The fog of commentary, marketing hype, and hysteria over the need for AT&T and the Open Software Foundation to merge has obscured the single most overlooked fact about the Foundation – it’s a software vendor, plain and simple. And the idea that it could act as a replacement or substitute in Europe for X/Open, which […]
The fog of commentary, marketing hype, and hysteria over the need for AT&T and the Open Software Foundation to merge has obscured the single most overlooked fact about the Foundation – it’s a software vendor, plain and simple. And the idea that it could act as a replacement or substitute in Europe for X/Open, which has been suggested more than once, is bad. It won’t even absorb some of X/Open’s responsibility, which is good. Fortunately, X/Open spokespersons themselves see X/Open as a real specifications body, while OSF is an implementations group. And that interpretation in itself says plenty about OSF’s emergence as a marketing organisation second to almost none. After the first formal OSF meetings in late summer, we’ve had a good look at OSF in action. The Foundation has set up strong marketing activities, including advertisements, public relations and so forth. Why? And why is OSF on a crash schedule to pick a standard user interface even before the market has selected a winner? Why, in the name of establishing an open standard, is a second Unix being offered, just when Unix versions were starting to merge? Why could it be bad for the industry if AT&T and OSF merge their development activities? Not for profit OSF has been set up as an independent corporation. It’s activities are useful, and membership appears to be worthwhile. The OSF staff is developing a sense of its mission, and a determination to accomplish it. OSF management is acting in good faith – but a key OSF characteristic is largely unstated. My impression is gained from attending the first meeting, speaking with management, and hearing and reading the information distributed by OSF. Keep in mind that I am trying to read between the lines; you won’t find an official OSF statement of purpose that agrees with my interpretation. Forget about foundation and not for profit and shared development effort and unifying the industry. OSF has a unique marketing strategy. It’s a fascinating and laudable experiment attracting the deep interest of the entire industry. But we need to understand that the OSF management and organisation will not act as philanthropists. They are not like the rest of us – their organisational imperative is to grow in a selected market segment, like any other business. This is not bad, but it differs from what many imagine. There is considerable supporting evidence. Let’s answer some questions from the viewpoint of OSF as a software vendor. For example, why the heavy marketing effort? We see extensive advertising, major public relations, market attitude surveys. All of this is well in advance of OSF providing any concrete member service or delivering any software. For a new non profitmaking foundation acting as a benefactor to the industry, its inexplicable. But for a new software vendor building image in order to carve out future market share, it makes plenty of sense. Why the panic to choose a user interface? Only now is open windowing technology reaching a significant number of Unix users. There are many good contenders, but none have yet had a chance to be endorsed by the market. If OSF represents an industry consensus to standardise on an interface, its too early. The market hasn’t spoken. A group of only four OSF staffers will examine 23 proposed technologies, take input from members, and make a decision in only 60 days. If they can’t decide in that time,the OSF president will choose. For an organisation setting shared industry standards, it’s not easily explained. But for an aggressive new vendor, there is an Zobvious need to grab market share by moving quickly with an early product. Why introduce AIX, seen as yet another version of Unix? Especially disturbing to independent software vendors is the reluctance to commit to binary compatibility with existing Unix standards on mass market chips such as the 80386. Wouldn’t it be more open to allow plug and play software applications on this popular architecture, since all 386 vendors now agree on one standard? The technical obstacles are trivial. But if OSF is a software v
endor, it makes sense to seize the initiative from its competition (AT&T and Sun) and sell OSF-ix as a new, incompatible wave of the future. Will OSF be the central source of supply for standard open platform software? Listening to the OSF presentations, one quickly learns that the Foundation presumes it will. The pitch involves the whole industry saving efforts by coming together around shared development. In this scenario, OEM customers will differentiate themselves only with proprietary applications, but will share OSF operating system and tools. Is this realistic? Certainly, those OEM customers must conform to open systems. But given a standard, there is no reason to think either should be the only implementation. The Foundation has ambitions in operating systems, networking, compilers and other areas – all platforms rather than applications. But in the world of pre-packaged software, these comprise the majority of revenues. Will the rest of the industry simply roll over and play dead to make room for the OSF? I don’t think so. Strange situation One good illustration is the OSF competition to select a Mac-like Unix user interface, with companies offering technology including DEC, AT&T, Hewlett and Adobe. Without a market leader, does any of us expect the losers to write off their investments and quit? Once people see OSF as a software vendor, it will have to compete just like other vendors. Even the OSF founders are not obligated to buy from OSF. There are other ways in which OSF behaves like a software vendor. At the first general meeting, there were some indications that OSF was considering licensing software specifications on a per-CPU basis. In effect, anyone trying to follow a standard set by OSF without buying OSF software might find themselves the target of a suit. This is not and won’t become OSF policy. But it is a strange situation for a standards body to contemplate. On the other hand, it is far from unusual for a proprietary vendor to think this way. OSF, whether or not it is not-for-profit, can be run like a business with an operating surplus fed into expansion of activities, including philanthropic activities such as the OSF Research Institute. It wouldn’t be surprising to find competitors one day contesting OSF’s tax-exempt status. In sum, OSF benefits the industry by keeping the original vendors of Unix honest. If AT&T and OSF merge, we’ll be back to the old single-source situation. And talk of OSF replacing X/Open is in nobody’s interest, and it won’t be, even when some of the results from the consortium start to show up on either side of the Atlantic.