The past several years have certainly brought consolidation among the integrated library systems (ILS) vendors. Sirsi merged with Dynix, Reed Elsevier sold Endeavor to Ex Libris (Francisco Partners), and OCLC PICA acquired Fretwell-Downing. Now, the largest vendor, SirsiDynix (which was recently acquired by San Francisco-based Vista Equity Partners), has announced a major consolidation of its core products in a new technology platform that the company said builds upon the best features of each. The single platform, code-named "Rome" and based largely on the Unicorn system, will allow the company to focus its development efforts and implement changes in a more timely way. While SirsiDynix clearly hopes its new offering will prove compelling to its customers and disruptive to competitors, it is clear that the company's library customers now face a tough and disruptive decision: migrate to Rome, stay with existing products that will no longer be developed, or opt for an alternative product.The first version of Rome won't actually be released until "later this year" with a subsequent version planned for late 2008. The platform is being built on the architecture of the company's Unicorn GL 3.2 system and will incorporate the best features that have been part of the company's Horizon 8.0/Corinthian development.
Talin Bingham, SirsiDynix chief technology officer, said: "Neither Unicorn nor Horizon systems are being eliminated. Instead, SirsiDynix is blending features, capabilities, user interfaces, modules, and more as the foundation of a single, new platform featuring best-of-breed solutions from both product families. The result will be a technology platform that can grow and evolve along with libraries and consortia."
The company said that Rome will offer integrated technology building blocks for libraries, including cost-saving software-as-a-service (SaaS) hosted options, portal and search solutions, new and easy-to-manage Web-based library staff software, electronic content management, presentation solutions, a suite of library management and productivity solutions, and more. Bingham said the company is working with technology from Fast Search & Transfer (FAST) on a faceted search engine with clustering capabilities and a visualization component.
Tom Gates, VP of marketing, said that the new management looked carefully during the last month and a half at "where we are as a company" and found it wanting in several areas. So, it decided to do a better job of allocating resources, setting priorities, and delivering products in a more stable and timely way. "Change was needed. We had too few resources working on too many standalone products to be successful. Our goal was to create a new holistic platform with a full range of building blocks needed by customers."
In its press release (http://sirsidynix.com/Newsevents/Releases/2007/20070313_technology_platform.pdf) and in a briefing call with media, executives of SirsiDynix stressed the benefits of its best-of-both worlds approach and spoke glowingly of all the new features that would be added to the "new" platform. One industry expert, however, who read between the lines and understands the ramifications, called it "a highly obfuscated announcement" that glossed over both the dropped development effort for Horizon 8.0 and the real impact on the company's customers.
Initially, at least, Rome will be a prettied-up version of Unicorn—a 25-year-old system that is valued for being mature and stable. In fact, users have been anticipating the release of the completely re-engineered Horizon 8.0 (promised for February 2007), and 10 libraries have been beta testing it. In addition, a number of legacy Dynix sites had been planning to migrate to Horizon. Now, development on Horizon version 8.0 has ceased; Horizon 7.4 will be the final version. However, the company stressed that it will continue to support Horizon 7.x indefinitely. It will support users of Unicorn GL 3.1 (3.2 will not be released as such), and it promises to work with customers to plan "optimal upgrade paths to Rome."
Library automation expert Marshall Breeding called the decision "very disruptive to say the least." He said that migrating to a new library system really involves a years' worth of hard work by many people, and the cost for staff time is probably equivalent to the cost of the new system. He doesn't think that the new private equity owners of the company really understood how much library customers would be impacted by this decision. (The new owners announced the consolidation decision just 55 days after taking ownership. And the company's CEO, CFO, COO, and chief marketing officer all resigned in mid-February, following the acquisition.) "This industry is fundamentally different [than other businesses]. The organizations are mostly non-profit and supported by tax dollars[,] and thus the libraries are less flexible. We're all running on a shoe string," Breeding explained.
He also noted that both Sirsi and Dynix were profitable endeavors before their merger and were even more profitable afterward—and the combined company has the largest development staff in the industry. He feels there are other strategies that could have been chosen. "What about providing front ends to both ILSs? It could have been done as a phased approach to get to a common system. It would have taken longer and cost more but would have been less disruptive to their library customers," he said.
(For more background on the situation from Breeding, see his article in the April Smart Libraries Newsletter at www.techsource.ala.org/sln/april-2007.html.)
One Horizon library systems manager at an academic library offered this blog comment on the news: "Rome has come at a very high price to many Dynix and Horizon sites (especially those further down the 8.0 migration path that than we were at [X]) and feels to me to be far from the ‘best of both worlds' that SirsiDynix is claiming. Amidst all the anger and frustration, quite a few libraries are seriously talking about a move to the open source Evergreen ILS."
Not all responses to the SirsiDynix news in the library blogosphere were negative. John Blyberg, head of technology and digital initiatives at the Darien Public Library in Connecticut, commented: "I applaud the SirsiDynix decision to punt on Horizon. You've got to admit, that took a serious gut-check. I can't, in good conscience, join the cacophony of hems and haws over this news after, myself, espousing the need for vendors to tear it down and build it back up. This is precisely the type of thing we ought to be expecting from our vendors every 10 years or so. I wish them the best of luck with Rome."
Systems librarian and blogger Ben Ostrowsky said his organization had counted on staying on Horizon 7.x for a while but planned to migrate to Horizon 8.x eventually. He was frustrated that Horizon 8.0 was joining the vaporware list, which, he noted, had been announced for release back in 2005. And, he added, "It's pretty certain, though, that the work I'm doing to improve our HIP catalog is going to be left by the side of the road. It probably won't translate. All roads may lead to Rome, but once you get there, you'd better be speaking Latin."
Andrew Pace, head of information technology for North Carolina State University Libraries in Raleigh, wrote in a blog post that he understands the SirsiDynix platform decision from a business standpoint, even if he might not agree with it 100 percent. The problem he sees "is that all the descriptions of Rome sound a lot like the descriptions of Horizon 8.0. Heck, they sound a little like Taos, for those who want to go back even farther. Rumors abound about the instability of Horizon 8. I cannot speak to those rumors. So, I think the decision about Rome is essentially a ‘time to market' decision. It's a choice between a sexy platform that is unstable and a stable platform that is unsexy." Given the migration requirements caused by this decision, he also feels there might be a "golden opportunity" for the recently released open source Evergreen system, as well as competitors of SirsiDynix.
Breeding agreed that Evergreen and other open source options could be attractive to some libraries. But he cautioned that Evergreen was a relatively immature system—one that is pretty good as a start but one that would require a lot of time and effort to build out.