Panaqqa Predictives

There is one nightmare that has been keeping Microsoft’s executive leadership awake for years now: that everything transfers to the web, from enterprise applications to personal data to games. No one will care what the underlying operating systems is. All that anyone will need is a computer with an operating system that includes a basic file system and a very sophisticated browser. And of course this means that the majority of new computer buyers would see no need to pay the “Windows tax”. Yes, a nightmare indeed.

The one single barrier to this that gets the most mention is the “no web = no software” problem. This is a major barrier, but not an insurmountable one. As of this writing there are multiple major development efforts underway to allow web based applications to operate offline, including work by most of the major players. Google is working hard to develop an offline Google Docs client. Our prediction here is a website or browser addon that works independent of network status rather than a standalone desktop client. Adobe’s AIR techology is another major effort underway to make web applications work in the offline environment. We suspect that Google will take most of the enterprise adopters and leave Adobe with most of the rest. Enterprise IT departments remember all too well Adobe’s many attempts over the years to achieve vendor lock-in, beginning with PostScript at the dawn of the desktop computer age.

Many people who trot out the online-offline problem as the only barrier fail to look a little more deeply to see the other problems inherent in moving to an exclusively web-based application software space. One word: standards. We are not referring to the W3C standards and the fact that compliance with them is sketchy. No, we’re referring more to database schema and portability of data between applications. Sure, we do have a few standards that ported across well from the desktop environment, such as CSV files, but apart from those, we are sorely lacking in standardized metadata. In terms of standardization, with the ability to import and export data between web based applications, we are at about the stage that desktop software was at in the mid 1980s: lots of proprietary file formats which their developers hope to impose as the standard for the software category.

XML has the potential to solve much of the standards issue. The medical industry is a prime example of how a push for industry wide standards takes much effort but produces good solid results. But this is only one industry. Users will want to be able to change software providers with as few migration headaches as possible.

Unfortunately, open standards have not gained as much traction as many would have liked, and we don’t see this situation improving anytime soon. As we increasingly move to a SaaS model (Software as a Service), we will see the same attempts at vendor lock in that we are familiar with from desktop software vendors. As recently as 2006, we have seen major players in the desktop software market threatening lawsuits over standards issues. We are referring specifically to Microsoft adding the ability to save PDF documents in Word 2007, rousing the anger of Adobe. We have also seen inferior and proprietary standards proposed to the ISO, along with attempts to force them through in underhanded ways including changing the rules, vote rigging and buying, etc. Microsoft again (OOXML). Isn’t it funny how their name keeps cropping up in any discussion of standards and interoperability?

We will have more to say on SaaS and the web as operating system in a near future post.

Stanford University has been the source of many innovations over the years, with the central principles behind the Google search engine being prominent among them. More recently Andrew Ng and Ashutosh Saxena, an assistant professor and a doctoral student respectively, have applied a machine learning algorithm to a tricky problem: creating a three dimensional model from a single photograph. Extracting 3-D from several images has been successfully done in a variety of ways (e.g. Microsoft’s Photosynth project), but the single image problem has been more difficult to resolve, partly because of the complexity of the underlying mathematics and the nature of the required learning algorithm.

Humans unconsciously use thousands of different visual properties to determine depth. Grass, for example, simply looks different from 20 meters away than it does from five. If you can see a whole mountain, it is likely quite a distance away. Sky and clouds have very distinct textures which identifies them as distant also. Earlier attempts to process an image had difficulty with arbitrary angles, relying on horizontal and vertical surfaces. The Stanford process, by contrast, breaks an image into many small pieces and attempts to determine the 3-D position, angle, and orientation of each piece.

While the Stanford algorithms really only work so far on outdoor scenes, with indoor scenes being much more limited, the likelihood is that this type of image processing will continue to improve quickly as the algorithm continues to learn. Images and their measured 3-D models are compared and correlated to match visual properties in the image to their depth values.

While it is easy to predict military interest in this new technique, what really interests us is the possibilities it offers in other areas. As the researchers work to extend the algorithm to a broader range of settings, we expect to see it applied to real-life environments for gaming and virtual reality, as well as such technologies as flight simulators.

Looking further ahead, we would like to predict one niche application that will be very important. With significant refinements to the algorithm, artificial vision will be developed for the blind that provides warnings when the person might be, for example, about to walk into some stationary object such as a table. This will require optimization of the algorithm coding plus advances in multi-core processors to offer the type of parallel processing power required to process an image stream in real time. This level of advance in processors will take several years, but we feel applications such as this are inevitable. When? Sometime between 2013 and 2016.

External links:

Make3D - The Stanford project made publicly accessible

Okay, perhaps we should complete that statement. Microsoft should quit worrying about Google so much and concentrate more on other companies that pose a much more serious threat to its business model. Today I’m going to look at one of them briefly. This company is neither a newcomer experiencing massive growth (such as Google), nor is it limited to Microsoft’s primary market areas (operating systems, desktop applications, online services, entertainment and devices). Oh, and it is many times bigger than Microsoft. The company? IBM.

IBM experienced strong revenue growth in 2007, and it is poised to lay down some aggressive challenges to Microsoft and others. At a recent trade show in Orlando, Kevin Cavanaugh (IBM’s VP of messaging and collaboration software) revealed that more than 400,000 copies of Symphony have been downloaded. Symphony, of course, is IBM’s free desktop productivity software. Said Cavanaugh: “we want to allow people to invest in innovation rather than spending money on commodity software”. If this is not a clear indication that Microsoft is in IBM’s gunsights, we don’t know what is.

So where is IBM going with this? Well, we predict that their next move will be to evolve Symphony into a credible - and free - competitor to Microsoft’s .NET development framework. One that will work on any hardware, running any operating system. And still longer term, we expect to see interoperability with other desktop and web based productivity tools, including Google Apps, through adoption of open standards. And we’ll bet that the open standards will not include OOXML, the badly flawed proprietary “standard” that Microsoft has recently been trying to force through the ISO acceptance process.

It is said in the computer industry that nothing succeeds like an installed user base. Over the next couple of years, IBM’s Symphony will continue to erode the user base for Microsoft Office, but not at the expense of web based apps, which will be interoperable. At the same time, we see IBM helping to clear the intellectual property minefield that Microsoft is attempting to lay around the Open Source Software movement. We think that IBM took down NCO just for practice.

Five years from now, Microsoft will still be around, will still be a large player. It just won’t have anywhere near the dominance it once did.

Every time an interesting new technology enters the realm of pure research, a huge array of well funded experiments is sure to follow. Brand new technology often starts as an observation of something unusual that only occurs in very specific circumstances, and determining the general principle behind it takes years of research and dozens or hundreds of experiments. The technology is extended, refined, and if researchers are very lucky, they are successful in discovering and developing applications.

We are at that exciting stage of developing applications for the new technologies that include nanotubes and nanowires. Many applications have been successfully demonstrated, but only a few have the promise of making the transition from laboratory to manufactured product. Some promising applications will not make it to market because a different technological approach will be there first; nanotube based NRAM comes to mind here. Still others will revolutionize a very specialized and narrow field, only to unexpectedly find broader application where it was not anticipated.

One technological advance that we predict will prove revolutionary is currently being developed with the improvement of microscopes and microscopic imaging in mind. Scientists at the University of California, Berkeley, and the Lawrence Berkeley National Laboratory have developed a nanowire laser smaller than a red bloodcell which holds great promise for the improved observation of living cells. But to our knowledge, this development is currently being applied exclusively to imaging in biology and chemistry. We predict this will change.

We believe that this major advance in the fledgling field of nanophotonics will revolutionize large segments of the computing equipment industry. This is, of course, many years in the future. Computing devices making use of this technology might be marketed in perhaps 12 years. It will take a couple of years just for the scope of applications under research to widen. But the potential for nanoscale lasers here is incredible. Optical computing, fiber optics, and looking more than 12 years out, nanoscale holographic storage.

Online social networks have been with us for almost as long as the Internet itself. From the earliest attempts using ARPANET and LISTSERV, to the first websites such as Classmates.com (1995) and SixDegrees.com (1997, no longer online), these services exploded with the launches of MySpace and Facebook in 2005. Estimates place the current number of social networking sites at close to 200, not including niche sites, with an estimated combined user base of over 340,000,000 users.

The popularity of social networking sites has raised some obvious privacy concerns, one of which is the risk of entrusting too much personal information to large corporations. These sites often contain a large amount of data which is difficult or impossible to obtain in traditional ways. Even though the data is public, using it in certain ways or placing it in a different context could be considered a breach of privacy and might be damaging to the individual.

Some users of social networks have struggled with a dilemma: what to do when certain people, such as a boss or a parent, ask to be listed as a friend on their profile. Adding someone as a friend gives that person access to the user’s profile, photos, and daily musings. Facebook does allow a user to grant new friends access to a “limited profile”, but the new friend can easily recognize that their access has been restricted. Both rejecting a friend request and granting restricted access can create issues for all involved.

So, one might ask, how will the social network friend issue be resolved? The answer is quite simple, and will likely start to be put in place by the developers of these websites within months. Certainly, before two years are up, failure to have these changes in place will condemn a social network to the dustbin of history. The changes? Multiple profiles per user. Within a short period of time, social network sites will allow a person to maintain one profile for their close friends, one for all their other friends, one for family, one for bosses and coworkers and so forth.

Once this change has happened, there will no longer be worries about rejecting the boss’ friend request because you don’t want her to see the pictures of your Hallowe’en party. The boss will see your “work friendly” profile. You need no longer worry about a friend of yours who happens to be a headhunter misusing your friends list to poach your employer’s staff. And none of your friends will see that you joined some weird group such as ”Accordion players with narcolepsy”.