MWAR jumped on Digital TV early on, starting with Philips and CDi. Like Laserdisks, CDi disappointed, but then MWAR turned to set-top boxes, dominating what also turned out to be a disappointing market to date. Both INTS and WIND took a more wait and see attitude about pursuing a market that held only promises; albeit very big promises. Nevertheless, MWAR has spent big and persevered, and deserves any and all set-top box design wins that comes its way.
Meanwhile, both INTS and WIND have set-top box design wins also. INTS won the HP set-top design a couple of years ago, and WIND scored with Hyundae of America. In addition, both INTS and WIND, but I don't think MWAR, are porting their RTOS to work with OpenTV from the Thompson/Sun partnership.
As I understand it, MWARs DAVID RTOS requires that titles be "authored" to be DAVID conversant, probably in a fashion very similar to how DVD titles will have to be authored. DVD titles not only involve exotic audio and visual compression (AC3 and MPEG-2), but DVD requires the allocation of bits about bits, called a bit budget. These bits provide various controls to a generic client, such as movement, menu selection governing choice of language, aspect ratio, piecemeal censoring or possibly even choice of ending. If Java applets are to be supported also, then the relationship between the server and client must be expanded even more.
No wonder Interactive TV has not been implemented on a large scale. Just the authoring of movie titles alone is a show-stopper (pun intended). However, if titles are authored on-the-fly digital, in analogue, or cheaply in digital with niggardly bit-budgets, then functionality would be limited, perhaps unduly. It would be unfortunate if someone digitized thousands of titles and authored them all for Interactive TV, only to discover that a critical bit was missed, or authored them consistent with a "Beta" standard when a "FHS" standard became dominate. For reference, it would take experienced technicians, with all their digitized inputs and a pre-conceived notion about a bit-budget, two or three days to author a DVD title, using three integrated workstations capable of real-time compression.
My suggestion would be to tailor a standard around DVD, which itself has already standardized throughout the world, and simply serve up DVD titles among other things to set-top boxes. Assuming the world goes DVD, DVD titles will soon be abundant, with extravagant bit budgets. If the world doesn't go DVD, then there must have been a world war, so it won't matter anyway.
The development group that sets the de facto industry standard for what transpires between the server and client is the key to who will ultimately dominate the set-top business. To find that key, look first at the servers. Who are the players providing server software? How does it comport with OpenTV, or DAVIC? What is the relationship/similarity between a server of interactive TV and an NC server?
I recall that Oracle and Sybase were each investing heavily on the server side of interactive TV, and we know Oracle is very serious about serving up the NC. Are they still the leading players on the server side? If so, what standards are they pursuing?
In the final analysis, it seems to me that the dominate RTOS player in the high-end set-top box space, whatever form it takes, will be the vendor whose product is maximally compatible with industry-standard server requirements on the most processor platforms, enabling the set-top box to provide a host of functions in concert with the server. The successful RTOS also will be compatible with the plethora of possible communication mechanisms and protocols implemented in software to facilitate configuration management by the server.
MWAR deserves this niche, but it is way to early to spot the winner until many complicated factors play out.
Allen, |
