of our eggs in one basket." If the Internet connection goes down, work stops (which is also true of networked client/server software). A client-based interface may offer greater control over data, but institutions may need to pay a fee for each computer on which the software is installed. Licensing models vary, however, so this is not always the case.

Many archives face difficulty not only in creating EAD files but also in publishing them online. As one archivist remarked, "There's been a big hole—people have been producing EAD for 10 years, but it's still kind of difficult." Some archival management systems address this problem by enabling archives to make available their finding aids on the Web. Indeed, a primary reason that Archon was developed was to facilitate publication of archival information online. Once an archivist enters information into Archon, it is automatically searchable and discoverable by Google (although archives can choose to defer publication of records until they have been approved). Likewise, many commercial systems offer support for online access to their collections, sometimes through the purchase of an additional module. However, some archives already have a mechanism for publishing their finding aids on the Web, so they may prefer software that enables them to easily export finding aids that they can then import into their existing Web-publication system. Since most browsers now provide support for XML, archives could simply upload their EAD files to a Web server, include a call-out to an XSLT stylesheet at the top of each file for the purposes of presentation, and display their finding aids without too much effort. Projects such as the EAD Cookbook have made stylesheets freely available. Although this simple approach does not offer sophisticated searching and other features, it enables archives to publish their finding aids online at minimal cost.

If archival management software does enable publishing archival collections online, archives should consider the quality and customizability of the end-user interface. Does it provide search and browse functions? Can users run advanced searches? Does it offer additional features, such as stored searches? Is the design clean and simple to navigate? Can it be easily customized to reflect the unique identity of the archive? Does the interface meet accessibility standards? Can it be translated into other languages?

In addition to providing access to archival collections, archives may wish to make available digital surrogates of items, such as images, texts, audio files, or video. Many archival management systems offer a "digital library" or "online exhibit" function to provide Web-based access to items in their collections. In evaluating these features, archives should consider what kind of media and metadata formats they support as well as how media are presented. For instance, CollectiveAccess has rich features for media support, including the automatic generation of MP3s upon loading an audio file to the server, an image viewer with pan and zoom, and the ability to mark time codes within video files. However, some archives may want to use a separate digital asset management system (DAM), such as ContentDM, DSpace, or Fedora, to provide online access to their collections, since they are using these robust systems for other digital collections. These institutions will want an easy way to batch export relevant metadata from their archival management system or, even better, a way to plug in their archival management system to their DAM. (ICA-AToM plans to use a plug-in architecture for exposing the application to Web services or allowing it to interface with other Web services, such as DSpace or Fedora.)

Some systems offer robust support for managing archival collections, including appraisals, locations, condition and conservation, and rights and restrictions. Some even allow users to create deeds of gift and location labels, track usage statistics, and manage requests for materials and reference help. Others focus more on archival description than on collection management. Many do both. Archives should determine what features are most essential to them, while noting that new versions of software often add features that they may desire.

Some software can enable institutions to run reports to, for example, track unprocessed collections or determine what is stored in a particular location. How easy is it to create and print out such reports? Through archival management software, organizations may also be able to track statistics such as the size of various collections, how many linear feet have been processed or deaccessioned over a year, and the most frequently requested collections.¹³ Such statistics can help archives determine how to set processing priorities and can be valuable in reporting to organizations such as ARL. Indeed, some software even allows institutions to mark accessions that are high priority for processing, helping them manage hidden collections.

Some archives are shying away from software that is still in development such as Archivists' Toolkit and Archon because "there are still bug reports." Users did report that there were some bugs or missing features for both tools, as well as for commercial systems. However, they also said that their error reports were taken seriously and that the development teams are responsive to user questions and suggestions. In the contemporary computing environment, software is continually evolving; witness the "permanent beta" status of Web 2.0 tools such as Google Documents. It is possible for software to be too mature, built using out-of-date technologies or approaches. On the other hand, some software never makes it out of beta or may not go in the direction anticipated, so institutions may lose time and resources if they adopt untested software.