| Thanks for engaging honestly with that forward comment. Yes, I have. No, I have not sold to engineers. To add some perspective: Airbrake/Sentry error monitoring; NewRelic monitoring; Pingdom uptime monitoring; Log management Ala Papertrail; Github/Assembla version control; are things that I reach for in the initial set up. I think engineers are/(should be) unwilling to give up on control when it comes to their core business areas. This is why I prefer open source (or at the very least known stable and long lasting) solutions for the data layer and core infrastructure pieces. Or at least compatible managed services; no introducing new layers of abstractions. Also they’re unwilling to pay for things like editors, from ideology and from the market today. I’ve seen the build it ourselves approach, primarily put forward by engineers who do not appreciate the costs of building and maintaining things, and from stakeholders with a vested interest in selling particular services. I would call both a circle jerk kind of situation. To reiterate, in my mind, great engineers would quickly outsource all non core competencies so that they can focus on the more important things. It costs me more to build something, both in terms of time and opportunity, than to have an expert do it for me. Herein lies understanding the problem that you’re trying to solve. To me, intelligent engineering is, cliched as it is, about understanding the tradeoffs, and building the bare minimum that delivers. Keeping in mind that the tradeoff could be about investing in future directions. One example that comes to mind was a data pipeline built using ifttt, dropbox, s3, and spreadsheets for data analysis: brutally simple, probably fragile, but brilliantly done for what was needed in that situation. Engineering is not writing code, engineering is solving problems. From a different angle, from a limited vantage point, I’ve seen companies that sell to engineers struggle because they fail to clearly articulate the value that they bring, and because they introduce too much risk to the rest of the system (What if the vendor disappears, or if I want to migrate to another? Why is there yet another API to learn that also strongly couples with my system?) I am curious about where you are coming from on this. Would you care to get into some specifics? |