Software megaprojects
Mar. 1st, 2015 03:32 pmRecently
sorhed posted an insightful observation that in the software field, all relevant and big software projects are designed in US, with very few exceptions. And that it is difficult to identify country #2.
Apparently both points are true. What we were arguing about was only definition of a megaproject. Actually, in the post I am referring to there was no meaningful definition. However the topicstarter referred Ruby on Rails as an IT megaproject, and Oracle/SAP, Step 7 and IEC61508 applications as trivial but large things.
So I'll build a list of s/w related megaprojects, and I will use the following criteria when listing megaprojects:
1. Requires x000+ engineering man-years to build.
2. Enables an ecosystems of x0000+ engineers who use it in their day to day coding job.
3. Global outreach. (Ruling out 1C, Yandex)
Let's start with some history. Just 30 years ago, there were not as many programming megaprojects as now:
megaprojects were minicomputers, mainframes, telco exchanges, and supercomputers. (Sorry PCs were joke at the time, no megaprojects there until 90s). All but telco exchanges were dominated by US firms. And big telco exchanges were built almost everywhere:
DMS-100/Protel(Canada), DX-200/TNSDL(Finland), AXE/PLEX, AXD/Erlang(Sweden), EWSD/CHILL(Germany), 5ESS(US)
OK, fast forward to now. Let's start with EDA - big three: US. FPGA programming: Both US. Lytho equipment: Netherlands, US, Japan. Contract manufacturing: Taiwan, US. CPUs: US, UK(ARM, MIPS now). GPUs: US, Canada. GPOSes: US. RTOSes: mostly US. CAE/CAD: US, France, Germany. General purpose programming ecosystems: US. Cloud programming ecosystems: US. ERPs: US, Germany. Industrial systems: Germany, US, Japan. Automotive s/w: Germany, Japan, US, UK. Safety and reliability engineering: Germany, US. Airspace: US, UK, France. Rail: Germany, France, Canada. Energy: Germany, Switzeland, US.
Did I miss something important?
To summarize, US is an apparent leader, but there are candidates for #2 from the list: UK, France, Germany, Canada, Japan.
![[livejournal.com profile]](https://www.dreamwidth.org/img/external/lj-userinfo.gif)
Apparently both points are true. What we were arguing about was only definition of a megaproject. Actually, in the post I am referring to there was no meaningful definition. However the topicstarter referred Ruby on Rails as an IT megaproject, and Oracle/SAP, Step 7 and IEC61508 applications as trivial but large things.
So I'll build a list of s/w related megaprojects, and I will use the following criteria when listing megaprojects:
1. Requires x000+ engineering man-years to build.
2. Enables an ecosystems of x0000+ engineers who use it in their day to day coding job.
3. Global outreach. (Ruling out 1C, Yandex)
Let's start with some history. Just 30 years ago, there were not as many programming megaprojects as now:
megaprojects were minicomputers, mainframes, telco exchanges, and supercomputers. (Sorry PCs were joke at the time, no megaprojects there until 90s). All but telco exchanges were dominated by US firms. And big telco exchanges were built almost everywhere:
DMS-100/Protel(Canada), DX-200/TNSDL(Finland), AXE/PLEX, AXD/Erlang(Sweden), EWSD/CHILL(Germany), 5ESS(US)
OK, fast forward to now. Let's start with EDA - big three: US. FPGA programming: Both US. Lytho equipment: Netherlands, US, Japan. Contract manufacturing: Taiwan, US. CPUs: US, UK(ARM, MIPS now). GPUs: US, Canada. GPOSes: US. RTOSes: mostly US. CAE/CAD: US, France, Germany. General purpose programming ecosystems: US. Cloud programming ecosystems: US. ERPs: US, Germany. Industrial systems: Germany, US, Japan. Automotive s/w: Germany, Japan, US, UK. Safety and reliability engineering: Germany, US. Airspace: US, UK, France. Rail: Germany, France, Canada. Energy: Germany, Switzeland, US.
Did I miss something important?
To summarize, US is an apparent leader, but there are candidates for #2 from the list: UK, France, Germany, Canada, Japan.