Extremely interesting question. If you look into it, Java has some major security problems, problems that probably (note probably, not looking for a debate here) cannot be solved without going to a more centralized model of processing. That is, more "dumb terminals" hooking to the net rather than smart boxes. This isn't in the cards either from the consumer perspective or from the MS perspective. Dumb boxes don't need anything other than the "Java OS" to connect them and to then dole out what the individual "needs" from the big iron. The Sun/IBM model of the web includes selling software by usage of CPu or some such rather than selling shrink wrapped packages, which is probably the major reason that Sun et. al like the government ideas related to regulation, etc. So, MS isn't anything like a champion of the individual, but then again, unless it champions the individual with his own processor and machine(s), it has to ignore the successful business model that built the company. MS has never fared well in the arena that requires centralized processing at the expense of the individual machine.
Telescript has a more complete model of remote execution, and if it hadn't been tied to the initial PDA/Newton craze by folks with an Apple background ("build a better mouse trap and then control all aspects of it") it wouldn't have fallen to the point it did. That GMGC has survived is both a testament to the company and the underlying technology. Odessy (sp?) is an alternative implementation of the Telescript/Agent idea, albeit one which is layered on Java and so can take advantage of that standard to leverage itself beyond it. MS will, IMHO, create new tools to enable individual developers to access the underlying technology within Portico as well as to take advantage of that technology for other reasons. In addition, MS may well embed the required agent arbitration and regulation protocols and tools within Windows NT. That would give it a leg up on the competition in more ways than one. It will get very interesting. It is already very interesting to me that GMGC has chosen NT as the basis for its' NOC, and that GMGC has chosen to not tie itself to the larger carriers right away as it did with the Magic-Cap deal for PDAs (one service provider - AT&T - and only a few licensed hardware vendors, with the utility being available only for some small subset of the hardware on earth), and last but not least, that it has chosen a wide open area of existing technology to use as a showcase for it's underlying efforts rather than betting on some "next great thing" technology that can only be accessed with the purchase of new hardware.
I get kind of tickled at the Portico / voice interface debate and all that "is their recognigion ...." type of question. That's a single interface, one of many, and one which gives good demo but isn't really the major interface everyone will use or be required to use. If you're familiar with the old days, then you may recall the great (and still ongoing) processor debates when the 8088 / 68000 systems were young. Sure the 68000 was pure, and better, and everything else. But the 8088 was in all the open systems. Same for Telescript/GMGC this time around (wow, so few companies get a second chance). It will be in all the open systems (open in a hardware sense, don't get into the wadded panties MS hatred mode), with GMGC there to add refinements and to build on its' success. Whether or not Telescript as such continues remains to be seen, but the underlying technology has refused to die and hasn't been well imitated. The alternatives are less complete and less secure as far as I can tell. They are also less refined and haven't been out of the lab much as yet. Telescript stuff has been through the wringer, albeit one abbreviated by bad business decisions surrounding product and service it was originally introduced with. Competitors did stuff like comparing agents to virsuses (during the virus hype phase of the internet) and explained how their good Java stuff was more confined and controlable, ignoring the security flaws in Java, of course. Other than that, I never saw anything about Telescript failures and foibles. They may have been around, but missed them if they were.
Giving the individual the keys to the web rather than the keys to a service is what agent technology is about IMO, and GMGC has the tools and talent. Alternatives are not without their good points, but they all appear to suffer from the same centralized control perspective that is actually the antithesis of the underlying technology. Sort of like killing the goose by keeping it in a cage designed for a hummingbird. Some companies don't want everyone with a PC being able to sell little software agents CPU cycles while they sleep. Such biz should be reserved for the vendors of big iron, in their opinion. Well, Telescript and/or its' kindred software could change that. Companies could bid for idle cpu time on any box anywhere on earth. Who knows what awful stuff would appear when J Q Public can purchase some serious CPU time on the web and even has a software agent going around and bidding on it for him. "Hey, leave your box on every night and it'll pay your ISP bill" could become a serious marketing tool. Maybe people would gang the cheap stuff into effective big iron rather than buying big iron. Dangerous thinking, eh ?
Disclaimer :
All opinions my own. I reserve the right to not give a damn if you agree or not, and will accept full credit or blame for my own mistakes and successes. I further do swear that I thought this crap when I typed it in but may not think the same thing later.
Regards |
