Vinod. Here are some musings of mine on the subject of Java, NC's and other sundry items (I said I would post tonight). These are my observations of the current "climate" for what it's worth.
======================================================================
On Java in general:
Should the dominant target API become something portable, then OS's become something that must compete solely based on their capabilities, and not on what apps they have available.
But portability is not the goal, it is the subterfuge. If Sun could convince the industry to standardize on its API, the underlying OS then becomes captive to their initiative and thereby becomes essentially irrelevant. At that point, Sun's API becomes -- for all intents and purposes -- *THE* OS.
I personally thought Sun's ruse was at least remotely possible for a short period of time. Or at least one always more credible than any of the previous major challenges to the WinTel duopoly. A series of events ending with the August 8 edition of _Information Week_ have convinced me this is no longer even *remotely* possible. The cover features the arrival within the next six months of the under $1000 (as low as $800), fully-featured, *manageable* PC from the major vendors. These machines will combine the best features of the brain-dead NetPC (the initial response to the NC) with those of a standard PC. They are targeted for corporate networks and specificly positioned *against* the NC.
It's going to work. The NC (as defined by Sun/Oracle) is now DOA.
But there was something more. A sub-heading for the cover could very well have been: "Sun Microsystems: RIP". I firmly believe this marks the beginning of the end for them. With the critical delay of their Java performance release until next spring coupled with increasing public skepticism for Java as a platform and the pending death of the NC while still in the cradle, they will be forced to increasingly
rely on high-end UNIX for decent revenue and growth. An omen for the future can be seen in SGI, who has all but thrown in the towel on the low-end workstation market by announcing a WinTel machine for next year. That's a significant development. Sun is currently relying on its high-end server business to counter losses in workstation revenue, having shrunk from over 70% of revenue to less than half in recent times with growth in that end flat. In order to stop Microsoft's relentless momentum up their tailpipe, they absolutely need to cut off its head. And that head is Microsoft's control over Win32, the industry's increasingly dominant OS API standard. Fail there and its only a matter of time before Microsoft pushes them over the cliff as they move into the final bastion still dominated by proprietary
vendors -- the high-end market.
In my mind, they have already failed. Like all of those testosterone-laden egos who have made this attempt in the past, the window of opportunity to shift the Big-MO into a new direction has passed them by and they have only succeeded in putting themselves -- and their company -- into Microsoft's crosshairs as it -- justifiably, BTW -- positions itself to diminish a serious threat to their future.
On Java, "specific":
An OS provides a lot more services than what the typical APIs are that an application programmer uses. Nobody buys an OS, they buy "applications" (not NLM's). That is the whole point of an OS -- to service that vital need. To the extent that a reliable and known infrastructure exists to service that need is the central force behind the drive to a single (or at least limited) standard and the push toward WinTel as a "natural monopoly". They (particularly Microsoft) are aware of this and very much exploit that natural tendency to their own advantage. Microsoft especially has been both lucky (by having been at the right place at the right time) and highly skillful in this regard. But that the tendency toward a natural monopoly exists is precisely what McNealy, Ellison et al are attempting to usurp for themselves. Like Cannivino, Noorda and many others before them, they are driven by ego and a desire to be "number one" -- to take out Microsoft as the dominant force. (read: Novell is now blowing smoke). Like I said before, their earnings will be "FLAT, at best...". Anyway, that ego can be fatal when judgement is impaired as a consequence of it, as demonstrated time and time again (read: Novell 95-present). The counter-attack was spotty in the beginning because Microsoft did not appear to understand quite how to respond. The NetPC was a pathetic and desperate first attempt, and doomed from the start. IBM wisely chose to drop out of this idiocy a short while ago. Others will probably follow before long, in spite of the fact that Microsoft is still giving lip-service to it. Hydra and Winterms (Wyse, etc.) will be a better direct counter in certain niche markets where an NC-like device makes sense, which it does in certain -- albeit limited -- environments. However, the low-cost standard PC response to the NC appears to be entirely a HW vendor driven initiative.
That one makes far more sense than either the NetPC or the NC (including the repackaged X-terminals currently being advertised and sold as NC's and using server products such as NTrigue (Citrix really) to run Win32 apps) for most corporate network situations.
On the $299 home (NetPC) home box:
These make absolutely no sense whatsoever to me. However, the corporate world might be tempted by the hype if decision makers happen to be prone -- as cited recently -- to MBM (Management by Magazine) or MBC (Management by Commercial). None-the-less, although these boxes may appear cheap on the surface, it is only because of cost-shifting. Servers and network infrastructure is where the expense winds up
traveling, so there is no real savings to write home about after that "other" stuff is factored in. There could be some short-term advantage in support -- but only if there were applications to run on them, which there aren't. By the time those critical applications could even hope to appear, the WinTel industry will have had more than enough time to shore up a genuine weak area, which relates to the expense and
difficulties of managing the desktop. As it stands, this area is a real PITA for anyone with responsibility for large numbers of them.
My conclusions:
If Eric Schimdt and company pull this off, then it will well be worth it for me to lose any gains I might make by pulling out of this stock completey by year-end. Why do I say that? Well, because I *do* want Novell to succeed, but I can't wait for them any longer.
Morose? No, I don't think so. Pragmatic, yeah, I think so.
BTW, I "charge" for this kind of stuff <G>, but I figure "what the hell" -- perhaps someone can use this kind of stuff in a case study somewhere <smile>.
And this is a company that is banking their entire future and existence "solely" on Java and JVM's to propel themselves back into prominence again ...
Again, "morose", no .... "pragmatic" at this point.
Regards,
Joe... |
