signal

- Based on a random survey of software engineers (who somehow had visibility to project inception & outcome)

- Doesn't define project

- Success was defined by not violating the iron triangle (so nobody knew what they were saying OR they're working on trivial "projects")

- If you read reviews for the book it's totally AI generated nonsense

- Based on a random survey of software engineers (who somehow had visibility to project inception & outcome)

- Doesn't define project

- Success was defined by not violating the iron triangle (so nobody knew what they were saying OR they're working on trivial "projects")

You don’t think DIY or ‘scrape by without a solution’ represents a significant dependency?

Sunk cost (amongst other factors) is a gravitational force on orgs that try to get things done without the right tool for the job.

What’s an example of a thriving tech company with home grown UX tooling?

How are you reconciling your take against the actual content of the article?

Hi folks! This is a writeup I published earlier in the year because I hear of a lot of people curious about the '4 Key Metrics of DevOps', what to measure, and how to get started making improvements to flow and performance.

TL;DR: Starting from a clear outcome is more important than chasing someone else's metrics, and mapping with your team can help you effectively translate your goals into action.

I'd love to hear thoughts, experience, and feedback!

Did I miss the target use case and user?

I'm a big fan of decision frameworks but without those two clarifications it's hard to give feedback. To make this practically valuable it would be great to have logging and an escape path for criteria that doesn't apply or when information isn't available. If I get to question 5/6 and I don't know the answer, can I get a valuable outcome? It would be ideal to capture those cases with logging so the tree can be improved.

This is a great idea and start though. I love how minimal it is. Once I use it, I immediately want it just built into Jira or wherever I'm making important/repeated decisions.

It would be interesting to see what you're saving on each tool and why they beat free alternatives. I think the only one that doesn't have a compelling free option is getresponse but there's sendfox for email that does the job well.

I'm in this camp - I want to share content with Pocket (and increasingly Notion for personal use, then outward to Twitter/LinkedIn/Automation and all of that is already in my browser.

Is that factoring in cost of living? There's an enormous difference between Toronto and SV there.

So, DevOps as you (mis)understood it is dead. Long live DevOps as it was meant to be understood.

This is super interesting! Are the product managers and designers a shared resource?

That's close, but still a problem. If you put change responsibility formally on the shoulders of a team or individual you're throwing them under the bus. It needs to be approached as a cross-functional team initiative (which is reproduced to scale out). A DevOps team might work if they're in a leadership position, but how does it work on the ground?

DevOps doesn't mean people doing multiple jobs. You can collaborate with separated roles and be productive. Furthermore you can build abstraction to simply remove manual, unrecorded interaction with systems to satisfy auditing and regulation.

:) Yes exactly! Deployment automation to the rescue!

Thanks for checking it out! It sounds like what you're looking for is a Ops or deployment checklist (but really just deployment automation). DevOps really is about "general engineering team principles", one of them being 'DevOps is not isolated to a specific role in our organization.' aka 'Don't have a "DevOps guy"'. There's a lot of great material in the references section at the bottom if you'd like more information on the reasoning there, but I think some investment in automation would relieve the checklist issue and make your deployments quicker and more flexible as a result.

Codeship really gets points for clarity and usability here. This is the best implementation of parallelism I've seen yet.

So here's what that actually is: DevOps is super popular with recruiters and managers since the venn diagram of people who care about DevOps tend to be good team players with general yet current competency with very valuable tools and techniques, and they happily pay very well for these people. People who like being paid very well and enjoy working interesting and current problems realize this and thus DevOps is a tag that allows these two groups to find eachother. It absolutely gets misused in job titles for visibility, and we don't care because it usually doesn't mean the company wants a person who knows and does everything (though it certainly happens). It is entirely possible for DevOps to refer to a position, but by the very nature of the contraction, it would be two jobs for one person. Right now we live in a time of sufficient complexity such that in all but the smallest organizations that doesn't make sense, and in those organizations job titles are moot anyways.

In reality, if you look at the history of the term and the current leadership of the development and discussion around DevOps, it still represents the following core areas: 3rd Order Cybernetics, Agile (Infrastructure + Software Development) and Continuous Delivery. I personally believe it's just a simpler way to refer to 3rd Order Cybernetics (3OC will never catch on) and DevOps the term will give way to something that better fits the nature of the movement, but until then, that's what DevOps is. I agree there's a lot of noise and confusion, but that hasn't erased the history, the progress or the true current state.

I disagree. I think the article speaks rather well about what DevOps is really about. There's actually a great deal of consensus within the community about it's meaning, and Mr. Kim has been a major contributing member for quite some time now. Every DevOpsDays since 2009 has critiqued and reaffirmed the ideals and principles. Have a look at vimeo.com/devopsdays or the wikipedia article for DevOps.

Swapping DevOps for PostOps or NoOps or FooOps doesn't change what it means.

Recruiters and managers starved for multidisciplinary, highly skilled T-shaped staff have started to throw the term around in job descriptions, but misuse of a term doesn't change it's meaning until the majority of its users adopt the evolution.

Speaking of speakers: The official deadline has just past (we just missed this!) but I'd love to have some HNers submit for DevOps Days Toronto here: http://devopsdays.org/events/2014-toronto/propose

I'm glad to see this freshened up. Curated deals ala appsumo minus spam is a welcome concept.

I have nothing against reading Huffington Post for entertainment or opinion pieces, but it's hardly a source for anything relevant to hacking(in the HN sense) or startups. I come here to see news that doesn't make the headlines of major media outlets. I probably shouldn't have commented on the article that I didn't want to see on HN, but I felt I should contribute if I was going to criticize.

Your options really depend on your current numbers, your estimated growth and your pain threshold for scale. Starting with a single box sounds like you either have nothing right now or you're on shared hosting, does 'dedicated box' mean a physical server or a virtual one? Virtual hosting nowadays with someone like Linode or Rackspace would be a good bet, but you can find cheap(and less reputable) options on Lowendtalk.

OS shouldn't be a dealbreaker, if you're comfortable with a distro and you see it go ahead, but if price is a consideration don't let rpm vs. deb concern you too much. You can run a stackscript to get off the ground without getting your hands dirty.

Can I ask why you wouldn't use a cloud provider? Are you already off the ground in the cloud and need to change to a single box? Or do you run a wonky stack? PHPfog runs Varnish in front of their stack, it would be plenty fast and save you tons of hassle setting up and maintaining a box. I'm on their free tier and could scale up any time.

I really dislike the way this has been handled. Yes, I'm sure the FBI is legally prohibited from notifying the victims of the problem directly or even fixing it remotely, but there are many other avenues available to address the issue than needlessly partnering with a single private contractor. Even the posts addressing the issue don't provide clear instructions to verify or clean the issue, and the external sites they reference are provided by DNS name rather than IP.

I think that at the very least the G20 need to agree to the creation of some form of unified body able to address these types of issues effectively. For that to happen the governments involved would need understanding and expertise in online security, which they are all sorely lacking. It's going to be a very rocky road.

Not to be a jerk, but I don't come to HN to see headlines from the Huffington Post. In restrospect I don't think I should have commented.

I like the idea and definitely agree with the premise, but I think this could gain more momentum as online events than off. Collaborative diagrams can be drawn up on a number of sites (Google apps being one - too bad diagram.ly is single user) and a regular event would eventually catch on. I'm no architect but I'd love to sit in and add value where I can.

I've been thinking about this lately too. I think it could be beneficial to have something like the co-founder wish list (spreadsheet) where people can post easy tasks they don't have time or skills to tackle (test this app on android, take inline CSS and consolidate it, add Ajax, document this etc). Maybe a twitter hashtag too. Put it together and link to Github through their API. Issues in Github can have labels, so maybe if you submit a repo to the site and have issues labelled (easy) they get pulled in and people can pick off issues they could help with.

Ultimately, of course, this doesn't solve your problem. However, it would address the issue nicely since this question is asked repeatedly in many places around the web.

OTRS? http://otrs.org

I think it has all these features but I only ran a quick search on the site, you may want to ask them for specifics.

Open source, free, you can even get it in vm appliance form: http://www.turnkeylinux.org/otrs

We use NewRelic to see everything from the bottom up: http://newrelic.com/

It's really easy to use, comprehensive and they have a 14day trial.

I also like yslow: http://developer.yahoo.com/yslow/

It's good for quick checks.

I think this is the natural reaction of the hacker news (intelligent and savvy) crowd, but I see more and more every day that the average human desperately wants to be seen by anyone. Hence the success of youtube, twitter and facebook.

It's both. You need to personally believe it's important for it to stick, and you have to train yourself to keep it in mind. It has to become like a demanding girlfriend you have to remember to txt or call. You can't dump her because she's loaded and you work for her dad.

The right system is totally secondary. It can be as simple as a note in your front pocket. In fact, a written note is good because you have to continually rewrite it, and that makes the tasks stick in your head. It also becomes a huge hassle to keep writing down the same things instead of getting them off the list. When I have to kill my list I do this.

Day to day I use google tasks (iphone-dock and firefox-sidebar) and it stays in my face all day between those two.

A long list is good. There's nothing that guarantees failure in todo world like having tasks like: "Profit". You have to break everything down to the smallest particle you can, then you knock off more things more often and you get used to using your system.

Determination is the most valuable skill in the world, stick with whatever you do and it will pay off in every aspect of your life.

You could sell it for $1200 on Flippa.