HNHacker News
TopNewBestAskShowJobs

mjfisher

620 karma · joined July 27, 2010

submissionscomments
mjfisher··on Managed Kubernetes Price Comparison
I think React would be a pretty good example of the kind of thing I'm talking about, actually. Just to be clear: I like React and I'm building a product with it right now of my own free choice. It does what it was built to do very well.

The point I'm trying to make is that people often think in absolute terms. The problem isn't thinking "React is good for SPAs" - it's thinking "React is good for all websites". I've come across a number of engineers now who genuinely think React is a good fit for e.g. totally static sites, "because it's fast".

I've come across similar persistent beliefs around NoSQL databases. Some teams wouldn't begin to consider an RDBMS for their primary data store, even while their data is heavily relational - because "SQL databases don't scale".

It's not that technologies are good or bad - it's the blind belief that a given technology is best for all situations. The hype drives the development. Not of the underlying tech itself, but in the teams using that tech.

mjfisher··on Managed Kubernetes Price Comparison
Yes, it is absolutely more limited. That, single-IP load balancers, and no direct equivalent of VPCs spring to mind as the biggest differences. AWS still makes a lot of sense in a lot of cases. It is worth noting DO has a decent API, so if wouldn't be _that_ hard to implement autoscaling yourself if you needed it.
mjfisher··on Managed Kubernetes Price Comparison
> Don't you think that a lot of wizbang software developers just want to use latest and greatest buzzword thing

Yes, absolutely, and it drives me up the wall. I've seen some incredibly unsuitable technology choices made, the complexity of which have absolutely sunk productivity on projects.

That said, it's easy to become cynical and associate anything that's become recently popular with hype-driven garbage. But that can blind you to some really great new stuff too. I tend to hang behind the very early adopters and wait to see how useful new tech becomes in the wild - the "a year with X" style blog posts tend to be really informative.

mjfisher··on Managed Kubernetes Price Comparison
That's exactly how I manage it too. It means there only needs to be one load balancer per cluster, and adding a new SSL cert is just a matter of adding a couple of lines to the ingress config.
mjfisher··on Managed Kubernetes Price Comparison
> Kubernetes feels like an insane complexity

There's two things there - the complexity of engineering the cluster itself, and the complexity of using it.

The former is where the pain is. If you can remove that pain by using a reliable managed offering, it changes the perspective a bit.

The complexity of using it is also non-trivial - you have to learn it's terminology and primitives, understand the deployment and scheduling model, know how volumes are provisioned, aggregate logs, etc.

However, the ROI for learning that complexity can be pretty good. If you get comfortable with it (maybe a month or so of learning and hacking?) you get sane rolling deployment processes by default, process management and auto healing, secrets and config management, scheduled tasks, health checks, and much, much easier infrastructure as code. Which means if things do go really sideways, it's usually not that hard just to stand up a replacement cluster from scratch. With a bit more reading and some additional open source services, you get fully automatic SSL management with let's encrypt too.

All that said, I absolutely agree with you in principle. No one should overcomplicate their deployments for no benefit. It's just worth reflecting on what those benefits are. Kubernetes had a bunch of them - and whether they're worth the effort of getting to know it depends very much on the system(s) you're running.

mjfisher··on Managed Kubernetes Price Comparison
Digital Ocean is still significantly cheaper (unsurprisingly). They don't charge for the control plane, so you just pay the normal prices for the droplets and resources you use. It's well integrated, allowing K8 to provision load balancers and volumes, and the Terraform provider for it works well.

My (admittedly small) cluster of 3x 4Gb droplets, an external load balancer, and volumes enough for logs, databases and filesystems costs about 70 USD/Month. It's been absolutely rock solid too. I have very few minor gripes and a lot of positive things to say about it.

mjfisher··on Questioning the Lambda Architecture (2014)
I believe the Lambda Architecture is a simplified refinement of a previous idea, the Kappa Architecture. It's just one letter further along. I have no idea why a Kappa Architecture was named as such.
mjfisher··on Thinking Journeys rather than Estimates
I wish as an industry we practiced quantitative risk management more. I run a small training workshop about producing better estimates, and the single idea that I want people to take away is that a single-valued estimate is a really terrible shorthand for describing a probability distribution.

If you describe estimates as ranges or distributions, a whole world of risk analysis tools open up - PERT, Monte Carlo, etc. More crucially, it allows you to start communicating and managing risk openly with everyone involved, instead of pretending it doesn't exist.

mjfisher··on Quarter of all tweets about climate change produced by bots
That's not the same thing at all. General relativity makes incredibly precise predictions about very specific phenomena. There have been hundreds of separate natural observational experiments (transit of mercury to gravitational lensing) and plenty of human created ones too (every GPS satellite in orbit).

Global climate science has it even harder than economics as an experimental science - we've got exactly one fully integrated system to observe, and we can't control for any of the variables. We can't re-run the same experiment again. We need to be objective and honest in how we communicate about climate science. Drawing false equivalences just undermines the very real and very urgent discussions we need to be having about climate as a species.

mjfisher··on Mini projects built with VanillaJS. No frameworks or libraries
> Sadly, I didn't get a chance to provide feedback to the criticisms

That's a shame. It sounds like the criteria were fairly well defined, but it would be nice if your assignment was used as a starting point for good in person discussion. E.g. "we think the code in index.js was too coupled - what could we do to break that out?"

mjfisher··on Mini projects built with VanillaJS. No frameworks or libraries
That's a nice, clean, simple approach - I learned about `createContextualFragment()`, so thanks!

I wanted to ask, did the company you were interviewing for tell you what they wanted to get out of the assignment on their end - specifically what they were assessing you on?

mjfisher··on Linode launches free DDoS protection
Two things that made me migrate (mostly) to DO from Linode:

1. For a long time, Linode did not have a terrform provider 2. DO's managed Kubernetes offering Just Works, and is very competitively priced

FWIW, I still run a small Linode box. It's been rock-solid, and the support they provide is absolutely top-notch.

mjfisher··on We ran the numbers, and there really is a pipeline problem in engineering hiring
I don't know if this will be of any use to you. If you're pitching to VCs, you're already operating way above my level. I also don't know if it's different over here. I live and work in the UK and I'm white. I genuinely don't think I've worked with anyone in the tech industry who cares what race or background their co-workers have.

There's definitely an underrepresentation problem (most grievously with female engineers), but I sincerely believe the problem is mostly systemic, rather than driven by personal prejudice.

From my (limited) perspective, the only thing anyone gives a damn about is whether you're a good developer. If you can code, if you can ship, if you can work well with your other colleagues. Please don't feel like you have you walk into meetings pretending you're somebody else. Walk into meetings with the confidence of a great engineer and founder - because anybody who gives a crap about anything else isn't worth impressing.

With love, from one engineer to another ️

mjfisher··on There Is No Cure for Burnout
I think there's two kinds of situation this occurs in. There's the "not yet learnt" kind where people don't know what they can deliver, over-promise, and end up in an awkward situation.

There's also a trickier, structural situation that I think is more common. If managers don't have real ownership of their client relationships, they learn they become effectively unable to say "no". If they try, someone else will say "yes" instead.

mjfisher··on How to Be an Epicurean
That looks absolutely perfect for me.

I've found some Stoic attitudes and practices a great help over the past year or so, but suspect there's some really valuable knowledge from the other schools I've yet to learn about.

I'm working my way through Seneca's letters at the moment. He keeps quoting Epicurus, so that's a pretty decent recommendation right there! Anything that builds out context around that era of philosophy is massively interesting to me.

Thank you very much :)

mjfisher··on How to Be an Epicurean
A nicely entertaining read. Does anyone else find the author's characterisation of Stoicism a little reductive? My reading of the Stoics has been limited so far, but at least as far as Marcus Aurelius writes one of the primary Stoic paths to peace and virtue is through helping others.
mjfisher··on How to sleep at night having a cloud service: common architecture do's
Storing secrets in env vars is very common in practice, although it presents a slightly bigger attack surface than using something like Hashicorp's Vault to just pull the secrets into memory.

You can sometimes find debug pages etc for apps and runtimes set up that will show all set environment variables, or have crash monitoring software that will capture env vars and send them elsewhere by default. Those risks can be managed, but having sensitive information not set in the process environment is more 'secure by default'. It also means in the event that someone finds a way to remotely execute code in your process (eval() on an unsantized input, anyone?) it's much harder to dump out secrets.

mjfisher··on GDPR fines were meant to rock the data privacy world
I'm curious about the economics of this - is it big enough to not be worth redeveloping when you consider over the income over say, 3-5 years?
mjfisher··on The pioneers of cultural anthropology who created “cultural relativism”
What a great example of the power of PCA! I absolutely love that you can give it nothing but a set of raw dimensional data, and it will give you back a set of axes that are (often) human interpretable.

I remember the first time I did that for a set of children's ratings data, and got to see games/movies/books spontaneously spread themselves out along fairly clear age and gender lines. I often wonder what other novel applications it can have.

mjfisher··on .NET Core 3.0 Concludes the .NET Framework API Porting Project
Can anyone closer to the .NET community than I am comment on the wider adoption of .NET core on standard line of business applications?

The few places I know that work with .NET are a long way from migrating yet. Are there any similarities with the Python 2/3 port? I imagine language level compatibility makes the transition much easier.

mjfisher··on How to write fast code in Ruby on Rails
Speaking as someone who uses Rails regularly, I agree - it is considerably slower than many other language choices. It probably isn't a good call for a performance critical bottlenecked service that can't scale horizontally.

For everything else, it's great, and it's almost always a good default to choose speed of development and engineer productivity over runtime performance.

mjfisher··on Show HN: Candidate Code – take home coding test management
I see this position fairly frequently. While I respect the place it comes from, I've never quite understood the unequivocal nature of it. Is it all take home tests you refuse?

Does your position change at all depending on the length of the take home, or how much effort has been put into designing the test? Or the trust you have in the company?

Personally, I'd rather do a short and focused take-home if it means I can save time on a longer in-person interview setting. It can be a really efficient way of demonstrating a set of skills, which leaves in-person interviews free to discuss higher impact topics like architecture, decision making and soft skills.

mjfisher··on Show HN: Candidate Code – take home coding test management
Also, because take home tests can be a controversial subject: I think they're a great way to evaluate some skills, but they're not the only way. There's pros and cons to every technique, for both candidates and the hiring organization. I'm personally also a big fan of pairing style interview questions (remote or otherwise). A lot of the bad experiences people have with take home coding challenges seem to be caused by poor design as much as anything else.

Mostly, I think if you design your hiring process to respect your candidates it's difficult to go too far wrong.

mjfisher··on Show HN: Candidate Code – take home coding test management
I've found well-designed take home coding tests to be a great way of assessing potential hires during a recruitment process. I've also found that take home tests can be really hard to manage for a growing team; lost emails from HR are common, incomplete submissions are returned ad-hoc via zipped attachments and google drive uploads, and reviews and feedback from your own engineers are often difficult to keep track of.

I created Candidate Code to make it trivially simple to issue your existing take home coding tests through Git, collect the results and review them collaboratively. Feedback is welcome.

mjfisher··on Dog-walking startup Wag raised $300M, then things got messy
I'm not sure that's the case with Javascript - there's plenty of microframeworks (express, hapi) and even a couple with bigger ambitions - like Sails.js

Last time I looked, none were comparable to Rails/Django/SpringBoot/Laravel

mjfisher··on How Firebase Interviewed Software Engineers
I think that's a great approach. I agree absolutely about not being able to fake it either; in practice I've never caught anyone putting forth someone else's code as their own, and I don't think it's a huge concern for the reason you mention.

Do also spare a thought for the people who don't have side projects, or any code that they can legitimately share. There are plenty of talented engineers in that position, and it'd be a shame to have a hiring process that leaves them out in the cold.

mjfisher··on How Firebase Interviewed Software Engineers
I get the sentiment - but at some point, a hiring company absolutely has to test your ability to make a judgement about whether to hire you. Speaking from experience on both sides of the table, I'd rather that be through writing actual running code than through some awkward or irrelevant whiteboarding exercise.
mjfisher··on How Firebase Interviewed Software Engineers
I have a strong dislike for the 'see how they think' or 'code like you would in production' style instructions for take-homes - they're not specific enough.

I don't think that necessarily applies to real-time pairing exercises though. The interviewer has significant latitude to interactively question and prompt to get at the skills and values a candidate possesses.

mjfisher··on How Firebase Interviewed Software Engineers
The issue there is the problem starts to become unbounded for the candidate; they can't tell what you're looking for, so they don't know how to prioritize their time.

That's a problem even if you have reasonable expectations yourself, because the candidate probably has experienced some tests where the expectations haven't been reasonable. For a (real) example: marking candidates down for not having an analytics integration.

How does your candidate know that's not important to you, even though it was important to the last company?

> testing your experience, not your literal skill in pumping out code

Past a certain point, it gets much more efficient to discuss experience in an interview setting. Good coding tests should be designed to test the skills you care about most to produce a strong hiring signal for the next phase - it's difficult to showcase a full decade's worth of experience in 4-8 hours' worth of coding challenge.

> Anyone can pump out code

I've found that to be depressingly untrue amongst a significant percentage of engineering applicants.

mjfisher··on How Firebase Interviewed Software Engineers
It's probably worth mentioning that any take home worth its salt will explicitly list what criteria you're being assessed on.

It's crazy how many coding tests are poorly designed and waste a candidate's time. I'm putting together a list of good practice for test design here: https://candidatecode.com/articles/coding_challenge_best_pra...

Feedback and contributions welcome

← PreviousPage 5 of 7Next →