All I Want For Christmas … Is A Census Of Software Developers!

By Frank SchirrmeisterThe year is almost over and in my homeland of Germany, kids are preparing to appease Santa Claus and Knecht Ruprecht (the guy who gives you a rod if you misbehaved during the year) with a Christmas rhyme. I remember that as a kid, I was at one point reciting from memory the second chapter from the book of Luke: “In those days Caesar Augustus issued a decree that a census should be taken of the entire Roman world. This was the first census that took place while^[a] Quirinius was governor of Syria. And everyone went to their own town to register.“ I remember I made it all the way to verse 20.A census? Great idea! My Christmas wish this year is a census of software developers! Oddly enough I thought of it during a recent talk given by EDA investor Jim Hogan. I wrote about it in my Blog post “Equine Anatomy, Pax Romana and the Reach of Standards.” Specifically I published a photo of a graph Jim used, a set of stacked triangles indicating the team-size of hardware and software developers. The upper triangle in that graph is wider, indicating a larger team size of software developers. It would be nice to draw this graph with some real project data. Hence my desire for a census …

Often when we talk about software developers and how EDA companies can sell to them, we cite that there are a “lot more of software developers than hardware developers,” often followed by “and their price expectations are much lower.” Proof would be nice and a census would certainly help. So do we have the proof perhaps already? Isn’t that what market research is for? Well, in a recent article I published called “Combining Prototyping Solutions to Solve Hardware/Software Integration Challenges”, I referred to research previously done by IBS and I drew a project flow averaged from a survey of 12 projects. It turns out that using the same data, a derivative of Jim’s stacked triangle graph would look like this:

The yellow portions on the right side show the relative effort of software development tasks measured in man weeks. The blue portion in the bottom shows the same for hardware development tasks. On the right side I drew the same graph, this time using the relative elapsed time per task measured in weeks. This is an important and different view because for a project not only the effort counts, but certain tasks like the specification need to be done before one can proceed. In both cases the graphs roughly resemble the Hogan triangles mentioned earlier. With the difference, though, that the software triangle is not that much wider. Shoot, as my almost seven year old daughter would say! Are there really not that many more software developers?

But wait, there is more!  I found more recent data from 2010, also researched by IBS, which surveyed the actual team sizes of hardware and software related tasks across 12 projects in the mobile media application domain, 4 each in 65nm, 45nm and 28nm. If I redraw the Hogan triangles again, I get the following as average:

Again, the measure results show graphs resembling the Hogan triangles, but judging from the width of the upper triangle, the number of software developers is not that much higher than that of hardware developers.

There is good news in all four graphs as well: The grey portion shows the area which we sometimes refer to as “hardware aware” software development, i.e. development which requires somewhat detailed knowledge of the hardware. That number actually represents a very fair share of the chip related software development and takes up significant time of project schedules in terms of the elapsed number of weeks. All of this is good news for technologies like Virtual Prototyping, Emulation and FPGA based prototyping, all of which appeal to this user type.

So this exercise leaves me somewhat relieved on the one hand that there are a fair amount of embedded software developers who can use our technologies, but on the other hand I am also wondering where the large number of software developers are hiding. For sure, as the data comes largely from semiconductor company surveys, the apps developers for IOS and Android may be missing here. But still, it would be nice to get better quantitative numbers. Which brings me back to my Christmas wish. Let me see whether “‘Tis the Season” and I can get some new data next year.

Happy Holidays!

–Frank Schirrmeister is group director for product marketing of the system development suite at Cadence.