The IBM ‘plan’ to destroy Windows (take 2)

Late last month something purporting to be an IBM Network Computing Division confidential strategy document was leaked to an OS/2 Web site, pulled from the site at IBM's request, because it contained "inaccuracies," and then replaced by a 'corrected' document furnished by IBM. In the gap in between, US mag Smart Reseller published a screamer based on the original document and, wouldn't you know it, the two versions don't seem to have all that much in common. Version one seems to have been a broad range strategy document describing a coherent, ambitious, broad-ranging and integrated plan to leverage relationships with companies including Sun, Netscape and Hewlett-Packard to wipe Microsoft off the desktop and supplant it with Java. It also includes some bad news for OS/2 users, e.g. "we now see OS/2 as a transitional platform over the next three years. Eventually, customers will have to move to another platform or accept a lower level of function if they stay on OS/2." Version two, on the other hand, just includes the bad news for OS/2 users. These sections of both documents have generated a lot of traffic in the OS/2 advocacy Usenet group, but the users aren't really being told anything they didn't know - IBM recommends that OS/2 Server users migrate to other server platforms as scalability requires, and that they migrate client development to 100 per cent pure Java and NC environments. Some of the more tragic postings reason that, as IBM says the original was inaccurate, they shouldn't believe any of it, but wait for the accurate version. The Register reasons that people waiting for the word from IBM had a lot to do with where OS/2 finds itself today. But the non-OS/2 stuff, which Smart Reseller interpreted as a 'plan for domination,' is a lot more interesting. Version one was a 12 page IBM strategy paper detailing a plan to blitz Microsoft off of corporate desktops via alliances with resellers, Hewlett-Packard and Sun, and by keeping together a focussed anti-Microsoft coalition that acts as a single competitor. IBM's angst at its publication is understandable, because besides sending OS/2 users ballistic, it sailed perilously close to all sorts of legal hot-buttons, and looked like it would be well worth a Redmond subpoena. It calls for the promotion of Netscape in order to keep Navigator up at a 30 per cent share of the browser market. This aim is at least arguably anticompetitive - if the whole of IBM standardised on Navigator and promoted it heavily the company still wouldn't have the heft to do this, so the implication is clearly that IBM has got to march in step with several other companies. So if the plan won the approval of IBM's high command and was implemented, this could be the anti-Microsoft alliance 'smoking pistol' that Bill Gates's lawyers are currently searching for. Even if it didn't, the Redmond attorneys could still claim IBM had been thinking about destroying Microsoft. But keeping Internet Explorer out of a dominant position is only a means to an end - the real strategy centres on Java, so unsurprisingly the recommendations include close relationships with Netscape, Sun and HP. The document calls for a partnership with HP to promote embedded Java, help for Sun to enhance Activator, which allows 100 per cent pure Java apps to run on IE, and it sets targets for Java uptake. By 2003, IBM should make sure at least 50 per cent of PCs run 100 per cent pure Java, and by 2001 50 per cent of new thin client types should support 100 per cent pure Java. The support of Netscape together with use of Activator as a 'secret weapon' to subvert Microsoft's browser and its Java strategy seems to come close to a viable strategy for achieving the PC industry target numbers, although as Microsoft has proved with IE, market penetration is a lot about promotion and distribution, neither of which IBM is particularly good at (e.g. OS/2, the last plot to wrest control of the desktop from Microsoft). But by recommending an alliance with HP, the authors are effectively suggesting IBM walk into a minefield. At the time the document was produced (July) there was considerable friction between Sun and HP over embedded Java. This now seems to have passed (HP and Sun make peace over Java), but even now dealing with HP on embedded Java would still destabilise IBM's relations with Sun. HP is producing its own JVM, but is swearing it will adhere to 100 per cent pure Java. It's also one of a number of companies who've wanted to loosen Sun's hold on Java, so an embedded Java alliance with IBM would look like (and would probably be) a bid to break away from Sun - so what price the alliance? An aside here though - as HP is both a down-the-line Windows NT supporter and an advocate of 100 per cent pure Java, is it not grooming itself as the company people are going to have to pay court to in the Java-Windows space? A desirable ally for all concerned... But back to the strategy document - is it really significant? We don't know what level it was produced at, and we don't know what its status was. Much of what it says is completely logical and obvious, and the sort of stuff you'd expect the Network Computer Division to be talking about. Obviously, if Microsoft does manage to lock the Windows market into its own semi-proprietary version of Java, IBM's ability to sell its own 100 per cent pure Java systems will be severely impaired. IE is a vital Java distribution mechanism for Microsoft, QED stop IE. HP is a possible cuckoo in Microsoft's nest, QED some more. The Network Computing Division looking for Java allies and binding them closer together is also a logical course. IBM is already allied with Sun and Netscape, and HP's known to have been looking for friends at the time the document was written, so this part of the plan is also rather than revolutionary than you might initially be led to believe. But what we think really undermines the document is its origins. It comes out of one IBM division, but if it was going to be an IBM master plan to destroy Microsoft, as our friends at ZD present it, it would have had to have the support of a whole range of different IBM operations - we love NCD dearly, but it doesn't have the weight to determine strategy for other IBM operations, never mind co-ordinate strategy for an industry-wide anti-Microsoft alliance. It's maybe a discussion document with elements of wish-list attached, but that's about all. If we were Microsoft, we'd still subpoena it though. ® Click for more stories