写于2001年7月

图书馆计算机集成管理系统正在发生根本性的变化，这些变化正在重新定义图书馆的”集成管理系统”的”集成”性，可能还没有反映到具体的产品中，但已经开始影响图书馆的具体实践，在诸如香港中央图书馆、加州数字图书馆等图书馆中得到了初步体现。

最 显著的特点是传统的基于图书馆业务管理自动化的软件系统与基于图书馆处理对象的数字化而兴起的数字图书馆系统两者之间的竞争、影响、融合。我曾经非常反对 对于图书馆的计算机应用”用一套系统解决所有问题”，认为这是不可能的，特别是对于数字图书馆，大家对其认识尚不统一，做法各异，怎么可能试图开发一套解 决方案呢？然而许多图书馆解决方案的供应商一直在进行不懈的努力。我认为的”不可能”，是把传统的图书馆业务流程自动化看得很绝对、很神圣，如果打破这一 点，从头设计一个适应数字化网络化时代的图书馆，其计算机解决方案是可以很好地集成起来的。

这样一个解决方案 难点还是在”系统需求”的把握上。数字图书馆众说纷纭，对不同的图书馆有不同的模式，如何统一？我们认为，统一的根本点还是在”服务”上，以最终的服务梳 理和重构整个业务流程。资源整合是现阶段的热门话题，整合资源是为了”知识导航”，这里面有大量的工作要做，目前CALIS计划二期正在积极筹划。对于未来，四川大学信息管理系的张晓林博士认为，将是个性化、智能化和大量”嵌入式”服务的时代，许多信息系统和个人对于图书馆中信息资源的利用是在不知不觉中完成的，当然，许多服务的增值效应也不知不觉地将体现在图书馆的银行账户中。

这样看图书馆就真正融入到发展前景看好的”信息服务产业”或”内容产业”中去了。这种情况下图书馆的形态发生任何变化都是有可能的，图书馆的计算机管理和服务可能外包，”超级”图书馆可能成为ISP/ICP/IDC，ERP/CRM的技术和理念将会得到同样的应用，电子商务的应用将是不言而喻的，现有图书馆条块分割的行业体制再一次被证明是一种很大的阻碍。

图书馆学理论宗师很早就定义了图书馆”为书找人、为人找书”的服务宗旨，传统的图书馆集成管理系统只解决了一半问题，另一半问题现在是解决的时候了。

前两年对数字时代图书馆集成系统的需求曾经想做一番调研，在馆所立了软课题而没有做。今天在Lorcan Dempsey’s blog上看到一个很好的总结，可以以此为蓝本加以完善。

http://orweblog.oclc.org/archives/000585.html

February 22, 2005

The integrated library system that isn’t

Libraries – systems and technologies, Libraries – distributed environments, Libraries – organization and services

One can read the phrase Integrated Library System (ILS) in two ways: as a system for the integrated library, or as an integrated system for the library. Although the latter is what was probably meant by the term, neither is an accurate description of what the ILS has become. In fact, it is a misleading term whose continued use is bemusing. It is clear that the ILS manages a progressively smaller part of the library activity. There has been a real shift in emphasis towards e-resource management (see the metasearch/resolver/ERM/knowledgebase suite of tools), and in some cases towards digital asset management. Libraries now manage a patchwork of systems which do not always play well together.

Think about the systems that support current library processes, and some possible directions. Items marked with ILS are typically a part of the ILS offering; those marked NILS are usually not.

Acquisitions (ILS). Libraries are part of larger organizations which ‘acquire’ a variety of materials and services, and have built enterprise systems to support this. It is likely that in many settings libraries will make more use of the generic institutional systems in the future.

Catalog (ILS). The local library catalog — let’s not use the user-unfriendly and jargonish ‘OPAC’ — is not a central part of most users’ information behavior. Users need to be able to discover items of potential interest to them and locate them in the library. The catalog does not do a great job with the former: it does not make its data work very hard. As it covers a part only of a user’s information needs, and a part only of the library collection, it may exert a progressively weaker gravitational pull on the user. More about this below.

Cataloging (ILS). Libraries have various cataloging workflows. They may acquire records from various resources, roll their own, participate in a shared cataloging environment. However, libraries are also creating metadata for other resources which may be poorly supported in their cataloging environment. They may have two workflows (their local system and a cataloging system such as Connexion from OCLC).

Circulation (ILS). This appears to be core function of the (current) ILS and it is where such systems started. However, even here there is an interesting trajectory in some environments towards group wide circuation systems. See Ohiolink for example, where the line between inter library loan and circulation becomes blurred.

Metasearch/portal (NILS). Much could be said about this intractably difficult challenge! (see here for a short and here for a long discussion of metasearch/portal activity). Some libraries are looking at wrapping a metasearch product around their various database offerings, including, in some cases, the catalog. Some libraries are putting a lot of effort into metasearch activity: it is useful, but metasearch will always be a partly broken service given the diversity of the target resource.

Resolver (NILS) The resolver is emerging as a critical part of library systems infrastructure, with the OpenURL binding resources together in various ways. The resolver moves the user from a discovered item to an instance of the item. We are likely to see OpenURLs used to tie together more systems in the future. See for example the current Google Scholar discussions, where a user might discover an item through Google and then be passed through to a resolver to locate local instances. This has been discussed in relation to articles; it could also happen with books, where the resolver talks to the catalog.

ERM/knowledgebase (NILS) A resolver or metasearch engine requires ‘intelligence’ about available collections. What is available? How do I connect to it? Under what terms and to whom is it available? Again, new services are emerging to help with this area, which may need to talk to acquistions systems and catalog.

Portable bibliography (NILS). Citation managers (Endnote, RefWorks, …) and reading lists are becoming more important. This light-weight bibliographic apparatus, a metadata bus if you like, provides interesting integration opportunities.

Digital asset management (NILS). Libraries are managing digitized local collections – maybe images from their special collections, historic newspapers, and so on. At the same time, many are looking at the systems infrastructure required to support institutional repository type services, where they provide management and disclosure services for research or learning materials produced within their institution.

Important other things Libraries may manage e-reserve systems, ILL systems, virtual reference systems, and so on. Various approaches to identity management may be in place.

So, there is a fragmented systems base, and service demands on some traditional service components are morphing as circumstances change. Here are some comments on this environment.

Thinking about collections

Reductively one can think about four collecting areas which are managed in different ‘boxes’:

1. The Bought collection. Print books and journals, CDs, DVDs, and so on. This has been the core of the library collection and it is around this that the ILS was built. These materials are catalogued and are ‘circulated’ by the library.

1. The licensed collection. A&I services, e-journals, and so on. These are now a major focus of investment and attention, and new systems components (metasearch/resolver/erm/knowledgebase) are being put in place to manage this. This area presents new metadata challenges in the form of data describing resources, the services through which they are made available, and the terms under which they are available.

1. The local digitized collection

Libraries are digitizing their rare or unique materials, releasing their research and learning potential in new contexts. This activity is in the cottage industry stage. Metadata creation may be expensive. Digital asset management solutions are not quite routine yet.

1. The managed institutional research and learning output. This is the institutional repository and learning object repository space.

The balance of investment between these collection areas is different in different libraries; but more interestingly it is probably changing within individual libraries. It is interesting to think what the relative balance between them will be in, say, ten years time, and what implications that has for systems support.

Thinking about services

One downside of this fragmented systems and collections environment is that it becomes more difficult to build services out on top of the collections. Too much effort is going into maintaining and integrating a fragmented systems infrastructure.

This becomes more of an issue as the pressure on the library to be seen to be ‘making a difference’ grows. Increasingly, the library needs to bring its services to the user within their work- or learn-flow, and be seen to be adding value to the collection of resources.

Thinking about directions

Here are some thoughts:

• The systems environment needs to become simpler. We will see more hosted solutions, better integration options in a ‘web services’ environment, and some consolidation of supply.

• For ILS vendors there seems to be an interesting shift away from their historic core towards e-resource management, and in some cases towards digital asset management.

• We will see less focus on the integration of library resources with each other as an end in itself, and more on the integration of library resources with user environments (personal, learning management system, etc).

• Following on from this, data and services need to be made available in ways which better facilitate their recombination in different user contexts. This touches on what I have called intrastructure, the applications tissue that allows us to more easily stitch together systems and services. RSS feeds, URL-based web services, bookmarklets, data import and export: these are all boundary crossing services which enable better stitching.

• I think that we are entering a period where opportunities to centralize services and data will be looked at more seriously again, as a way of reducing cost and complexity, and of releasing resources to focus on user experience.