DOS isn’t dead

Like the resident psychopath in an 80’s horror franchise, DOS, despite many claims to the contrary over the years – simply refuses to die.

Describing DOS as having “gone away” after the release of Windows 95, even the mighty Microsoft couldn’t have predicted how important its later raison d’être, far away from the glitz of multimedia applications, at the very heart of embedded systems, was to become.

An exponential enhancement of eye-candy ensued, increased detail, colour depth and frame rates freed the desktop developer to incrementally deliver 3D gaming and HD video as the world looked on in awe – but what of the embedded developer?

The embedded developer looked on, ever twitchy – as he saw his control and monitoring needs change little. The ‘Hare’ of the desktop computing world began pressuring him to migrate to ever ostentatious Operating Systems (with ever ascending hardware requirements) that offered little benefit to his world of 24/7 reliability and many risks.

Mocked as the ‘Tortoise’, scaremongering of false DOS obsolescence increased this pressure, he drowned in confusion and desperation of how, his often headless, applications must unnecessarily now piggy back onto a resource hungry GUI based Operating System.

Feeling ever the Luddite, it then occurred to our friend many new hurdles he must overcome. How could his product remain profitable after absorbing the multiplying cost this superfluous, devourer of hardware performance, would demand? How exactly can his compact product dissipate the now seemingly enormous TDP as a result of vastly increased wattage…and hold on – exactly where is this additional power magically appearing from!?

As if this wasn’t headache enough, today’s embedded operating systems do not work out of the box and require building, the level of effort required to reach a working build very much inversely proportional to the end user license cost the product’s market price must absorb.

Concluding the least effort required resulted in a license cost that’s double the cost of previous his entire system, he naturally sways towards the purported ‘license free’ route – “This sounds more like it!” he beams. Though his enthusiasm soon wanes as he realises the mammoth development task ahead of him, made worse as the realisation that via this route, an easy scalable hardware configuration is a long forgotten myth and he now needs to produce multiple builds to cover the slightest variation in his product.

He bathes in a momentary daydream of the good old days, where he could instruct devices directly, use the full capabilities of his hardware without having to budget for OS resource overhead and not have to worry about bewilderingly complex license agreements.

He longs to return to the times he could bet his life on his product’s reliability for decades forward, not lay awake at night anxiously dreading a humiliating public Blue Screen of Death (who can forget the Beijing Olympics 2008!), or if a headless system, complete non-functionality with even less explanation to an important customer – he can’t decide which is worse!

His hand is forced now though, his core customer base are now demanding such ‘extravagancies’ as 100Mbit LAN and USB Support, or is it?

Surprisingly, few software developers appear au fait with the advancements made with DOS since (as a desktop user) they said their goodbyes decades ago. Both USB and high speed networking are now as standard in your embedded DOS environment, alongside other new features and a thriving developer community, which DSL are heavily involved in.

Ability to operate entirely from ROM isn’t news, but it’s this factor that ranges from a comfort factor to a necessity in approval stringent industries (i.e. the Military). Today’s DOS can also operate in a tiny footprint (the majority of DSL’s modules have FreeDOS pre-installed on low cost, high reliability on-board SPI flash) eliminating the need for separate, expensive, industrial flash memory.