InRhythm

Your partners in accelerated digital transformation.

Software Engineering

Iterate Your Way to Success

“A learning experience is one of those things that says, ‘You know that thing you just did? Don’t do that.’” — Douglas Adams, The Salmon of Doubt

Here’s a thinking tool: when you’re having difficulty finding a solution, guess. Then you have something to work with.

In other words, when you don’t know what to do next, oversimplify and self-monitor. Iterative development strategies like The Agile Method are just versions of this idea. It’s also fundamental to many machine learning (or AI) strategies.

Daniel Dennett once gave a nice breakdown of how evolution works:

How To Create a Chicken:

1. Take a dinosaur
2. Let it replicate
3. Select for flightworthiness
4. Repeat steps 1-2-3
5. Repeat step 4
6. …until you get a chicken

The thing about evolution is that nobody really knows what will be selected for, or how future conditions will impact individual survival, or just how all of that will produce chickens or ocelots or the particular way an owl turns its head. There is no intelligence to it — it’s a thoughtless process. The lesson is to keep moving, to keep iterating. Progress is always better than stagnation.

The great news is that you are an intelligent designer. You can observe and change reality at will; you need not wait on randomness to discover working designs—but don’t be afraid to guess. The mistake is in assuming that you’re designing the perfect final state of a solution (complex) instead of the first working state (simple).

Try to promote continuous improvement through ongoing positive change, keeping on an open mind about unconventional solutions. Foster a culture of accountability, have a clear purpose, and keep checking if you’re any closer to your goal. Trust that useful properties will emerge out of creative interactions, in unpredictable ways. Be open to surprises:

1. Intelligently analyze complex systems to find weaknesses and inefficiencies.
2. Create a range of solutions and intelligently determine both the right solution and the right success criteria.
3. Model and test the solution in collaboration with all stakeholders to validate the solution
4. Analyze the results, measure against success criteria, and restart the process on failure.
5. Standardize success into repeatable guidelines and share that knowledge with all stakeholders
6. Repeat 1-5, continuously.
7. …until you get a chicken (Okay, maybe not a chicken. But you get the point.)

At InRhythm, we work tirelessly to ensure that our engineers are prepared to succeed in any engineering culture. Our training program, InRhythm University (IRU), is designed to ensure that every graduating engineer has mastered the technologies and techniques that we know work at the highest levels of the industry through continuous repetition of the above process. As we continue to train and teach new engineers through IRU, we will also continue to pass along what we learn. Keep learning and keep growing.

 

WebAssembly Will Change the Way You Code

WebAssembly (WASM) is a new type of code that runs in browsers. Previously, developers of browser-based web applications had only one choice of programming language and/or environment: JavaScript. This was good and much gold did the people earn; nevertheless, some weren’t happy that they had only one choice. WASM is here to change that. WASM, according to Mozilla, “is not primarily intended to be written by hand, rather it is designed to be an effective compilation target for low-level source languages like C, C++, Rust, etc.” This is the key point. You can now write “web scripts” using the language of your choice. That’s huge.

Is WASM trying to kill JavaScript? No, says the inventor of JavaScript, Brendan Eich. The goal of WASM is to enable a more linguistically-diverse development ecosystem delivering extremely fast (native or near-native speed) code through efficient compilation and parallelization (the ability for web software to take advantage of multiple processors). More huge.

This last point is maybe the most important. InRhythm engineer Will Bratches pointed out in a recent lightning talk that increases in speed are no longer achieved at the chip level—Moore’s Law is no more. What we do now is spread work across cores, leveraging parallelization and managing concurrency. Client-side JavaScript is not designed for that world. New languages like Go and Rust are. WASM unifies the two. 

Here comes the “hello world!” moment. Head over to WebAssembly Explorer—it’s sort of a CodePen for WASM. On the left panel, paste this code:

int plusOne(int num) {
return num + 1;
}

Next, click “compile”. You’ll see something like this:

WASM WebAssembly Screenshot

Yes, that’s assembly code on the right (I know!). A more human, readable version is in the center column, the WAT view, which you can read more about here.  Just scan through it—you’ll see our function name plusOne represented as $_27plusOnei. You’ll also see that we’re establishing 32-bit integer parameters, setting their scope to “local” and so forth.  You can, for example, see how the assembly code  (add)s one to the (ex)tended (a)ccumulator register for 32-bit numbers:

add eax, 1

…is represented in the WAT:

(i32.const 1)

You can now hit the “download” button, which will deliver a .wasm file.  That’s really it—you have the WASM binary to use. There’s a tiny bit of work you have to do to compile the WASM so it works in your browser. The easiest thing to do is clone this simple example and load the index.html in your browser. Check the console. Now read through the code.

There’s a lot more to this subject, so start exploring. You’ll be creating first-person shooters in your browser in no time!

GraphQL: The API Query Language


Overview

GraphQL is an API query language used to obtain data from a server. It serves a similar purpose to a RESTful API, but with some distinct advantages.

The first advantage is that the you will only receive the data that you request. Part of the payload of a GraphQL request is the list of fields to return. This can help reduce load time for a request by only including the data that the consumer needs.

Another advantage of GraphQL is that you can get all the data you need in a single request. By defining the data you want from related fields, you can get all that data you need from those in one API call.

You can also very simply update the API without versioning. Adding new data to your endpoint will not cause any issues as the client will still only receive the data requested.

You can also use GraphiQL to test and document your API.

Example

Creating and consuming a GraphQL API is very simple. Let’s briefly go through setting up the server and then creating a simple client. In this article, we will be using Apollo and Node.js for the server, but GraphQL libraries are available in many languages.

First we need to define our data. We do this in GraphQL by defining type`s. Let’s consider a simple contact model:

type Contact {
name: String!
email: String!
}

This tells GraphQL that we have a type that is Contact that will include two strings, name and email. The next thing we need to do is define our queries:

type Query {
contacts: [Contact]
contact(name:String!): Contact
}

In this case we have two different queries. The first will return a list of Contacts and the other will return one Contact identified by the required name parameter.

Third we can define any Mutations that will allow for any changing of the data:

type Mutation {
addContact(name:String!, email:String!): Contact
}

The last step here is letting GraphQL know how to resolve these queries and mutations. To do this we will define resolvers:

var contacts = [
{name: "Test User", email: "test@user.com"},
{name: "Other User", email: "other@user.com"}
];
const resolvers = {
Query: {
contacts: () => contacts,
contact: (_, { name }) => contacts.find((c) => c.name === name)
},
Mutation: {
addContact: (_, { name, email }) => {
newContact = {name: name, email: email};
contacts.push(newContact);
return newContact;
}
}
};

Now all that is left is to hook it up. This depends on your exact setup, so I won’t go into detail. here.

Client

Creating a client to consume your new GraphQL endpoint is very simple.

Let’s create some GraphQL queries:

{
contacts{
name
email
}
}
query Contact($name:String!){
contact(name: $name){
name
email
}
}

Mutations will look something like this:

mutation addContact($name:String!, $email:String!) {
addContact(name: $name, email: $email) {
name
email
}
}

You can find a simple version of this application here: https://github.com/thomascking/graphql

More information: https://graphql.org/ https://www.apollographql.com/ https://www.robinwieruch.de/graphql-apollo-server-tutorial/

 

ES6 IIFEs with fat-arrow functions

Writing Immediately Invoked Function Expressions or IIFEs in ES6(Also known as ES2015) just got a lot easier with the introduction of fat arrow functions.

(global => {
  const MY_CONSTANT = 'api key or something'
  let counter = 0
  let some_array = [1,2,34,5,6,7]

  counter = some_array.reduce (total, item) => total + item
})(window)

This is great. Much shorter than the old ES5 way of writing IIFEs. Or if you don’t need to reference the window/global object and you just hate extra characters:

()=> {
  // code goes here...
}()

I would be hesitant to use this second approach in some cases because it’s not immediately clear that the function is going to be invoked immediately. The containing parens(while not super pretty) let you know from line one that this is a function that will be invoked right away. Also be careful using this in fat arrow functions in general. They have a slightly different context than normal functions which can bite you.

As a bit of extra credit there’s an even shorter way to write IIFEs in ES6 which is to use the block context syntax (not super recommended. there’s a reason you probably haven’t ever seen this before) all on its own like so:

// some code...
{
  // whatever happens here is private and invoked immediately
  console.log("I'm executing code immediately!")
}
// more code...

This approach is very fun because it requires the absolute minimum amount of code. It doesn’t allow us to take/pass arguments though so it’s a bit limited in terms of the features we usually use in IIFEs but that shouldn’t stop you if all you want to do is create a new closure to hide private variables from outside code.

This post originally appeared on Jack of Spades. Check out his more recent writing and follow him on twitter!

Where my girls at?

This post is another by InRhythm’s own Sari Morninghawk. For the full post and additional links, check out the original Where My Girls At?” on her website.
Women – Major Contributors to Computer Science

Many computer science pioneers were women. The first computer programmer and the people who programmed the first digital computers were all women. In the early days of computers, programming was considered “women’s work” while building and architecting the machines was mostly left to the men.

Continue reading on Sari’s blog…