XP Embedded: Alive and kicking, or dead and twitching?

I have a longstanding frustration with commercially pushed “obsolescence”; I define this as a migration path pushed for nontechnical motivations, sometimes even flying in the face of technical reason.

The retail industry has been guilty of this for decades, cleverly scripted advertisements to convince one that the usually minor, usually superfluous features of the shiny “new” product are the opposite, with the purpose of leaving the individual wondering how they had coped so long without them.

However, the embedded/industrial computing industry is a far cry from that world. Reliability, longevity, and stability are kings here, but the ever increasing need to drive profit from new products constantly threatens to corrupt these fundamentals – particularly when coupled with falsified scaremongering about the “old” product’s impending demise, designed to panic the customer. This then starts bordering on the unethical.

Whilst parallels can increasingly be drawn in so many industrial product ranges, as an MCTS of XP Embedded/WES 2009 (from here on “XPe”), I must fight on the side of one of the most popular GUI-based embedded operating systems of our time.

Following a call from a panicked client, having heard on the grapevine of XPe’s looming quietus, it transpired he had seen press releases relating to the retail, desktop version, XP Professional. This is an ambiguity we should first resolve, such that we can be confident in what is supplier scaremongering and what is more client confusion.

The purely desktop versions of XP, “Professional” and “Home,” had an announcement of a cessation of “support” of April 8, 2014. Whilst in my learned view Microsoft’s announcement wasn’t ambiguous, a proportion of the embedded public erroneously took this to cover all versions of XP, including our own, cherished, XPe.

It’s important that we understand what is meant by the announcement itself, before we concern ourselves with what versions it affects: Simply, this refers to no further Service Pack or security updates. Embedded developers abhor updates, which can threaten the stability of a deployed, inaccessible, system and one that is rarely connected to the Internet directly, such that a need to protect itself (at source) is necessary, akin to the very small percentage of such systems running resource hungry antivirus software.

The other aspect is the advised lack of new hardware driver support; this is ambiguous and could easily be perceived to be equally true of new embedded hardware. Embedded manufacturers still ensure their systems support long-consumer-forgotten legacy OSs (even DOS) as the continuation of the OS platform is so often desired for customers forced to upgrade, through hardware obsolescence, for example.

Changes to both naming conventions and licensing arrangements by Microsoft have made a disjointed product range more coherent to fresh faces, though to seasoned developers, the road map became somewhat perplexing.

It’s worth noting now that a well-known global electronics distributor adds fuel to the fire of ambiguity by stating openly on their website “Windows XP Embedded (Now Windows XP Professional for Embedded Systems),” which is simply false.

Having addressed those two aspects, we find neither has significant bearing in the embedded marketplace. With no announcements regarding cessation of availability of licenses (and why would they, logically?), one has to question exactly in what way is this highly flexible OS now suddenly unsuitable? Its low-performance requirements, wide driver support, low licensing costs, and excellent embedded-only features make it ideal for so many mid-range applications.

Thus, it’s disappointing that XPe is dragged into the cloud of confusion regarding its obsolescence to create fear and pressure customers to upgrade. One wonders if actually, having perceived they “must,” would a developer first consider the latest variants too resource and storage hungry? Could these concerns be a major shot in the foot for those pushing these agendas? A recent embedded survey suggested only 15 percent of developers intend to use a Microsoft OS for their next project. This must be addressed.

Whilst writing, I received a newsletter from another major electronics distributor entitled “XP is about to expire: Discover the alternatives,” pushing unsurprisingly, Windows 8 Embedded. At a risk of being branded a Luddite (though I would counter this with an in-depth understanding of what our industry needs), I would suggest that a very small percentage of embedded applications would utilize any of the new features offered by these latest versions, whilst all having to embrace higher hardware and licensing costs in addition to increased power consumption, heat dissipation, and storage needs. All this reduces market competitiveness – purely because one was pushed to jump on the new technology bandwagon. How is this good for our industry?