Conference organizer Eric Norlin began by stating the premise behind Glue is “what will happen if (or when) everything moves to the web.” Panel moderator Stuart Allsop later in the day tried to put an acronym – which signals a concept with staying power – around Glue and came up with ‘Globally Logical User Environment.’ Put in broad terms Glue could be described as everything that is needed to allow all computing to move to web and everything needed to allow for seamless portability and scalability within and among all the different platforms and applications that exist in this environment. Perhaps the difficulty in defining Glue is that at this point it is not simply sticking components in the right place based on a set of clearly defined instructions. Rather there are a bunch of pieces (applications, platforms, protocols, etc.) available but no definition for what they should be assembled into or what the best method or materials are to attach them together. With this kind of uncertainty, the day’s discussions may best be framed around common questions.
The first question comes down to what users are looking for glue to accomplish. There is agreement on some basic principles. Users generally desire a simpler sign-on process without the need to create redundant credentials for different sites and services that at the same time securely verifies identity. Ease of pushing out data to selected contacts by group (i.e. family, co-workers, close friends) regardless of platform or application, as well as better filtering to ensure each user is fed relevant data are also generally recognized as necessary. Beyond just identity the idea of an online reputation, a somewhat qualitative measure of credibility, was also discussed as desirable to travel with a user online. What makes the implementation of reputation more challenging is that it is highly dependent on context. For example, a person’s reputation as an ebay seller has little relevance to expectations of his credibility in a java forum. Moreover, this reputation is dependent on the reader of the information who is more likely to believe it if it includes references from the reader’s network.
The discussion of what users want in future products is also limited by the fact that few can clearly articulate a vision beyond tweaks to the products they already understand. A panel on web apps agreed that a full migration to SaaS products from computer-housed software would be a disappointment if all it meant was that a small business owner had several tabs open in a browser instead of several different windows for programs. They agreed some sort of new type product that brought together the functionality of different applications would offer a greater value.
The next question is how to technically accomplish these “goals” (e.g. data portability, online identity) – with goals (of course) being something of a moving target. Core to this challenge is simplifying and standardizing communication enough so that “it’s as easy for machines to interact as humans,” according to Aaron Fulkerson – a challenge made more difficult because machines are “unforgiving” and unable to abstract and interpret in the same way as people. While the communication methods must be simple for maximum applicability, they must also be rich enough to contain all the information that needs to be pushed and pulled by different programs. When considering execution, the communication must not only occur between applications but also among layers, as Software as a Service products must interact with Platform as a Service and Infrastructure as a Service Technologies which may still have to communicate with physically hosted products (e.g. legacy mainframes).
The final question is how to gain user acceptance – and in this case users mean the general public, not the tech enthusiasts in the Glue crowd who attempted to sign up for beta versions of the non-existent web apps mentioned by Gnip’s Eric Marcoullier during his presentation. Users have fears, both real and imagined, about data and applications moving to the “cloud.” Enterprises moving essential data outside their walls face questions like, “What happens if a provider goes out of business or faces subpoena?” There is also just the fear that goes along with one’s data being out of one’s physical possession. One suggestion to mitigate this problem was transparency about any issues (security, down-time, etc.) and providing customers with as accurate estimates as possible on when problems would be fixed. The thought behind this idea is it could give individuals the more satisfying “the IT department is working on it” feeling rather than the uncertainty that currently exists when a product like Gmail goes down. Salesforce.com’s Peter Coffee noted it can be challenging to assess how much information to give users about exactly what stands behind an application. On one hand, does a customer really care whether an application is housed in the cloud as long as it functions as desired, but on the other hand more knowledge can lead to comfort and sometimes the ability to self-diagnose issues that may arise.
The Glue Conference continues all day today during which time I expect all the uncertainty and questions from Day 1 will be definitively resolved
