Enterprise Search Profiles: A Service of ArnoldIT.com
Over the years, we have written profiles about vendors engaged in search and search-related software. As i approach 70 years of age, I am taking drafts of these profiles and posting them on this Xenky page. Some of these profiles have been refined for various publishers. Others were prepared for interested parties. My intent is to provide a person looking for information about search and content processing with basic descriptions of companies, products, and technologies. Looking for information about defunct firms and some firms that have been acquired is frustrating due to the changes in the online access systems. Google is better at presenting information about a popular topic than as a tool for supporting certain types of specialized information scanning.
As my team and I work through my “foul papers” of the company profiles in my files, we have learned that descriptions of systems, benefits, and features have been consistent for decades. Some of the systems dating from the 1980s (ISYS) and Verity (now part of HP Autonomy), for example, are essentially in step with modern vendors’ sales pitches and technical descriptions. Systems which went out of business sported technology that was positioned in a manner comparable to today’s most successful systems.
If you want to add information to a profile, you may navigate to Beyond Search and the specific story announcing the availability of a specific profile. Once at that story, you m ay use the Comments function to additional information, links, argue with me, or add your own commentary to a particular profile.
I am not “unfreezing” these drafts. I am not updating them. These are provided on an “as is” basis. Please, note that these are provided as reference only and commercial use is not permitted without prior written permission. Like our reference material on the frozen site www.arnoldit.com, I am building an archive because I am cutting back on the work that I have been doing for 40 years. I assume my estate will keep the site online once I log off for the last time.
Let me offer several observations:
Descriptions of defunct vendors’ systems are surprisingly timely. The reason a vendor fails seems to correlate with the actual software failing, not the marketing which sold a system. Great marketing does not mean a flawed system will survive.
The “lingo”, jargon, and buzzwordage is quite fresh across the profiles which span the 1980s to the present day. Developers have been given the job of making “science fiction” work in an organization. Search vendors can paint a word picture and often the systems are described as meeting needs which are essentially unchanged over the decades
Search has not made significant strides comparable to the progress in mobile device hardware. Search is a very difficult problem to solve and turn into a sustainable business.
Search developers like Sisyphus may face a thankless task. The marketers will just recycle glittering generalities. System users will face findability challenges. That’s one hypothesis these profiles touch upon.
Profiles will be added on a periodic basis but not a fixed schedule.
Stephen E Arnold, October 8, 2013
Autonomy [autonomy-01-14] Autonomy is an important vendor. The company was a pioneer in its use of Bayesian, Laplacian, and Volterra methods as key components in its Integrated Data Operating Layer (IDOL) and Digital Reasoning Engine (DRE). The company generated almost $1 billion in revenues for fiscal 2011 when Hewlett Packard acquired the company, its patents, and customers. No other enterprise-centric search vendor has come close to matching Autonomy’s revenue generation. Autonomy’s technology foreshadowed the approach of other vendors like Google and Recommind. Endeca, when Oracle acquired that company in 2011, tallied about one-tenth the revenue of Autonomy. Prior to the sale to HP, Autonomy stands as the pre-eminent marketer of search technology to the enterprise. This report covers the technology, the marketing, the strengths and weaknesses of the company. [2014-02-05]
Convera. [convera-2013-10-08.pdf] The company is important because it was one of the first firms to articulate a comprehensive vision for information access. Convera promised video search, text search, text mining, automatic indexing, clustering, work flow, analytics and more. The company also pioneered the jumping from enterprise search to Web search to vertical search in an effort to generate sustainable revenues. The company’s interactions with firms like Intel and the NBA made clear the risks inherent in trying to make search deliver on marketing assertions. Vestiges of Convera exist today. Some integrators still support licensees of RetrievalWare, Convera’s amalgam of proprietary software, acquired technology, and scripts created to add features to a particular installation. [2013/10/09]
Delphes. [delphes-10-18-final.pdf] Delphes and its Dio range of search system is an example of academic theory colliding with business realities. Delphes was based on linguistic methods developed at a number of high-profile universities. The company Delphes sought to commercialize the chief technology officer’s concept. The resulting system was described in terms of “linguistic soul.” The system included most of the early-2000 information retrieval methods, including natural language processing, work flow, federated search, and automatic indexing of content in different languages. Delphes listed some notable clients. The company went offline in 2007 and its senior management dispersed to academic and other pursuits. [2013/10/22]
Dieselpoint. [Dieselpoint-10-31-final.pdf] Dieselpoint has been an alternative to the business-chasing, XML-centric data management systems for many years. The company offers a wide range of features in a pure Java search-and-retrieval system. Dieselpoint’s system can process structured or unstructured information. The system offers the Java specialist an extensive playground. Multiple languages, facets, and speedy query processing for eCommerce, search-based applications, or basic Web search make the system interesting. If you have not heard of Dieselpoint, that is because the company maintains a low profile. The company is in business even though portions of its Web site have not been updated for months. [2013/11/04]
Entopia. [entopia-10-14-Final.pdf] Entopia took advanced methods like semantics, mixed in some MBA buzzwords, and set out to make search the infrastructure of the licensee’s organization. Tacit Software (acquired by Oracle) was active in the email monitoring and analysis business. Other search vendors combined analytics and automatic indexing. Entopia, however, was not able to get the commitment a company would make to SAP R/3 or comparable software system. Information is a big problem. When a company like Entopia sells a large company on a big solution, the company has to see value in the system. Entopia was unable to deliver on its vision. Search, no matter how it is packaged, was difficult to implement and demonstrate measurable value to the licensee. The company went offline in 2006. [2013/10/16]
Fulcrum. (PCDocs, Hummingbird, OpenText). [Fulcrum-12-23-13.pdf] Fulcrum Technologies opened for business in 1983. Its technology is incorporate into OpenText’s solutions in 2013 and probably for the foreseeable future. The company’s Ful/Text and Search Server are not widely known today. The code is now three decades young. What makes Fulcrum interesting is its technology, its marketing of search and a knowledge system, and its numerous owners. The Fulcrum case provides a Google Map style view of how a search company’s journey leads to financial peaks and valleys, through marketing deserts and oases, and from owner to owner with little change in the actual search technology itself. Fulcrum brushed open source, linguistic methods, and almost every possible way to apply search to an organizational problem. Fulcrum is instructive for would-be search entrepreneurs relaxing as the New Year approaches. [2013/12/23]
iPhrase. [iphrase-jan-2014.pdf] Now a unit of IBM, iPhrase continues to provide some functions to OmniFind, a search and content management system. iPhrase approach enterprise content with a “return on investment” argument. Some of the company’s rhetoric borrowed from Autonomy’s earlier “black hole” analysis. From a technology point of view, iPhrase demonstrated that Extensible Markup Language could be used to re-render source content and impose significant computational burdens and generate a voracious appetite for storage. The company landed some big accounts and elected not to go public. The firm sold itself to IBM. The lessons of iPhrase are ones I find instructive, particularly with the valiant effort to “prove” that search delivers ROI, a theme that other vendors embrace but, like iPhrase, struggle to back up with credible financial data. [2014/01/07]
SchemaLogic. [SchemaLogic-Final-11-27.pdf] SchemaLogic provided a content management system for taxonomies and controlled vocabularies. The company found sales success in organizations in which an appreciation and understanding of tight indexing existed. Taxonomy became a marketing buzzword, but the number of organizations willing to spend six figures or more to index content was modest in relation to the number of companies with no indexing or “good enough” indexing control. The company, like many content processing-centric vendors, struggled to deliver strong top line growth and a healthy net profit. The firm was sold in 2011 to a competitor, Smartlogic. [2013/12/2]
Siderean. [Siderean-B.pdf] Siderean Software’s Seamark Navigator embraced semantics. The system provided a more modern approach to Endeca’s aging system. The company’s system anticipated the later success of MarkLogic as an way to wring value from content in SGML or XML formats. The interfaces that Siderean’s team developed provide an interesting glimpse into the ways in which related content can be explored. Metadata may offer too many options, thus making a system’s outputs too confusing for some applications. Siderean made sales to some high profile clients, but the revenue generated from license fees was not able to support the company’s operations. Siderean went offline in 2008. [2013/11/11]
TeraText (formerly a product of SAIC, now Leidos pty). [TeraText-Feb-21.pdf] TeraText’s technical foundation dates from the 1980s. Few of the enterprise search vendors I encounter know much about the information retrieval work of Dr. Ron Sacks-Davis or the forward-looking system he developed decades ago. TeraText is an important product today because it continues to provide infrastructure, content processing, and information retrieval applications in a number of low-profile organizations. The system allows a user to search. More importantly, TeraText may be one of the first, if not the first content processing system, to make it possible to make information retrieval the central function of enterprise applications. Search vendors asserting that certain functionality is “new” or a “revolutionary approach” may signal their lack of knowledge about an important system. [2014-03-03]
Verity. [Verity-Dec-2013.pdf] Verity was one of the leading providers of enterprise and OEM (original equipment manufacturing) search from the late 1980s until it was purchased by Autonomy in 2005 for $500 million. Verity’s system evolved into a complex, expensive, and often sluggish rascal. The company, like many other “one size fits all” search vendors rode a financial teeter totter as it tried to find a way to make search into a higher-value enterprise solution. The company rapidly added layers of functionality to its core search-and-retrieval system. Like Fulcrum and other search vendors, the quest for revenues reached into the fuzzy domain of knowledge management. The company purchased the Ultraseek search system in a tactical move to have a lower-cost option for prospects and respond to potential customers’ price resistance. Verity was one of the first search vendors to emphasize consulting and engineering services as a major source of revenue. Since Verity became part of Autonomy, the company’s brand identity has diffused. [2013/12/10]