"Always certain, sometimes correct" was the way I was once described by an editor who twice fired me. Consider those words as a caveat, then, surrounding these year-ending, year-beginning predictions for the upcoming twelve months. Remember, I am a professional, so don't attempt this at home.

The wise prognosticator knows we all have a tendency to overestimate change in the short term and underestimate it in the long term. What this means for us in 1997 is that major players are unlikely to be either born or die. That is, if a significant new company comes along in 1998, it will probably not achieve that significance before 1999. And while a big company or two may die this year, their brand names at least are likely to survive.

Nevertheless, here are the trends:

1) Consolidation, consolidation, consolidation as merger and acquisition fever continues, this time pushed by rising wages, Asian economic jitters and a cooling economy. Don't be surprised to see a major networking company acquired, as well as a major database company. I'm not saying Cabletron and Informix are for sale, but they look mighty attractive to the right buyer. More M&A news awaits further down.

2) Rising from the ashes is good, and we can expect some of that from Digital Equipment and Oracle. In the case of DEC, it couldn't go much lower. The recent deal with Intel gives Digital not only some needed cash and lower expenses, but a completely new outlook on its Alpha system business. DEC is already selling Alpha motherboards that offer a price-performance ratio Intel can only dream of, and with Intel now building the Alphas, DEC is going to try very seriously to stick it to their new manufacturing partner.

As for Oracle, the recent financial jitters reflect some real problems with the company -- MAJOR problems. For one thing, Oracle's new vision is based on paper-thin substance. Network Computers? No, thanks, I already did the "mainframe thing" and have no desire to regress. Network Computer Architecture? Oracle's own applications are not even NCA-compliant, and when a vendor starts pushing "wrappers" as its key architectural element, it's time to get a new candy bar. Verticalization? Great idea, but Oracle has little in the way of products to back it up. The only verticalization that has been accomplished is Larry Ellison getting out of bed....One Team? Oracle is more fragmented now than ever. Applications? When you have to emphasize "wrappers," you know there can't be too much beef.

If it's all so bad, why then do I see 1998 as a better year for Oracle? Because the company has been here before and recovered smartly. Remember the financial debacle of 1990 when Larry Ellison had a 23 year-old controller handling the books? Then there's the fact that Informix's problems are even worse. That company's demise could only help Oracle. Sybase, too, is no pillar of corporate strength. And I actually view Oracle's recent VP losses as a potential gain for it, because there was SO MUCH dead weight, particularly at the Corporate VP level. Oracle should recover smartly (if miraculously) by the second half of 1998.

3) Intel, which should be hit with ever more slender margins in this dawning age of the $500 PC, will instead ride its new Merced processor to profit margins that are about the same as this years'. This is a big one, because Intel really needs something to combat AMD and Cyrix/IBM, and that something is Merced, which simultaneously eliminates the RISC threat from HP while bringing-in another major customer -- Apple. It's 1999 when I'd worry for Intel.

4) Remember Java? Well 1998 may well be the Year of Java, but probably only for IBM, which has put more people and more money behind the programming language than probably all other developers put together (even Sun, which owns Java). Look for a succession of product introductions, both in applications and programming tools, that will give IBM a major voice in the world of Java. And the point of it all is not to beat Microsoft, but to further refine Big Blue's superbly profitable consulting operation, where Java is the strategic CASE in point.

5) If The Graduate taught us that one word -- "plastics" -- described the 1960s, the late 1990s can be described by one acronym -- DSP. Digital Signal Processors will appear in 1998 that for the first time combine massive computational capability with, in some cases, the ability to run general purpose operating systems. Even those high-end DSPs that can't run a mainstream OS will be tightly-coupled with cheap RISC CPUs that can. The result will be a new generation of sound, video, and networking products -- many of them standalone -- that will start in 1998 and continue through the end of the decade. We're talking about 1400 MIPs Internet telephones, real-time MPEG encoding/decoding, and super high-end programmable modems. We're also talking about yet another generation of video games at today's price points but offering 3-D rendering where individual pixels can't even be seen.

In 1999, DSPs and general-purpose CPUs will start fighting over the new digital TV market. This could be more bad news for Intel.

6) Godzilla meets Megalon: Compaq and Dell go head-to-head in the made-to-order PC business in 1998. Oddly, while there is unlikely to be a clear winner between these two, there will be clear losers among the second-tier cloners. Micron has chips and memory modules to fall back on, but Ted Waitt at Gateway should probably have taken Compaq's offer from earlier this year, because his company will never be worth more.

7) The Internet makes money, finally. Not tons of money, but some significant coin. And it's about time. But what about the supposed shakeout among ISPs? The Baby Bells were supposed to take over the Internet service business, but now they see selling backbone services and T-1 connections as a lot more profitable than getting $12 per month from you and me. There will be lots of fumbling here as cable TV companies start installing phone lines because cable modems still aren't ready, while traditional ISPs turn themselves into CLECs (local phone service providers) to claim cheap xDSL connections from local phone companies. This means lots of action, but again no clear winners or losers until 1999.

8) Windows 98 is late. No surprise here, but if Microsoft does its job right, this will be the first OS from Redmond that can be upgraded over the Net, moving Bill G that much closer to his goal of renting, rather than selling, software. One thing Bill will stop paying rent on, though, is MSN, which should bite the dust by mid-year. The rest of Microsoft will rumble on, generating billions in profits and more or less ignoring the Department of Justice.

9) Novell still won't find its marbles. It's sad, but not even a smart guy like Eric Schmidt can turn this one around. Novell will be looking for a merger by late spring. But don't expect the merger partner to be Apple, which will probably return to marginal profitability by mid-year, but will still need a miracle to regain its old magic. Of course Steve Jobs is not a stranger to miracles.

10) I will continue my relentless one percent per year weight increase. Lardbutt lives! Please pass me a doughnut.