Here are a few items, all fairly disparate, that caught my attention yesterday that I want to call out up top.
Docker buried the lede on Project Moby:
If you read Docker’s blog post on this from Tuesday
, the most important sentence might be the penultimate one: “All our open source collaboration will move to the Moby project.” Basically, Moby will be the home of the various components that comprise Docker’s Community Edition, such as containerd, LinuxKit, etc., and ideally they will be detached enough from Docker’s, (the company’s) products that anybody else can use them to build something competitive to Docker’s products.
This is actually kind of a big deal in terms of branding, commercial development and community engagement, and Docker really did not communicate it clearly to begin with. For more clarification from Docker creator Solomon Hykes, and to get a sense of community confusion, check out this this discussion on the Moby GitHub repo
Amazon CodeStar has sysadmins scared again: Amazon CodeStar
is a new tool, announced yesterday, to help users easily spin up AWS resources for their application, and then transition them into the world of agile development. I viewed it as kind of an incremental new service, packaging up a lot of what AWS is already offering and has been pushing (at least implicitly) for years, with a much-needed shine job to help less-tech-savvy users get started.
However, the Hacker News discussion
of CodeStar centered around how it will or won’t kill lots of DevOps jobs at SMBs that are just now making the transition to the cloud. I come down on the side of the debate that says “this won’t be such a big deal for jobs” (at least as currently constituted) but it’s a good discussion to read to get a sense of how folks are reacting to the increasing automation in the already-automated cloud. FWIW, I think container orchestration will have a much bigger impact on DevOps.
Steve Ballmer’s grand, possibly ineffective, data service:
Steve Ballmer’s USAFacts
website is, by all accounts (including mine), relatively nice to use and full of lots of good data. Its biggest problem—as pointed out by authors at both Ars Technica
and The Atlantic
—is that facts and data haven’t proven remarkably good at influencing public opinion lately. For many, many reasons, data is treated differently in business than it is in politics or personal matters, and that might never change.
A few things, though: (1) This is really a people problem, not a technology problem (I think, for example, that Facebook takes an inordinate amount of flack for its users actions); (2) companies obviously want to protect their platforms, but they should consider what legal liability they open themselves up to by actively monitoring content instead of relying on safe harbor laws; and (3) the machine learning technologies on which these service are built are not foolproof; humans still must be in the loop.
For evidence on that latter point can be found here
and probably in many other places. People being people, they will actively try to fool the systems.