We’re about to put out a BJSS white paper on DevOps - hopefully sometime in the next week, barring delays getting it into appropriately pretty PDF format. During the review process, one of the people reading asked about a new term - “NoOps” - they’d recently heard.
I tend to agree with the latter. For me, DevOps is about closing the cultural gap between Development and Operations and promoting a common understanding of and respect for the complex job each does.
The original NetFlix article (and the term NoOps itself) stereotypes Ops as universally unhelpful box-rackers, helpdesk jockeys and tape-changers to be eliminated – apparently mostly based on experience of an environment with a dysfunctional Ops team.
For me, this is a neat illustration of the problem rather than the solution. It would be equally misguided to push a mirror argument of “NoDevs”, suggesting Ops people who can code should oust any dev who’s never been on a pager, on the grounds that they all write awful code that falls over the whole time. The fact that many developers do write awful code does not mean this is universally applicable.
Very good Ops people (like very good people in any area of IT), are rare but do exist. When you find them they often know everything from the business process down to the rusty metal on the disk, and can fix at every layer. These people usually start from a very similar background to good developers/BAs/TAs/technical project managers, but prefer to apply those skills in a different way and develop their career within a different specialism. Project-orientated people tend to get their kicks out of delivering something brand new and transformational. Ops-orientated people get theirs from delivering fantastic ongoing service, keeping entropy at bay, and working to perfect the imperfect. Both are equally valuable.
Because I missed it the first time, it may be worth pointing out that in the comments on the original NetFlix article, Adrian seems to have (very graciously) come to the same conclusion himself.
Of course, there are a bunch of BOFHs out there too. I suspect those on the hardware side will have to adapt or die over the next ten years as cloud infrastructure really kicks in. However, while the Ops industry is continually changing - no-one has to pick out insects from the relays in modern datacentres - I have no doubt that people will continue to describe themselves on a spectrum from “devs” to “ops” for a long time to come.
The main thing is for everyone to get on and work together for the benefit of the business, rather than using negative and offensive labels like “NoOps”. That’s why I don’t mention the term in the white paper.