I think this is largely true, but Microsoft is large enough that it really isn't a single entity - the various divisions are outputting different levels of suckiness/goodness. For instance:

DevDiv is on a pretty good roll right now. It is still many steps behind other communities in a lot of regards (and likely always will be), but LINQ was a huge milestone in 3.5 and the new dynamic integration options of 4.0 is a big step forward for a mainstream static language.

The XBox division continues to do cool things. Last year's dashboard revamp was a huge face lift for the system, XBox Live may continue to be a wretched hive of scum and villainy but it is a well-implemented one. I'm skeptical of Kinect's value for the type of games I play, but I believe it'll be very popular amongst casuals and should be a compelling addition to the XBox's UI for non-gaming apps. Oh, and the "instant-on" 1080p streaming of pay-per-view and Netflix movies is really sick and feels like living in the future.

The Zune may always be an niche choice, but its the best option if you'd prefer to spend $15/mo for "unlimited" music rather than purchasing music at $10/album. Which situation makes the most sense will vary from person to person. Last year Zune started handing out 10 song credits per month to each subscriber as well, which helped take the sting out of the normal arguments against subscription services.

I don't particularly follow the Windows Phone 7 / Kin stuff, but it seems obvious they screwed up royally on at least half that proposition (and I don't expect big things out of the Win Phone 7 half either).

Windows 7 has been very well perceived, though it still lags behind Linux/OSX.

Office 2007 was a pretty good redesign of the suite, but Office 2010 doesn't seem to have raised the bar much. I don't think OpenOffice will ever replace MS Office, rather Office will continue to lead until GoogleDocs is fully ready to take over (its getting closer & closer every day and I don't think MS's new web-based Office will be able to compete).

In retrospect, after reviewing what I've written so far, my opinion seems to be that Microsoft's offerrings haven't been that bad execution-wise lately, but that most are plagued by the fact that they are "also rans": decent products that have some strong merits, but offer no compelling reason to leave the market leader or community favorite.

Yeah I was just gonna quote that line as well. Nowadays invention is often meaningless anyway since there are so many ideas and mashups floating around. And even when it does matter, it has little correlation with success.

In any case, Microsoft's success was based on innovation in the business model, and they've done a lot of innovation with their software over the years.

The real problem with Microsoft if the insider stories are anything to go by is that it's full of bureaucratic middle management self-sabotaging with no real leadership. All you have to do is listen to Steve Ballmer bluster his way through an interview to realize that being a good sales guy in no way confers the ability lead a technology company.

However Microsoft still has a ton of successful products and great people. I imagine if they were to do some kind of massive internal housecleaning and layoff of underperforming and territorial middle management based on interviews of the rank and file that they could restore quite a bit of their product development savvy overnight.

I think Microsoft's biggest innovation was not a thing but a process, and it that served them very well for most of their history. Namely "fast follow" - the ability to take a competitor's product and very quickly produce their own, better version. DOS, Windows 3, Word, Excel, MSVC++, Internet Explorer, etc. etc.

Windows 7 seems noticeably absent from the list of stuff the R&D department produced.