Many years ago, I spent countless nights downloading disk images over the internal SNA network, burning the images to actual 3.5 inch floppy disks, and installing the latest build of OS/2 2.0.
It was exciting to be part of something which seemed to be a genuine jump forward in technology.
Windows had stagnated at 3.1, Apple had released System 7 in 1990–1991, compared to these OS/2 seemed to be a major jump forward.
A lot of time (both on the clock and off) and money was poured into OS/2 above and beyond anything that IBM officially tracked.
At the same time that OS/2 2.0 was coming together, the corporation was melting down.
We had AdStar, Eduquest, Pennant, Lexmark, Axciom, Advantis.
Some of these divisions actually split off into separate companies, others operated financially separate but never formally left IBM (until years later in the case of AdStar).
A very destructive environment ensued where divisions billed each other market rates for services.
Morale dropped through the floor.
There was the AMSROUND discussion (think of it as a free–wheeling blog in 1991 era IBM without the comment spam), there were various side discussions in the hallways at Myers Corners Lab, Kingston, Raleigh, and elsewhere.
Putting a hard price on internal resources changed the focus of development efforts and management away from customer facing to internal concerns.
We'd bicker about how much it
cost to host forums (IBM's internal discussion boards), or to transfer files (finding it theoretically cheaper to fly two guys to Boulder in first class than to ship the MVS distribution via RSCS over SNA, and faster too).
We lost focus as company, and posted stunning losses in 1992–1993 as a result.
And as I read the discussions at Mini–Microsoft I can't help but wonder if we're seeing a similar meltdown at Microsoft.
IBM's problem then, and I think Micrsoft's problem today, is the sense that they had to compete at the same level of intensity, support, marketing, profit across a very wide product set.
When I joined IBM in 1990 it manufactured all sorts of computer hardware from PCs through mainframes, desktop printers through high–end high–speed printers, networking equipment, cash registers and point–of–sale terminals, various peripherals to support all of these pieces of equipment; plus software to drive, manage, and connect all of these disparate services together.
It was not only an internal expectation, but external through customers and the media that anything that came from IBM would have the same level of security, service, support, etc. as anything else that they'd purchased from IBM.
It meant there was very little margin for error, products took years to be developed and introduced and were frequently killed off before introduction because whatever revenues they were forecast to earn would not recoup the development, marketing and service cost (even if the development cost had already been incurred).
I remember this era as I watch Microsoft struggle with Windows Vista and other products.
Vista may well have some revolutionary features, but everything I have read about it makes me think twice about upgrading to it.
I held out for close to two years before switching from OS/2 to Windows '95 (because as IBM's webmaster it was getting embarrassing that I could not easily check out our web site given the lack of decent browsers on OS/2).
But I've now gotten to a stable point with Windows XP SP2 and I really can't see changing simply for the sake of changing.
I don't see any value to me in upgrading to Vista.
And that is a similar judgement that I heard from many people in the pre–Windows '95 era when OS/2 had a couple of years' headstart on what Windows '95 would introduce.
The penalty for switching to OS/2 seemed to be too high for people to accept, it seemed to be more than whatever benefits people would gain from switching (in the pre–browser, pre–widespread email era did you really need to run more than one complex application at a time?).
Microsoft now have multiple operating systems across multiple hardware platforms.
It builds some hardware, not much.
It has worldwide operations, commitments, requirements.
It has embraced and extended its organization across the most popular areas of technology today, covering video games, internet operations and applications, operating systems and various desktop and server applications.
And I wonder if Microsoft, like the IBM of 1992, are simply too big, too complex to manage all the interconnections and interdependencies effectively.
We have different expectations of technology conglomerates than other conglomerates.
No one expects interoperability between GE turbine engines and GE clock radios.
Perhaps quality, support, but not interoperability.
We can accept GE as a brand that describes disparate products, but we don't accept Microsoft or IBM (or, I wonder, open source) as brands describing products.
We expect something more, interoperability, interconnections, infinite scalability.
When I was at IBM I questioned the wisdom of bundling a bunch of disparate products together under the WebSphere brand.
In retrospect I think it may have been a smart move, though for reasons that I doubt IBM Software marketing would appreciate: it lowered the footprint of expectations from all of IBM Software and technology to just WebSphere.
So I wonder if the solution to Microsoft's malaise, if there is such a thing, is not to break up the company but break up the brand footprint.
Lower the expectations, you can't manage them as it is across the company, but lessing the span or breadth of expectations they might make it easier to control projects, market products, and improve their image in the eyes of their customers.
And keep Windows XP around for a while longer than planned.
There is no reason to kill a good, stable product (I know, I also got 18 security patches last week, so stable may not be a great adjective to use; but the APIs, driver dependencies, system requirements are pretty stable).
e.p.c. posted this at 19:11 GMT on 18-Feb-2007 from Amagansett, NY.