On Jan 1, 2015, at 9:52 AM, Steve Clark <sclark at netwolves.com> wrote: > On 12/29/2014 09:04 PM, Warren Young wrote: >> >> The vast majority of software developed is in-house stuff, where the developers and the users *can* enter into an agile delivery cycle. > Where did you get the 5% from An industry publication, probably 10-15 years ago. (You know, back when they printed these things on paper.) I haven’t bothered looking into it again since. If you want a fuzzy but more current picture of the state of things, go to a technology job board, and narrow the search to an area of the country that isn’t Silicon Valley, New York City, Boston, or Austin. (i.e. Not one of the hotbeds of commercial packaged software development.) Ideally, somewhere you don’t live now, never have lived in, never have visited, and have no desire to visit/live in. Anchorage, Buffalo, Cleveland... Now count how many jobs are for positions on a team creating packaged commercial software, then make a ratio of that vs the total job count. Then discount for the fact that internal software development has a lot less turnover than all these flash-in-the-pan App Store darlings. Don’t mistake a job for a NASDAQ Top 100 company for one developing commercial packaged software. Apple, IBM, and Microsoft need internal software, too. I think you’ll find a lot of companies you’ve never heard of. Obscurity doesn’t remove these companies’ need for custom software. One of the biggest movements in software development in recent years (Agile/XP) got started as a result of work done for *Chrysler.* > according to google there are > > "over 200 billion lines of existing COBOL code, much of it running mission-critical 24/7 applications, it is simply too costly (in the short run) for many organizations to convert." > > And what about Fortran, RPG etc. Yeah, that’s all going to be in-house stuff. The fact that this software is irreplaceable doesn’t affect my point at all. If this software is neither perfect nor dying, there is an in-house staff working on its maintenance and enhancement. The developers could instead be consultants, but that’s a question of outsourcing. There is still a tight coupling between user requests and what gets developed. My point is that that loop is much larger and looser with packaged commercial software, to the point that user requests are nearly disconnected from visible effects. I don’t know about you, but my success rate at convincing commercial packaged software producers to add my pet features and fixes is probably under 5%, and it often takes 12-24 months to get a change into a released version. At least with open source, I have the *option* of developing the fix/feature myself. With in-house software, you should be getting nearly 100% of your requests implemented, within a reasonable schedule given the staffing provided. (If you put 1,000 “top priority” features on a wish list and give it to a single developer, it will necessarily take a long time to get a feature through the pipeline.) > Sounds like you have no shortage of help Ahaha. No.