I have been lurking on this thread for many months, and have enjoyed and benefited from the comments on SUNW and the industry. I feel I must comment on the possible impact of Windows XP, from the point of view of a software developer. One warning: this is a long post.
Windows XP is, from a development point of view, the most significant OS release since Windows 95. I'll explain why, again from a software point of view. I'll leave it up to others to decide how this affects the business side (in fact, I hope to benefit from such a discussion :)
Prior to Windows 95 was Windows 3.1, which some of you will remember as an unholy, unstable mess. In fact, Windows 3.1 barely qualified as a stable OS, but it was quite frankly the FIRST version of Windows that could be successfully used in a professional environment. Windows 3.1 was a "16 bit operation system", meaning that it used 2 byte (16 bit) quantities to represent memory. This architecture was a severe restriction, reducing the amount of memory that the OS and programs could use, and introducing inherent difficulties in programming that caused huge instabilities. Programs that only crashed once a day were considered 'stable'.
Windows 95 was a significant improvement because it changed the OS from a 16 bit to a 32 bit (4 byte) architecture. Without getting into gory details, this meant that programming was much easier, and stability improved substantially. It is my opinion that the success of Windows 95 was led not by consumer desires and features, but by the fact that corporations led the migration through business necessity. Businesses found it necessary to upgrade because of the huge productivity savings in software support and development (fewer crashes and lockups, cheaper software development). Software suppliers were quick to drop support for Windows 3.1, because it required code that was substantially different from Windows 95.
By contrast, Windows 98, 2000, and ME were, from a software point of view, actually counterproductive. They did not offer substantial savings or advantages (again, from a software point of view). In fact, they were counterproductive in that software vendors found that they had to support an increasingly complex matrix of platforms (hardware and OS versions). Since Windows 98 et al were not 'compelling' upgrades to businesses, many still run Windows 95, forcing software vendors to support the whole gamut: Windows 95, 98, 2000, me.
As if the story were not convoluted enough, adding to this complexity is the Windows NT OS. Originally conceived, I believe, as an ultimate 'true' multi-tasking business OS, it turned into its own 'branch' of development at MSFT. Windows NT is a 'real' OS, in that it has the features necessary for a business to rely on 7x24x365. For business applications that must run continuously (any server applications and many critical business applications), Windows NT is a MUST for businesses.
For some bizarre reason, Windows NT continued on a life of its own, with regular upgrades and new releases. Windows NT 4.0 Service Pack 5 is the 'standard' release that is popular with many businesses. So, this increased the complexity of the software support matrix; in addition to all the flavors of Windows desktop OSs, add Windows NT, further increasing the software support and development costs.
At this point (if you're still there :) , I have to point out the tremendous costs associated with development, maintenance, and support across this huge OS matrix. IT organizations go crazy trying to maintain a stable environment with these combinations, but in many cases have their hands tied. There is no compelling reason to upgrade users, and in fact no 'ultimate' destination to upgrade them TO. End users would and could not use Windows NT; it is a developers OS, and has few of the advanced end user features that users demand today. Windows ME/2000 are not compelling enough to convince the CIO/CFO/CEO to undergo the cost of upgrade.
Now comes Windows XP. Windows XP is actually the next verison of Windows NT (5.1). It has the roots of a 'true' operating system; stability, true multitasking, true network directory support, real backup support, etc. It will run for days, weeks, months without rebooting (of course, after they work out the bugs :)
Windows XP will become the replacement for all prior Windows OSs. I believe it will become the 'compelling' event for upgrading. IT managers/CIO will finally be able to convince the rest of the corporation that they will all benefit from the upgrade to a singular OS that contains the necessary user and IT features to run a business.
I do not believe this will happen all at once. Obviously, first it has to be released. Then, it has to become completely stable (why this doesn't happen in reverse order is somewhat complex; suffice to say that the world is part of the Microsoft QA organization). Next, software vendors must fully support it, and provide an upgrade path to it. At that point, the CIO makes his/her argument about such-and-such savings in support and increase in user productivity, at which time a migration plan is constructed and executed. Finally, support for prior releases is dropped in the corporation.
I believe that the consumer migration will either parallel or actually lag this corporate migration. Believe it or not, many users 'borrow' software from work, and want their home work environment to match their work environment. Once they see the benefits of Windows XP (stability, backup), they will also want to upgrade. Additionally, newer versions of consumer software will only support Windows XP (thus reducing the software development cost, and therefore the software product cost). This will provide the compelling event that finally pushes the consumer market over the edge.
So, what does this mean? As others have pointed out, it means that certain companies will benefit. Obviously, Microsoft will sell a huge number of OSs and software upgrades. But, Windows XP requires more computing resources, so other beneficiaries will be PC hardware vendors (P4s and memory), and ultimately software vendors (who can charge a one-time upgrade version, and will have lowered development/support costs).
Who will lose? That is a good question. The Sun/MS battle has little to do with the desktop; MS has already won that battle. Java is not the lynchpin; the crux is OS performance and stability. UNIX is still perceived to be the king in that arena, but MS is pouring huge resources into changing that, both from perception and reality. Ultimately, I'm afraid that SUN must lose and MSFT must win. Once IT professionals accept a (future) version of a MS OS that is as fast and stable as UNIX, the walls around the SUN castle will indeed crumble.
So, you say, just another MS clone spouting kool-aid? Actually, currently I am long on SUNW and have no MSFT holdings. Why? Because I believe SUNW is a 'value' play, and MSFT is overvalued. I know that sounds funny; SUNW is certainly not cheap (but getting cheaper everyday :) , and MSFT will certainly not seem overvalued in a few years. But, I believe the migration I've outlined about, and the resulting demise of SUN, will take a number of years to occur. In the meantime, I believe SUN will have one final rising over the next 1-2 years, and then slowing set forever...
These, of course, are only my own opinions. I apologize for the length of this message, especially as a first post, and congratulate any who have read through to this end. |
