Marshall Breeding is well known in the field of library automation,  maintaining the website 
http://www.librarytechnology.org/.  His talk is based on a recent survey of librarians attitudes to current 
systems/vendors and the emerging open source market.
  His scan of the current state of the market pointed out that no successful  recent new systems (e.g. latest version of Horizon was abandoned), the current  crop date mainly from
the mid-90s. Recent mergers have reduced choice, and  even newer systems such as Evergreen are doing the same things as existing  systems.
  
 This has been accompanied with increasing dissatisfaction as they fail to  keep pace with customers expectations of innovation. Nowadays there are very few  voluntary migrations to lateral systems. There is a need to focus on  e-resources, the user experience etc, and  this could encourage less integrated  systems with a core system to which is bolted on a discovery layer, link  resolver, federated search tool.
  
 Companies are beginning to get involved in library automation who are not  traditional system vendors e.g. OCLC, Bowker.
  
 Marshall spoke at length about the Open Source alternatives which are  garnering a lot of attention. Ultimately the Total Costs of Ownership are  similar to that of a proprietary system. OS will penetrate the mainstream when  TCOs are well documented enough to stand up to objective procurement. Currently  OS systems are very similar in functionality to commercial  systems but their  impact could disrupt the status quo by
  
 injecting competition into the market
increase pressure to  innovate
increase pressure to decrease costs
make systems more open
 
We need to work towards a new ILS vision, e.g. current systems are  based on workflows cast > 25 years ago, e-resources now represent >50% of  our resources, many systems have large gaps e.g. ILL, collection development, 
binding, remote storage.
  
 Alongside OS software we need to consider more open access to data open  APIs, ideally industry standard which would allow access to all components of  functionality
 Can these be open and commercial?  Marshall advocates ILSs becoming more  lightweight as modules become interoperable, with a single point of management  for each function.
 Our current legacy systems have created artificial boundaries we need to  redefine these i.e. 
 PAC / portal 
 Circulation / ILL / Remote Storage 
 Collection Development / Acquisitions / Budget management 
 The first stage of this has begun with the separation of  the front-end (PAC)  by using next-genration discovery tools/interface. Technology cycle is much  faster at the front-end and this is only a small part of the library ILS
 We should see a move towards service-oriented business architecture where web  services allow the flexibility to weave a fabric of changing applications. This  could lead to greater enterprise operability and open the door to massively  consolidated implementations, of scaled up consortia. 
 The 'Global Enterprise' of Google, OCLC, Worldcat etc has to be tackled - what  is our relationship with these? How can we leverage our content in enterprise  discovery systems to drive users toward library resources e.g. by exposing the  metadata.
 We have to consider the place of MARC not only in an XML- based world, but  also in a post-metadata world where users are searching the  digital objects  themselves.
 'Web destinations' such as  Amazon are now competing with libraries,  increasing the pressure on us.
 We have to break out of the marketing/consumer model when dealing with  vendors and move towards dialogue and increased partnership. 
 Evolution or Revolution? Web 2.0 has invigorated libraries and it may be has  provided the catalyst for the latter.