I have to say that I really appreciate the folks in the Microsoft Data Platform Group, and the SQL Server team in particular. No, really! Not only because SQL Server 2012 is such a solid product, but also because they leave features in SQL Server even though they probably cause themselves all sorts of grief as a result.

However, in so doing, they avoid causing anguish to their customers who, through no fault of their own, have to use the sort of archaic technology that one associates with flares and bad hairstyles. I like SQL Server Rules, for example. Although they threaten to take them out, the butcher's hand has always paused. DMO? They kept it all the way to SQL Server 2012. The sp_makewebtask extended procedure survived for ages.

Microsoft's SQL Server team looks after its customers even though it often means leaving in features that don't make commercial sense to accountants, or technical sense to those developing the product. The requirements of customers generally come first. If enough customers want something supported, they'll usually support it. We all make the odd complaint, but overall they seem to listen.

My renewed appreciation more "customer-focused" culture of the Data Platform Group springs from its stark contrast to recent decisions in other parts of Microsoft. I was particularly shocked with the demise of Silverlight as an active platform so soon after the initial hoopla, but there is quite a list of products, most of which I use or have used, that are no longer supported or have become 'legacy'. Without thinking too hard, there's Visual Foxpro, Live Mesh, Silverlight, Zune, Netmeeting, Windows Media Center, Flight Simulator, MS Money, Expression Web, Expression Mesh and Windows Live Messenger.

I am particularly grieved about Expression Web, aka FrontPage, which is an extraordinarily good product, even though it drifted from its designer roots into being a bit developer-oriented before it expired. Sure, it is a free download now and I will use it as long as I can, but I prefer to use applications that get the love.

There are always good reasons for discontinuing a software product or a feature. Mistakes, or acquisitions and mergers, can make a product line difficult to explain. Sometimes, circumstances force us to rename a product or consolidate two developments. Occasionally, as with Silverlight, technology just reaches a tipping-point. It is possible, as Microsoft's server group has demonstrated, to do this whilst causing as few problems to customers and supporters as possible. Somehow, the "Metro" naming farce, the dropping of Expression Web, the sudden renaming of the Azure products, and other eccentric decisions all make one appreciate it the more when things are done right.

Phil Factor