[Editor’s Note: New contributor Shion Deysarkar, the CEO of 80legs, sent in his thoughts on key trends to watch next year. Prior to joining 80legs, he founded and ran EXC, a predictive software consulting firm. 80legs is a service that resells idle time on users’ computers (similar to how the SETI@home project works. Find more at 80legs.com. –mrh]

Real-time listening garners huge attention but struggles to deliver actionable data

“Real-time” has already emerged as a hot topic, but most of the buzz has centered around accessing the real-time stream.  Access to data is always an important step, even though it may be taken for granted later on, but an equally (if not more) important step is making sense of the data.

A fair amount of the conversation around real-time has centered on making sense of the data, but I’ve heard very little about how to actually do it.  I’m not sure why this is, but I suspect the root cause is that making sense of this data is a very, very hard computational problem.  It’s going to require serious, sophisticated algorithms to convert 1 million variations of “RT @friend21 coldplay concert wuz awesome” to “Opinion on Coldplay Spring Tour: 78% positive, 15% negative, 7% neutral or ambivalent; major positive contributors: music, Chris Martin; major negative contributors: price, ticket availability”.

Until we get to that point, there will be many disgruntled folks questioning what the value of the real-time stream is.  Access is awesome, but action is where the value lies.

Continued demise of database-driven storage. Standardization of non-database storage.

As the sheer amount of data available to us grows at an exponential rate, database-driven storage will fail to keep up, and alternative systems like Hadoop and other file-based structures will dominate large data systems.  An important thing to watch will be how standardized the management of large data systems becomes.  Right now Hadoop has all the buzz, but there are several instances of companies building their own proprietary systems.  In as much as I’ve heard, the proprietary systems perform better because they are designed specifically with the company’s tasks at hand.  Standardized systems have obvious long-term benefits though.  Which will hold up?

Emergence of the Semantic Web

Related to making meaning out of data, I strongly hope we will finally see the emergence of the Semantic Web in 2010.  Adoption of semantic technologies, such as RDF, n-tuple stores, NLP, and so on will go a long way to turning the Web into what we really want and need it to be – a gigantic query-able database.  There are many pockets around the world, from big corporations to local meet-ups, working hard on making the Semantic Web a reality.  It will be a huge boon if we see the fruition of this work next year.

Large Internet companies solidify their market position as competitive scale becomes unattainable for smaller players

As the amount of data on the Web grows, it will be increasingly difficult for smaller companies to achieve and/or offer the scale of infrastructure necessary to deal with that data.  Although the cost of storage, CPU time and bandwidth continues to decrease (well maybe not the last one with tiered pricing), the amount of data to churn through is increasing at a faster rate.  Only large companies (think Google, Facebook, etc.) have the capital required to work with data at this scale.  We will need innovative technologies to figure out how to overcome the current financial realities of dealing with big data.

Startups learn how to charge for things

Startups will learn that giving away free stuff does not make a business and we’ll see fewer startups emerge without a clear path to revenue.  It’s become clear that advertising revenue is only significant at scale.  Either the products and services offered by startups will start being priced appropriately, or alternative revenue models will emerge.  We may see more freemium models or “tangential” models like Mint’s.

This trend will be reinforced by pressure on venture capitalists to perform better as an asset class (one of the worst in the last 10 years if you take away 2000, I believe).  This will drive a stronger focus on paths to revenue, and less on spray-and-pray approaches.

Startup culture spreads to the rest of the U.S.

Last year we saw the emergence of incubators in several cities outside of California.  This will continue to happen, but in addition to incubators, we’ll see city governments get in on the action as they try to encourage startup growth through co-working facilities, event sponsorships and more.  We’re starting to see this in Houston, where the local government is actively engaging the local startup scene for advice.