The notion of information is central to our mission. Depending on its purpose, an information system may have to be able to retrieve information from databases or web services, receive it by mail or collect it in structured forms or in less structured conversations, extract it from natural language texts, store it in databases, combine it with other kinds of information, and present it to users as (any combination of) text, written or spoken conversations and graphics. For this to work, information has to be represented. We use RDF(S) and to a certain extent OWL for this purpose.

Interaction is another key notion in our philosophy. The interaction between user and system, the interaction between two or more users (mediated by a system), the interaction between one information system and another, and the interaction between different loosely coupled components of a single system. Interaction may be single-modal or multi-modal. Our platform handles all these forms of interaction in a uniform way — by means of interaction managers specified in SCXML.

Our expertise in Natural Language Processing (NLP) is perhaps our greatest asset. If a particular application requires part-of-speech tagging, word sense disambiguation or parsing, then we know how to do it. In the context of our platform, such components are set up as web services, and communication with them is established using messages (or "events" as we prefer to call them) encoded in XML and/or JSON.

When good standards for what we do exist, we are committed to them. All in the interest of portability and interoperability — things we know are important to our clients as well. We are concentrating on RDF(S), VoiceXML and SCXML in particular, and are also actively involved in a W3C working group aiming at creating and looking after the SCXML standard.

We always choose mathematically well-founded approaches when they are available. That is, we choose to work with RDF(S) and OWL, not only because they are standards, but also because they are based on formal logic. And we choose to work with SCXML because it is based on the theory of state machines.

Before implementing our target applications, we do a lot of modelling of both the knowledge needed (a.k.a. knowledge engineering) and interaction flow required (a.k.a interaction design). Graphical notations of various kinds makes it easy to involve client and/or end users in the process. Once we are done modelling, the step to a working system is usually short and painless. The charts and diagrams produced during the modelling process make up an important part of the documentation of the system.

Our platform is highly distributed and very flexible. In the interest of achieving a load balance optimal for a particular application, core components can be run on either servers or web clients (e.g. browsers). Note that the standards and methods ("best practices") that we advocate are not tied to a particular platform which means that using them with our platform is optional.

Just as we insist on using standards, we insist on openness and transparency. We are therefore willing to lecture about our methods, in courses that can be targeted towards different kinds of audiences. We are also willing and able to act as consultants to companies and organisations that are about to invest in new information systems, or want to enhance already existing ones. We furthermore believe in open source software, and we are indeed open to the extent that even our business plan is left open for everyone to see (only in Swedish at this time).

Here at Spyderbrain we have our minds set on programming, not just on programming single computers, but on programming the web. As one giant computer, if you will. As we see it, notions such as "Web Services", "the Semantic Web", "AJAX", and "Mashups" are all subsumed by this vision.