Loading...

Follow Smashing Magazine | For Professional Web Design.. on Feedspot

Continue with Google
Continue with Facebook
or

Valid
Sending Emails Asynchronously Through AWS SES Sending Emails Asynchronously Through AWS SES Leonardo Losoviz 2018-11-13T14:30:53+01:00 2018-11-13T14:40:42+00:00

Most applications send emails to communicate with their users. Transactional emails are those triggered by the user’s interaction with the application, such as when welcoming a new user after registering in the site, giving the user a link to reset the password, or attaching an invoice after the user does a purchase. All these previous cases will typically require sending only one email to the user. In some other cases though, the application needs to send many more emails, such as when a user posts new content on the site, and all her followers (which, in a platform like Twitter, may amount to millions of users) will receive a notification. In this latter situation, not architected properly, sending emails may become a bottleneck in the application.

That is what happened in my case. I have a site that may need to send 20 emails after some user-triggered actions (such as user notifications to all her followers). Initially, it relied on sending the emails through a popular cloud-based SMTP provider (such as SendGrid, Mandrill, Mailjet and Mailgun), however the response back to the user would take seconds. Evidently, connecting to the SMTP server to send those 20 emails was slowing the process down significantly.

After inspection, I found out the sources of the problem:

  1. Synchronous connection
    The application connects to the SMTP server and waits for an acknowledgment, synchronously, before continuing the execution of the process.
  2. High latency
    While my server is located in Singapore, the SMTP provider I was using has its servers located in the US, making the roundtrip connection take considerable time.
  3. No reusability of the SMTP connection When calling the function to send an email, the function sends the email immediately, creating a new SMTP connection on that moment (it doesn’t offer to collect all emails and send them all together at the end of the request, under a single SMTP connection).

Because of #1, the time the user must wait for the response is tied to the time it takes to send the emails. Because of #2, the time to send one email is relatively high. And because of #3, the time to send 20 emails is 20 times the time it takes to send one email. While sending only one email may not make the application terribly slower, sending 20 emails certainly does, affecting the user experience.

Let’s see how we can solve this issue.

Front-end is messy and complicated these days. That's why we publish articles, printed books and webinars with useful techniques to improve your work. Even better: Smashing Membership with a growing selection of front-end & UX goodies. So you get your work done, better and faster.

Explore Smashing Membership ↬
Paying Attention To The Nature Of Transactional Emails

Before anything, we must notice that not all emails are equal in importance. We can broadly categorize emails into two groups: priority and non-priority emails. For instance, if the user forgot the password to access the account, she will expect the email with the password reset link immediately on her inbox; that is a priority email. In contrast, sending an email notifying that somebody we follow has posted new content does not need to arrive on the user’s inbox immediately; that is a non-priority email.

The solution must optimize how these two categories of emails are sent. Assuming that there will only be a few (maybe 1 or 2) priority emails to be sent during the process, and the bulk of the emails will be non-priority ones, then we design the solution as follows:

  • Priority emails can simply avoid the high latency issue by using an SMTP provider located in the same region where the application is deployed. In addition to good research, this involves integrating our application with the provider’s API.
  • Non-priority emails can be sent asynchronously, and in batches where many emails are sent together. Implemented at the application level, it requires an appropriate technology stack.

Let’s define the technology stack to send emails asynchronously next.

Defining The Technology Stack

Note: I have decided to base my stack on AWS services because my website is already hosted on AWS EC2. Otherwise, I would have an overhead from moving data among several companies’ networks. However, we can implement our soluting using other cloud service providers too.

My first approach was to set-up a queue. Through a queue, I could have the application not send the emails anymore, but instead publish a message with the email content and metadata in a queue, and then have another process pick up the messages from the queue and send the emails.

However, when checking the queue service from AWS, called SQS, I decided that it was not an appropriate solution, because:

  • It is rather complex to set-up;
  • A standard queue message can store only up top 256 kb of information, which may not be enough if the email has attachments (an invoice for instance). And even though it is possible to split a large message into smaller messages, the complexity grows even more.

Then I realized that I could perfectly imitate the behavior of a queue through a combination of other AWS services, S3 and Lambda, which are much easier to set-up. S3, a cloud object storage solution to store and retrieve data, can act as the repository for uploading the messages, and Lambda, a computing service that runs code in response to events, can pick a message and execute an operation with it.

In other words, we can set-up our email sending process like this:

  1. The application uploads a file with the email content + metadata to an S3 bucket.
  2. Whenever a new file is uploaded into the S3 bucket, S3 triggers an event containing the path to the new file.
  3. A Lambda function picks the event, reads the file, and sends the email.

Finally, we have to decide how to send emails. We can either keep using the SMTP provider that we already have, having the Lambda function interact with their APIs, or use the AWS service for sending emails, called SES. Using SES has both benefits and drawbacks:

Benefits:
  • Very simple to use from within AWS Lambda (it just takes 2 lines of code).
  • It is cheaper: Lambda fees are computed based on the amount of time it takes to execute the function, so connecting to SES from within the AWS network will take a shorter time than connecting to an external server, making the function finish earlier and costing less. (Unless SES is not available in the same region where the application is hosted; in my case, because SES is not offered in the Asian Pacific (Singapore) region, where my EC2 server is located, then I might be better off connecting to some Asia-based external SMTP provider).
Drawbacks:
  • Not many stats for monitoring our sent emails are provided, and adding more powerful ones requires extra effort (eg: tracking what percentage of emails were opened, or what links were clicked, must be set-up through AWS CloudWatch).
  • If we keep using the SMTP provider for sending the priority emails, then we won’t have our stats all together in 1 place.

For simplicity, in the code below we will be using SES.

We have then defined the logic of the process and stack as follows: The application sends priority emails as usual, but for non-priority ones, it uploads a file with email content and metadata to S3; this file is asynchronously processed by a Lambda function, which connects to SES to send the email.

Let’s start implementing the solution.

Differentiating Between Priority And Non-Priority Emails

In short, this all depends on the application, so we need to decide on an email by email basis. I will describe a solution I implemented for WordPress, which requires some hacks around the constraints from function wp_mail. For other platforms, the strategy below will work too, but quite possibly there will be better strategies, which do not require hacks to work.

The way to send an email in WordPress is by calling function wp_mail, and we don’t want to change that (eg: by calling either function wp_mail_synchronous or wp_mail_asynchronous), so our implementation of wp_mail will need to handle both synchronous and asynchronous cases, and will need to know to which group the email belongs. Unluckily, wp_mail doesn’t offer any extra parameter from which we could asses this information, as it can be seen from its signature:

function wp_mail( $to, $subject, $message, $headers = '', $attachments = array() )

Then, in order to find out the category of the email we add a hacky solution: by default, we make an email belong to the priority group, and if $to contains a particular email (eg: nonpriority@asynchronous.mail), or if $subject starts with a special string (eg: “[Non-priority!]“), then it belongs to the non-priority group (and we remove the corresponding email or string from the subject). wp_mail is a pluggable function, so we can override it simply by implementing a new function with the same signature on our functions.php file. Initially, it contains the same code of the original wp_mail function, located in file wp-includes/pluggable.php, to extract all parameters:

if ( !function_exists( 'wp_mail' ) ) :

function wp_mail( $to, $subject, $message, $headers = '', $attachments = array() ) {

  $atts = apply_filters( 'wp_mail', compact( 'to', 'subject', 'message', 'headers', 'attachments' ) );

  if ( isset( $atts['to'] ) ) {
    $to = $atts['to'];
  }

  if ( !is_array( $to ) ) {
    $to = explode( ',', $to );
  }

  if ( isset( $atts['subject'] ) ) {
    $subject = $atts['subject'];
  }

  if ( isset( $atts['message'] ) ) {
    $message = $atts['message'];
  }

  if ( isset( $atts['headers'] ) ) {
    $headers = $atts['headers'];
  }

  if ( isset( $atts['attachments'] ) ) {
    $attachments = $atts['attachments'];
  }

  if ( ! is_array( $attachments ) ) {
    $attachments = explode( "\n", str_replace( "\r\n", "\n", $attachments ) );
  }
  
  // Continue below...
}
endif;

And then we check if it is non-priority, in which case we then fork to a separate logic under function send_asynchronous_mail or, if it is not, we keep executing the same code as in the original wp_mail function:

function wp_mail( $to, $subject, $message, $headers = '', $attachments = array() ) {

  // Continued from above...

  $hacky_email = "nonpriority@asynchronous.mail";
  if (in_array($hacky_email, $to)) {

    // Remove the hacky email from $to
    array_splice($to, array_search($hacky_email, $to), 1);

    // Fork to asynchronous logic
    return send_asynchronous_mail($to, $subject, $message, $headers, $attachments);
  }

  // Continue all code from original function in wp-includes/pluggable.php
  // ...
}

In our function send_asynchronous_mail, instead of uploading the email straight to S3, we simply add the email to a global variable $emailqueue, from which we can upload all emails together to S3 in a single connection at the end of the request:

function send_asynchronous_mail($to, $subject, $message, $headers, $attachments) {
  
  global $emailqueue;
  if (!$emailqueue) {
    $emailqueue = array();
  }
  
  // Add email to queue. Code continues below...
}

We can upload one file per email, or we can bundle them so that in 1 file we contain many emails. Since $headers contains email meta (from, content-type and charset, CC, BCC, and reply-to fields), we can group emails together whenever they have the same $headers. This way, these emails can all be uploaded in the same file to S3, and the $headers meta information will be included only once in the file, instead of once per email:

function send_asynchronous_mail($to, $subject, $message, $headers, $attachments) {
  
  // Continued from above...

  // Add email to the queue
  $emailqueue[$headers] = $emailqueue[$headers] ?? array();
  $emailqueue[$headers][] = array(
    'to' => $to,
    'subject' => $subject,
    'message' => $message,
    'attachments' => $attachments,
  );

  // Code continues below
}

Finally, function send_asynchronous_mail returns true. Please notice that this code is hacky: true would normally mean that the email was sent successfully, but in this case, it hasn’t even been sent yet, and it could perfectly fail. Because of this, the function calling wp_mail must not treat a true response as “the email was sent successfully,” but an acknowledgment that it has been enqueued. That’s why it is important to restrict this technique to non-priority emails so that if it fails, the process can keep retrying in the background, and the user will not expect the email to already be in her inbox:

function send_asynchronous_mail($to, $subject, $message, $headers, $attachments) {
  
  // Continued from above...

  // That's it!
  return true;
}
Uploading Emails To S3

In my previous article “Sharing Data Among Multiple Servers Through AWS S3”, I described how to create a bucket in S3, and how to upload files to the bucket through the SDK. All code below continues the implementation of a solution for WordPress, hence we connect to AWS using the SDK for PHP.

We can extend from the abstract class AWS_S3 (introduced in my previous article) to connect to S3 and upload the emails to a bucket “async-emails” at the end of the request (triggered through wp_footer hook). Please notice that we must keep the ACL as “private” since we don’t want the emails to be exposed to the internet:

class AsyncEmails_AWS_S3 extends AWS_S3 {

  function __construct() {

    // Send all emails at the end of the execution
    add_action("wp_footer", array($this, "upload_emails_to_s3"), PHP_INT_MAX);
  }

  protected function get_acl() {

    return "private";
  }

  protected function get_bucket() {

    return "async-emails";
  }

  function upload_emails_to_s3() {

    $s3Client = $this->get_s3_client();

    // Code continued below...
  }
}
new AsyncEmails_AWS_S3();

We start iterating through the pairs of headers => emaildata saved in global variable $emailqueue, and get a default configuration from function get_default_email_meta for if the headers are empty. In the code below, I only retrieve the “from” field from the headers (the code to extract all headers can be copied from the original function wp_mail):

class AsyncEmails_AWS_S3 extends AWS_S3 {

  public function get_default_email_meta() {

    // Code continued from above...

    return array(
      'from' => sprintf(
        '%s ',
        get_bloginfo('name'),
        get_bloginfo('admin_email')
      ),
      'contentType' => 'text/html',
      'charset' => strtolower(get_option('blog_charset'))
    );
  }

  public function upload_emails_to_s3() {

    // Code continued from above...

    global $emailqueue;
    foreach ($emailqueue as $headers => $emails) {

      $meta = $this->get_default_email_meta();

      // Retrieve the "from" from the headers
      $regexp = '/From:\s*(([^\
Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Using Visual Composer Website Builder To Create WordPress Websites Using Visual Composer Website Builder To Create WordPress Websites Nick Babich 2018-11-12T12:55:03+01:00 2018-11-12T16:40:43+00:00

(This is a sponsored article.) WordPress has changed the way we make websites and millions of people use it to create websites today. But this tool has a few significant limitations — it requires time and coding skills to create a website.

Even when you have aquired coding skills, jumping into code each time when you need to solve a problem (add a new UI element or change styling options for existing one) can be tedious. All too often we hear: “We need to work harder to achieve our goals.” While working hard is definitely important we also need to work smarter.

Today, I’d like to review a tool that will allow us to work smarter. Imagine WordPress without design and technical limits; the tool that reduces the need to hand-code the parts of your website and frees you up to work on more interesting and valuable parts of the design.

In this article, I’ll review the Visual Composer Website Builder and create a real-world example — a landing page for a digital product — just by using this tool.

What Is Visual Composer Website Builder?

Visual Composer Website Builder is a simple and powerful drag-and-drop website builder that promises to change the way we work with WordPress. It introduced a more intuitive way of building a page — all actions involving changing visual hierarchy and content management are done visually. The tool reduces the need to hand-code the theme parts of a website and frees you up to work on valuable parts of the design such as content.

(Large preview)

Content is the most important property of your website. It’s the primary reason why people visit your site — for content. It’s worth putting a lot of efforts in creating good content and use tools that help you to deliver the content in the best way to visitors with the least amount of effort.

Visual Composer And WPBakery

Visual Composer Website Builder is a builder from the creators of WPBakery Page Builder. If you had a chance to use the WPBakery Page builder before you might wonder what the difference between the two plugins is. Let’s be clear about these two products:

There are a few significant differences between the two:.

  • The key difference between WPBakery Page builder and Visual Composer is that WPBakery is only for the content part, while with Visual Composer Website Builder you can create a complete website (including Headers and Footers).
  • Visual Composer is not shortcode based, which helps generate clean code. Also, disabling the plugin won’t leave you with “shortcode hell” (a situation when shortcodes can’t be rendered without an activated plugin).

You can check the full list of differences between two plugins here.

Now, Visual Composer Website Builder is not an ‘advanced’ version of WPBakery. It is an entirely new product that was created to satisfy the growing needs of web professionals. Visual Composer is not just a plugin; it’s a powerful platform that can be extended as user needs continue evolving.

A Quick List Of Visual Composer’s Features

While I’ll show you how Visual Composer works in action below, it’s worth to point out a few key benefits of this tool:

  • It’s a live-preview editor with drag-and-drop features, and hundreds of ready-to-use content elements that bring a lot of design freedom. You can make changes instantly and see end-results before publishing.
  • Two ways of page editing — using frontend editor and tree view. Tree view allows navigating through the elements available on a page and makes a design process much easier.
  • Ready-to-use WordPress templates for all types of pages — from landing pages and portfolios to business websites with dedicated product pages, because editing an existing template is a lot easier than starting from scratch with a blank page.
  • Visual Composer works with any theme (i.e. it’s possible to integrate Visual Composer Website builder into your existing themes)
  • Responsive design out-of-the-box. All the elements and templates are responsive and mobile-ready. You can adjust responsiveness for each independent column.
  • Header, footer, and sidebar editor. Usually the header, footer and sidebar are defined by the theme you’re using. When web professionals need to change them, they usually move to code. But with Visual Composer, you can change the layout quickly and easily using only the visual editor. This feature is available in the Premium version of the product.
  • An impressive collection of add-ons (it’s possible to get add-ons from the Hub or get them from third-party developers)

There are also three features that make Visual Composer stand out from the crowd. Here they are:

1. Visual Composer Hub

Visual Composer Hub is a cloud which stores all the elements available to the users. It’s basically like a design system that keeps itself updated and where you can get new elements, templates, elements, blocks (soon).

(Large preview)

The great thing about Visual Composer Hub is that you don’t need to update the plugin to get new elements — you can download the elements whenever you need them. As a result, your WP setup isn’t bloated with a myriad of unused elements.

2. New Technical Stack

Visual Composer Website builder is built on a new technology stack — it’s powered by ReactJS and doesn’t use any of the WordPress shortcodes. This helps to achieve better performance — the team behind Visual Composer conducted a series of internal tests and showed that pages created with Visual Composer load 1-1.5s faster than the same layouts re-created with WPBakery.

3. API

Visual Composer Website builder has a well-documented open API. If you have coding skills, you can extend Visual Composer with your own custom elements which may be helpful for some custom projects.

How To Create A Landing Page With Visual Composer

In this section, I’ll show how to create a landing page for a digital product called CalmPod (a fictional home speaker device) with the new Visual Composer Website Builder.

Our journey starts in a WP interface where we need to create a new page — give it a title and click the ‘Edit with Visual Composer button.’

(Large preview) Creating A Layout For A Landing Page

The process of creating the page starts with building an appropriate layout. Usually building a layout for a landing page takes a lot of time and effort. Designers have to try a lot of different approaches before finding the one that works the best for the content. But Visual Composer simplifies the task for designers — it provides a list of ready-to-use layouts (available under the Add Template option). So, all you need to do to create a new page is to find the appropriate layout from the list of available options and see how it works for your content.

You can start with a blank page or select a ready-to-use template. (Large preview)

But for our example, we’ll select the Startup Page template. This template applies automatically as soon as we click the + symbol, so all we need to do is to modify it according to our needs.

(Large preview)

Each layout in Visual Composer consists of rows and columns. The row is a base that defines the logical structure of the page. Each row consists of columns. Visual Composer gives you the ability to control the number of columns in a row.

(Large preview)

Tip: Notice that Visual Composer uses different colored borders for UI units. When we select a row, we see a blue-colored border, when we select a column, we see an orange-colored border. This feature can be extremely valuable when you work on creating complex layouts.

(Large preview) (Large preview)

The great thing about Visual Composer is that we can customize all properties of the layout — add/remove elements or change their styling options (such as margins, padding between elements). For example, we don’t need to dive into the code to alter the sizes of columns; we can simply drag and drop the borders of individual elements.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
CSS Frameworks Or CSS Grid: What Should I Use For My Project? CSS Frameworks Or CSS Grid: What Should I Use For My Project? Rachel Andrew 2018-11-09T12:30:30+02:00 2018-11-09T15:49:47+00:00

Among the questions I am most frequently asked is some variety of the question, “Should I use CSS Grid or Bootstrap?” In this article, I will take a look at that question. You will discover that the reasons for using frameworks are varied, and not simply centered around use of the grid system contained in that framework. I hope that by unpacking these reasons, I can help you to make your own decision, in terms of what is best for the sites and applications that you are working on, and also for the team you work with.

In this article when I talk about a framework, I’m describing a third party CSS framework such as Bootstrap or Foundation. You might argue these are really component libraries, but many people (including their own docs) would describe them as a framework so that is what we will use here. The important factor is that these are something developed externally to you, without reference to your specific issues. The alternative to using a third party framework is to write your own CSS — that might involve developing your own internal framework, using a bunch of common files as a starting point, or creating every project as a new thing. All these things are done in reference to your own specific needs rather than very generic ones.

Why Choose A CSS Framework?

The question of whether to use Grid or a framework is flawed, as CSS Grid is not a drop-in replacement for the things that a CSS framework does. Any exploration of the subject needs to consider what of our framework CSS Grid is going to replace. I wanted to start by finding out why people had chosen to use a CSS framework at all. So I turned to Twitter and posted this tweet.

A question: if you have chosen to use a CSS framework (Bootstrap, Foundation etc.) for your project, what were the main reasons for doing so?

— Rachel Andrew (@rachelandrew) October 16, 2018

There were a lot of responses. As I expected, there are far more reasons to use a framework than simply the grid system that it contains.

Getting workflow just right ain’t an easy task. So are proper estimates. Or alignment among different departments. That’s why we’ve set up “this-is-how-I-work”-sessions — with smart cookies sharing what works well for them. A part of the Smashing Membership, of course.

Explore Smashing Membership ↬
A Framework Gives Your Team Ready Made Documentation

If you are working on a project with a number of other developers then any internal system you create will need to also include documentation to help your team members use it effectively. Creating useful documentation is time-consuming, skilled work in itself, and something that the big frameworks do very well.

The Bootstrap Documentation. (Large preview)

Framework documentation came up again and again, with many experienced front-end developers chipping in and explaining this is why they would recommend and use a CSS framework. I sometimes hear the opinion that people are using frameworks because they don’t really know CSS, many of the people replying, however, are well known to me as expert CSS developers. I’m sure that they are sometimes frustrated by the choices made by the framework, however, the positive aspects of that choice outweigh that.

Online Communities: Easy Access To Help

When you decide to use a particular tool, you also gain a community of users to ask for help. Unless you have a very clear CSS issue, and can produce a reduced use case to demonstrate it, asking for help with CSS can be difficult. It is especially so if you want to ask how to approach building a certain component. Using a framework can give you a starting point for your question; in general, you will be asking how to modify or style a particular component rather than starting from scratch. This is an easier thing to ask, as well as an easier thing to answer.

The Grid System

Despite the fact that we have CSS Grid, many people replied that the main reason they decided to use a framework was for the grid system. Of course, many of these projects may have been started a long time before CSS Grid was available. Even today, however, concerns about backwards compatibility or team understanding of newer layout methods might cause people to decide to use a framework rather than adopting native CSS.

Speed Of Project Delivery

when having a unique design and efficient css were a much lower priority than the due date

— CodingBlocks.net (@CodingBlocks) October 16, 2018

Opting for a framework will, in general, make it far quicker to deliver your project, in particular if that project fits very well with the way the framework does things and doesn’t need a lot of customization.

In the case of developing an MVP for a new idea, a framework may well be an excellent choice. You will have many things to spend time on, and be still testing assumptions in terms of what the project needs. Being able to develop that first version using a framework can help you get the product in front of users more quickly, and save burning up a lot of time developing things you then decide not to use.

I go with a framework on anything that’s an MVP — minimal dev effort on tooling and frameworks to optimise for time working on the actual concept.

I make my own CSS “framework” on anything that’s a redo of an existing large-scale site/app.

— Ben Bodien (@bbodien) October 16, 2018

Another place where speed and a bunch of ready built components can be very useful is when developing the backend admin system for a site or application. In the case where you simply need to create a few admin screens, a framework can save a lot of time styling form fields and other components! There are even dashboard themes for Bootstrap and Foundation that can give a helpful starting point.

Collections of dasboard components make it quicker to build out the admin for an app. (Large preview) I’m Not A Designer!

This point is the reason I’ve opted for a CSS framework in the past. I’m not a designer, and if I have to both design and build something, I’ll spend a long time trying to make design decisions I am entirely unqualified to make. It would be lovely to have the funds to hire a designer for every side project, however, I don’t, and so a framework might mean the difference between shipping the thing and not.

Dealing With CSS Bugs And Browser Compatibility Issues

Mentioned less than I thought it might be was the fact that the framework authors would already have dealt with browser issues, be that due to actual bugs or lack of support for certain features. However, this was still a factor in the decision-making for many people.

To Help With Responsive Design

Responsiveness of web pages . I found it difficult for me to decide breakpoints for web pages.

— Faisal Ali (@f_a_akhtar) October 18, 2018

This came up a few times; people were opting for a framework specifically due to the fact it was responsive, or that I made decisions about breakpoints for them. I thought it interesting that this specifically was something called out as a factor in choosing to use a framework.

Why Not Use A Framework?

Among positive reasons why frameworks had been selected were some of the issues that people have had with that choice.

Difficulty Of Overriding Framework Code

Many people commented on the fact that it could become difficult to override the code used in the framework, and that frameworks were a good choice if they didn’t need a lot of overriding. The benefits of ease of use, and everyone on the team understanding how to use the framework can be lost if there are then a huge number of customizations in place.

All Websites End Up Looking The Same

The blame for all websites starting to look the same has been placed squarely at the door of the well known CSS frameworks. I have seen sites where I am sure a certain framework has been used, then discover they are custom CSS, so prevalent are the design choices made in these frameworks.

The difficulty in overriding framework styles already mentioned is a large part of why sites developed using a particular framework will tend to look similar. This isn’t just a creative issue, it can be very odd as a user of a few websites which have all opted for the same framework to feel that they are all the same thing. In terms of conveying your brand, and making good user experience part of that, perhaps you lose something when opting for the generic choices of a framework.

Inheriting The CSS Problems Of The Entire World

Whether front or back-end, any tool or framework that seeks to hit the mainstream has to solve as many problems as possible. Unless the tool is tightly coupled to solving one particular use-case it is going to contain a lot of very generic code, and code which solves problems that you do not have, and will never have.

You may be in the fortunate position of only needing your full experience to be viewed in the most current browsers, allowing for a more limited experience in Internet Explorer, or older versions of Chrome. Using a framework with lots of built-in support going back to IE9 would result in lots of additional code — especially given the improvements in CSS layout recently. It might also prevent you from being creative, as everything in the framework is assuming this requirement for support. Things which are possible using CSS may well be limited by the framework.

As an example, the grid systems in popular frameworks do not have an ability to span rows, as there isn’t any concept or rows in layout systems prior to Grid Layout. CSS Grid Layout easily allows for this. If you are tied to the Bootstrap Grid and your designer comes up with a design that includes elements which span rows, you are left unable to implement it — despite the fact that Grid might be supported by your target browsers.

Performance Issues

Related to the above are performance issues inherent in using fairly generic code, rather than something optimized for the exact use cases that you have. When trying to improve performance you will find yourself hitting up against the decisions of the framework.

Increased Technical Debt

Speed, mainly, though we quickly found out it was a bit of a false economy because it created substantial technical debt in the long run.

— Tim Cthulhuegdon (@nefarioustim) October 16, 2018

While a framework might be a great way to get your startup quickly off the ground, and at the time of making that decision you are sure that you will replace it, is there a plan to make that happen?

Learning A Framework Rather Than Learning CSS

When talking to conference and workshop attendees, I have discovered that many people have only ever used a framework to write CSS. There is nothing wrong with coming into web development via one of these tools, given the complexity of the web platform today I imagine that will be the route in for many people. However, it can become a career-limiting choice, especially if the framework you based your skills around falls out of favor.

Having front-end developers without CSS knowledge should worry a company. It makes it incredibly hard to move away from that framework if your team doesn’t actually understand how to do CSS without it. While this isn’t really a reason not to use a framework, it is something to bear in mind when using one. When the day comes to move away you would hope that the team will be ready to take on something new, not needing to remember (or learn for the first time) how to write CSS!

The Choice Doesn’t Consider End Users

Nicole Sullivan asked pretty much the same question a few days prior to my question as I was thinking about writing this article, although she was considering front-end frameworks as a whole rather than just CSS frameworks. Jeremy Keith noted that precisely zero of the answers concerned end users. This was also the case with the responses to my question.

In our race to get our site built quickly, our desire to make things as good as possible for ourselves as the designers and developers of the site, do we forget who we are doing this for? Do the decisions made by the framework developer match up with the needs of the users of the site you are building?

Can We Replace Frameworks With “Vanilla” CSS?

If you are considering replacing your framework or starting a new project without one, what are some of the things that you could consider in order to make that process easier?

Understand Which Parts Of The Framework You Need

If you are replacing the use of a framework with your own CSS, a good place to start would be to audit your use of the current framework. Work out what you are using and why. Consider how you will replace those things in the new design.

You could follow a similar process when thinking about whether to select a framework or write your own. What parts of this could you reasonably expect to need? How well does it fit with your requirements? Will there be a lot of code that you import, potentially ask visitors to download, but never make use of?

Create A Documented Pattern Library Or Style Guide

I am a huge fan of working with pattern libraries and you can read my post here on Smashing Magazine about our use of Fractal. A pattern library or a style guide enables the creation of documentation along with all of your components. I start all of my projects by working on the CSS in the pattern library.

You are still going to need to write the documentation, as someone who writes documentation, however, I know that often the hardest thing is knowing where to start and how to structure the docs. A pattern library helps with this by keeping the docs along with the CSS for the component itself. This approach can also help prevent the docs becoming out of date as they are tightly linked to the component they refer to.

Develop Your Own CSS Code Guidelines

Consistency across the team is incredibly useful, and without a framework, there may be nothing dictating that. With newer layout methods, in particular, there are often several ways in which a pattern could be built, if everyone picks a different one then inconsistencies are likely to creep in.

Better Places To Ask For Help

Other than sending people in the direction of Stack Overflow, it seems that there are very few places to ask for help on CSS. In particular there seem to be few places which are approachable for beginners. If we are to encourage people away from third-party tools then we need to fill that need for friendly, helpful support which comes from the communities around those tools.

Within a company, it is possible that more experienced developers can become the CSS support for newer team members. If moving away from a framework to your own solution, it would be wise to consider what training might be needed to help bridge the gap, especially if people are used to using the help provided around the third party tool when they needed help in the past.

Style Guides Or Starting Points For Non-Designers

I tie myself in knots with questions such as, “Which fonts should I use?”, “How big should the headings be in relationship to the body text?”, “Is it OK to use a drop shadow?” I can easily write the CSS — if I know what I’m trying to do! What I really need are some rules for making a not-terrible design, some kind of typography starting point, or a set of basic guidelines would replace being able to use the defaults of a framework in a lot of cases.

Educating People About The State Of Modern Browser Interoperability

I have discovered that people who have been immersed in framework-based development for a number of years, often have a view of browser interoperability which is several years out of date. We have never been in a better situation in terms of CSS working cross-browser. It may be that some browsers don’t support one new shiny bit of CSS, but in general, CSS (when supported) won’t be full of strange bugs. For example, in almost all cases if you use CSS Grid in one browser your CSS will work in exactly the same way in another.

If trying to make a case for not using a framework to team members who believe that the framework saves them from browser bugs, this point may be a useful one to raise. Are the browser compatibility problems real, or based on the concerns of the past?

Will We See A New Breed Of Frameworks?

Something that interests me is whether our new layout methods will help usher in a new breed of tools and frameworks. Will we see tools which take advantage of new layout methods, allow for more creativity but still give teams and individuals some of the undeniable advantages that came out of the responses to my tweet.

Perhaps by relying on new layout methods, rather than an inbuilt grid system, a new-style framework could be much lighter, becoming a collection of useful components. It might be able to then get away from some of the performance issues inherent in very generic code.

An area a framework could help with would be in helping to create solid fallbacks for browsers which don’t support newer layout methods, or by having really solid accessibility baked into the components. This could help provide guidance into a way of working that considers interoperability and accessibility, even for those people who don’t have these things in the forefront of their minds.

I don’t think that simply switching the Bootstrap Grid for CSS Grid Layout will achieve this. Instead, authors coming up with new frameworks, should perhaps look at some of the reasons outlined here, and try to solve them in new ways, using the new functionality we have in CSS to do that.

Should You Use A Framework?

You and your team will need to answer that question yourself. And, despite what people might try to have you believe, there is no universal right or wrong answer. There is only what is right or wrong for your project. I hope that this article and the many responses to my original question might give you some things to discuss as you ponder that question.

Remember that the answer will change over time. It might be a useful thought experiment to not only consider what you need right now, in terms of initially doing the development for the site, but consider the lifespan of the site. Do you expect this still to be around in five years? Will your choice be a positive or negative one then?

Document your decisions, don’t be afraid to revisit them, and do ensure that you and your team maintain your skills outside of any framework that you decide to use. That way, you will be in a good place to move on in future and to make the best decisions for the next phases of your project.

I’d love the conversation started in that tweet to continue. Let us know your stories in the comments — they may well help..

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Sharing Data Among Multiple Servers Through AWS S3 Sharing Data Among Multiple Servers Through AWS S3 Leonardo Losoviz 2018-11-08T12:30:30+01:00 2018-11-08T11:47:40+00:00

When providing some functionality for processing a file uploaded by the user, the file must be available to the process throughout the execution. A simple upload and save operation presents no issues. However, if in addition the file must be manipulated before being saved, and the application is running on several servers behind a load balancer, then we need to make sure that the file is available to whichever server is running the process at each time.

For instance, a multi-step “Upload your user avatar” functionality may require the user to upload an avatar on step 1, crop it on step 2, and finally save it on step 3. After the file is uploaded to a server on step 1, the file must be available to whichever server handles the request for steps 2 and 3, which may or may not be the same one for step 1.

A naive approach would be to copy the uploaded file on step 1 to all other servers, so the file would be available on all of them. However, this approach is not just extremely complex but also unfeasible: for instance, if the site runs on hundreds of servers, from several regions, then it cannot be accomplished.

A possible solution is to enable “sticky sessions” on the load balancer, which will always assign the same server for a given session. Then, steps 1, 2 and 3 will be handled by the same server, and the file uploaded to this server on step 1 will still be there for steps 2 and 3. However, sticky sessions are not fully reliable: If in between steps 1 and 2 that server crashed, then the load balancer will have to assign a different server, disrupting the functionality and the user experience. Likewise, always assigning the same server for a session may, under special circumstances, lead to slower response times from an overburdened server.

A more proper solution is to keep a copy of the file on a repository accessible to all servers. Then, after the file is uploaded to the server on step 1, this server will upload it to the repository (or, alternatively, the file could be uploaded to the repository directly from the client, bypassing the server); the server handling step 2 will download the file from the repository, manipulate it, and upload it there again; and finally the server handling step 3 will download it from the repository and save it.

Getting workflow just right ain’t an easy task. So are proper estimates. Or alignment among different departments. That’s why we’ve set up “this-is-how-I-work”-sessions — with smart cookies sharing what works well for them. A part of the Smashing Membership, of course.

Explore Smashing Membership ↬

In this article, I will describe this latter solution, based on a WordPress application storing files on Amazon Web Services (AWS) Simple Storage Service (S3) (a cloud object storage solution to store and retrieve data), operating through the AWS SDK.

Note 1: For a simple functionality such as cropping avatars, another solution would be to completely bypass the server, and implement it directly in the cloud through Lambda functions. But since this article is about connecting an application running on the server with AWS S3, we don’t consider this solution.

Note 2: In order to use AWS S3 (or any other of the AWS services) we will need to have a user account. Amazon offers a free tier here for 1 year, which is good enough for experimenting with their services.

Note 3: There are 3rd party plugins for uploading files from WordPress to S3. One such plugin is WP Media Offload (the lite version is available here), which provides a great feature: it seamlessly transfers files uploaded to the Media Library to an S3 bucket, which allows to decouple the contents of the site (such as everything under /wp-content/uploads) from the application code. By decoupling contents and code, we are able to deploy our WordPress application using Git (otherwise we cannot since user-uploaded content is not hosted on the Git repository), and host the application on multiple servers (otherwise, each server would need to keep a copy of all user-uploaded content.)

Creating The Bucket

When creating the bucket, we need to pay consideration to the bucket name: Each bucket name must be globally unique on the AWS network, so even though we would like to call our bucket something simple like “avatars”, that name may already be taken, then we may choose something more distinctive like “avatars-name-of-my-company”.

We will also need to select the region where the bucket is based (the region is the physical location where the data center is located, with locations all over the world.)

The region must be the same one as where our application is deployed, so that accessing S3 during the process execution is fast. Otherwise, the user may have to wait extra seconds from uploading/downloading an image to/from a distant location.

Note: It makes sense to use S3 as the cloud object storage solution only if we also use Amazon’s service for virtual servers on the cloud, EC2, for running the application. If instead, we rely on some other company for hosting the application, such as Microsoft Azure or DigitalOcean, then we should also use their cloud object storage services. Otherwise, our site will suffer an overhead from data traveling among different companies’ networks.

In the screenshots below we will see how to create the bucket where to upload the user avatars for cropping. We first head to the S3 dashboard and click on “Create bucket”:

S3 dashboard, showing all our existing buckets. (Large preview)

Then we type in the bucket name (in this case, “avatars-smashing”) and choose the region (“EU (Frankfurt)”):

Creating a bucket through in S3. (Large preview)

Only the bucket name and region are mandatory. For the following steps we can keep the default options, so we click on “Next” until finally clicking on “Create bucket”, and with that, we will have the bucket created.

Setting Up The User Permissions

When connecting to AWS through the SDK, we will be required to enter our user credentials (a pair of access key ID and secret access key), to validate that we have access to the requested services and objects. User permissions can be very general (an “admin” role can do everything) or very granular, just granting permission to the specific operations needed and nothing else.

As a general rule, the more specific our granted permissions, the better, as to avoid security issues. When creating the new user, we will need to create a policy, which is a simple JSON document listing the permissions to be granted to the user. In our case, our user permissions will grant access to S3, for bucket “avatars-smashing”, for the operations of “Put” (for uploading an object), “Get” (for downloading an object), and “List” (for listing all the objects in the bucket), resulting in the following policy:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "s3:Put*",
                "s3:Get*",
                "s3:List*"
            ],
            "Resource": [
                "arn:aws:s3:::avatars-smashing",
                "arn:aws:s3:::avatars-smashing/*"
            ]
        }
    ]
}

In the screenshots below, we can see how to add user permissions. We must go to the Identity and Access Management (IAM) dashboard:

IAM dashboard, listing all the users we have created. (Large preview)

In the dashboard, we click on “Users” and immediately after on “Add User”. In the Add User page, we choose a user name (“crop-avatars”), and tick on “Programmatic access” as the Access type, which will provide the access key ID and secret access key for connecting through the SDK:

Adding a new user. (Large preview)

We then click on button “Next: Permissions”, click on “Attach existing policies directly”, and click on “Create policy”. This will open a new tab in the browser, with the Create policy page. We click on the JSON tab, and enter the JSON code for the policy defined above:

Creating a policy granting 'Get', 'Post' and 'List' operations on the 'avatars-smashing' bucket. (Large preview)

We then click on Review policy, give it a name (“CropAvatars”), and finally click on Create policy. Having the policy created, we switch back to the previous tab, select the CropAvatars policy (we may need to refresh the list of policies to see it), click on Next: Review, and finally on Create user. After this is done, we can finally download the access key ID and secret access key (please notice that these credentials are available for this unique moment; if we don’t copy or download them now, we’ll have to create a new pair):

After the user is created, we are offered a unique time to download the credentials. (Large preview) Connecting To AWS Through The SDK

The SDK is available through a myriad of languages. For a WordPress application, we require the SDK for PHP which can be downloaded from here, and instructions on how to install it are here.

Once we have the bucket created, the user credentials ready, and the SDK installed, we can start uploading files to S3.

Uploading And Downloading Files

For convenience, we define the user credentials and the region as constants in the wp-config.php file:

define ('AWS_ACCESS_KEY_ID', '...'); // Your access key id
define ('AWS_SECRET_ACCESS_KEY', '...'); // Your secret access key
define ('AWS_REGION', 'eu-central-1'); // Region where the bucket is located. This is the region id for "EU (Frankfurt)"

In our case, we are implementing the crop avatar functionality, for which avatars will be stored on the “avatars-smashing” bucket. However, in our application we may have several other buckets for other functionalities, requiring to execute the same operations of uploading, downloading and listing files. Hence, we implement the common methods on an abstract class AWS_S3, and we obtain the inputs, such as the bucket name defined through function get_bucket, in the implementing child classes.

// Load the SDK and import the AWS objects
require 'vendor/autoload.php';
use Aws\S3\S3Client;
use Aws\Exception\AwsException;

// Definition of an abstract class
abstract class AWS_S3 {
  
  protected function get_bucket() {

    // The bucket name will be implemented by the child class
    return '';
  }
}

The S3Client class exposes the API for interacting with S3. We instantiate it only when needed (through lazy-initialization), and save a reference to it under $this->s3Client as to keep using the same instance:

abstract class AWS_S3 {

  // Continued from above...

  protected $s3Client;

  protected function get_s3_client() {

    // Lazy initialization
    if (!$this->s3Client) {

      // Create an S3Client. Provide the credentials and region as defined through constants in wp-config.php
      $this->s3Client = new S3Client([
        'version' => '2006-03-01',
        'region' => AWS_REGION,
        'credentials' => [
          'key' => AWS_ACCESS_KEY_ID,
          'secret' => AWS_SECRET_ACCESS_KEY,
        ],
      ]);
    }

    return $this->s3Client;
  }
}

When we are dealing with $file in our application, this variable contains the absolute path to the file in disk (e.g. /var/app/current/wp-content/uploads/users/654/leo.jpg), but when uploading the file to S3 we should not store the object under the same path. In particular, we must remove the initial bit concerning the system information (/var/app/current) for security reasons, and optionally we can remove the /wp-content bit (since all files are stored under this folder, this is redundant information), keeping only the relative path to the file (/uploads/users/654/leo.jpg). Conveniently, this can be achieved by removing everything after WP_CONTENT_DIR from the absolute path. Functions get_file and get_file_relative_path below switch between the absolute and the relative file paths:

abstract class AWS_S3 {

  // Continued from above...

  function get_file_relative_path($file) {

    return substr($file, strlen(WP_CONTENT_DIR));
  }

  function get_file($file_relative_path) {

    return WP_CONTENT_DIR.$file_relative_path;
  }
}

When uploading an..

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
How To Build A Virtual Reality Model With A Real-Time Cross-Device Preview How To Build A Virtual Reality Model With A Real-Time Cross-Device Preview Alvin Wan 2018-11-06T23:50:16+01:00 2018-11-07T10:45:16+00:00

Virtual reality (VR) is an experience based in a computer-generated environment; a number of different VR products make headlines and its applications range far and wide: for the winter Olympics, the US team utilized virtual reality for athletic training; surgeons are experimenting with virtual reality for medical training; and most commonly, virtual reality is being applied to games.

We will focus on the last category of applications and will specifically focus on point-and-click adventure games. Such games are a casual class of games; the goal is to point and click on objects in the scene, to finish a puzzle. In this tutorial, we will build a simple version of such a game but in virtual reality. This serves as an introduction to programming in three dimensions and is a self-contained getting-started guide to deploying a virtual reality model on the web. You will be building with webVR, a framework that gives a dual advantage — users can play your game in VR and users without a VR headset can still play your game on a phone or desktop.

Developing For Virtual Reality

Any developer can create content for VR nowadays. To get a better understanding of VR development, working a demo project can help. Read article →

In the second half of these tutorial, you will then build a “mirror” for your desktop. This means that all movements the player makes on a mobile device will be mirrored in a desktop preview. This allows you see what the player sees, allowing you to provide guidance, record the game, or simply keep guests entertained.

Prerequisites

To get started, you will need the following. For the second half of this tutorial, you will need a Mac OSX. Whereas the code can apply to any platform, the dependency installation instructions below are for Mac.

  • Internet access, specifically to glitch.com;
  • A virtual reality headset (optional, recommended). I use Google Cardboard, which is offered at $15 a piece.

Front-end is messy and complicated these days. That's why we publish articles, printed books and webinars with useful techniques to improve your work. Even better: Smashing Membership with a growing selection of front-end & UX goodies. So you get your work done, better and faster.

Explore Smashing Membership ↬
Step 1: Setting Up A Virtual Reality (VR) Model

In this step, we will set up a website with a single static HTML page. This allows us to code from your desktop and automatically deploy to the web. The deployed website can then be loaded on your mobile phone and placed inside a VR headset. Alternatively, the deployed website can be loaded by a standalone VR headset. Get started by navigating to glitch.com. Then,

  1. Click on “New Project” in the top-right.
  2. Click on “hello-express” in the drop-down.
(Large preview)

Next, click on views/index.html in the left sidebar. We will refer to this as your “editor”.

(Large preview)

To preview the webpage, click on “Preview” in the top left. We will refer to this as your preview. Note that any changes in your editor will be automatically reflected in this preview, barring bugs or unsupported browsers.

(Large preview)

Back in your editor, replace the current HTML with the following boilerplate for a VR model.

<!DOCTYPE html>
<html>
  <head>
      <script src="https://aframe.io/releases/0.7.0/aframe.min.js"></script>
  </head>
  <body>
    <a-scene>
      
      <!-- blue sky -->
      <a-sky color="#a3d0ed"></a-sky>
      
      <!-- camera with wasd and panning controls -->
      <a-entity camera look-controls wasd-controls position="0 0.5 2" rotation="0 0 0"></a-entity>
        
      <!-- brown ground -->
      <a-box shadow id="ground" shadow="receive:true" color="#847452" width="10" height="0.1" depth="10"></a-box>
        
      <!-- start code here -->
      <!-- end code here -->
    </a-scene>
  </body>
</html>

Navigate see the following.

(Large preview)

To preview this on your VR headset, use the URL in the omnibar. In the picture above, the URL is https://point-and-click-vr-game.glitch.me/. Your working environment is now set up; feel free to share this URL with family and friends. In the next step, you will create a virtual reality model.

Step 2: Build A Tree Model

You will now create a tree, using primitives from aframe.io. These are standard objects that Aframe has pre-programmed for ease of use. Specifically, Aframe refers to objects as entities. There are three concepts, related to all entities, to organize our discussion around:

  1. Geometry and material,
  2. Transformation Axes,
  3. Relative Transformations.

First, geometry and material are two building blocks of all three-dimensional objects in code. The geometry defines the “shape” — a cube, a sphere, a pyramid, and so on. The material defines static properties of the shape, such as color, reflectiveness, roughness.

Aframe simplifies this concept for us by defining primitives, such as <a-box>, <a-sphere>, <a-cylinder> and many others to make a specification of a geometry and its material simpler. Start by defining a green sphere. On line 19 in your code, right after <!-- start code here -->, add the following.

      <!-- start code here -->
      <a-sphere color="green" radius="0.5"></a-sphere>  <!-- new line -->
      <!-- end code here -->

Second, there are three axes to transform our object along. The x axis runs horizontally, where x values increase as we move right. The y axis runs vertically, where y values increase as we move up. The z axis runs out of your screen, where z values increase as we move towards you. We can translate, rotate, or scale entities along these three axes.

For example, to translate an object “right,” we increase its x value. To spin an object like a top, we rotate it along the y-axis. Modify line 19 to move the sphere “up” — this means you need to increase the sphere’s y value. Note that all transformations are specified as <x> <y> <z>, meaning to increase its y value, you need to increase the second value. By default, all objects are located at position 0, 0, 0. Add the position specification below.

      <!-- start code here -->
      <a-sphere color="green" radius="0.5" position="0 1 0"></a-sphere> <!-- edited line -->
      <!-- end code here -->

Third, all transformations are relative to its parent. To add a trunk to your tree, add a cylinder inside of the sphere above. This ensures that the position of your trunk is relative to the sphere’s position. In essence, this keeps your tree together as one unit. Add the <a-cylinder> entity between the <a-sphere ...> and </a-sphere> tags.

      <a-sphere color="green" radius="0.5" position="0 1 0">
        <a-cylinder color="#84651e" position="0 -0.9 0" radius="0.05"></a-cylinder> <!-- new line -->
      </a-sphere>

To make this treeless barebones, add more foliage, in the form of two more green spheres.

      <a-sphere color="green" radius="0.5" position="0 0.75 0">
        <a-cylinder color="#84651e" position="0 -0.9 0" radius="0.05"></a-cylinder>
        <a-sphere color="green" radius="0.35" position="0 0.5 0"></a-sphere> <!-- new line -->
        <a-sphere color="green" radius="0.2" position="0 0.8 0"></a-sphere> <!-- new line -->
      </a-sphere>

Navigate back to your preview, and you will see the following tree:

(Large preview)

Reload the website preview on your VR headset, and check out your new tree. In the next section, we will make this tree interactive.

Step 3: Add Click Interaction To Model

To make an entity interactive, you will need to:

  • Add an animation,
  • Have this animation trigger on click.

Since the end user is using a virtual reality headset, clicking is equivalent to staring: in other words, stare at an object to “click” on it. To effect these changes, you will start with the cursor. Redefine the camera, by replacing line 13 with the following.

<a-entity camera look-controls wasd-controls position="0 0.5 2" rotation="0 0 0">
  <a-entity cursor="fuse: true; fuseTimeout: 250"
            position="0 0 -1"
            geometry="primitive: ring; radiusInner: 0.02; radiusOuter: 0.03"
            material="color: black; shader: flat"
            scale="0.5 0.5 0.5"
            raycaster="far: 20; interval: 1000; objects: .clickable">
    <!-- add animation here -->
  </a-entity>
</a-entity>

The above adds a cursor that can trigger the clicking action. Note the objects: .clickable property. This means that all objects with the class “clickable” will trigger the animation and receive a “click” command where appropriate. You will also add an animation to the click cursor, so that users know when the cursor triggers a click. Here, the cursor will shrink slowly when pointing at a clickable object, snapping after a second to denote an object has been clicked. Replace the comment <!-- add animation here --> with the following code:

<a-animation begin="fusing" easing="ease-in" attribute="scale"
  fill="backwards" from="1 1 1" to="0.2 0.2 0.2" dur="250"></a-animation>

Move the tree to the right by 2 units and add class “clickable” to the tree, by modifying line 29 to match the following.

<a-sphere color="green" radius="0.5" position="2 0.75 0" >

Next, you will:

  • Specify an animation,
  • Trigger the animation with a click.

Due to Aframe’s easy-to-use animation entity, both steps can be done in quick succession.

Add an <a-animation> tag on line 33, right after the <a-cylinder> tag but before the end of the </a-sphere>.

<a-animation begin="click" attribute="position" from="2 0.75 0" to="2.2 0.75 0" fill="both" direction="alternate" repeat="1"></a-animation>

The above properties specify a number of configurations for the animation. The animation:

  • Is triggered by the click event
  • Modifies the tree’s position
  • Starts from the original position 2 0.75 0
  • Ends in 2.2 0.75 0 (moving 0.2 units to the right)
  • Animates when traveling to and from the destination
  • Alternates animation between traveling to and from the destination
  • Repeats this animation once. This means the object animates twice in total — once to the destination and once back to the original position.

Finally, navigate to your preview, and drag from the cursor to your tree. Once the black circle rests on the tree, the tree will move to the right and back.

Large preview

This concludes the basics needed to build a point-and-click adventure game, in virtual reality. To view and play a more complete version of this game, see the following short scene. The mission is to open the gate and hide the tree behind the gate, by clicking on various objects in the scene.

(Large preview)

Next, we set up a simple nodeJS server to serve our static demo.

Step 4: Setup NodeJS Server

In this step, we will set up a basic, functional nodeJS server that serves your existing VR model. In the left sidebar of your editor, select package.json.

Start by..

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
The Value Of Storyboarding For Product Design The Value Of Storyboarding For Product Design Joshua Bumcrot 2018-11-06T14:15:20+01:00 2018-11-06T13:55:06+00:00

When you think of the word “storyboarding,” you probably think about film, media, and video creation. Historically, this is what storyboards have been used for, but who says we can’t use them for product development as well?

Storyboards are an effective communication and product development tool for digital marketers, content creators, user experience specialists, and product managers. In this article, I’ll teach you why storyboarding is valuable to the product development process, as well as why it can reduce interdepartmental miscommunications as well as overhead costs.

Below is a comic of a clear and common situation experienced by most companies. Words can only go so far as to effective explanations and miscommunications are are a common occurrence in the product development world. As you read the article, keep this image in mind and notice examples where storyboarding or visual communication may be able to prevent misunderstandings.

The ‘Tree Swing’ communication problem (Large preview)

Web forms are such an important part of the web, but we design them poorly all the time. The brand-new “Form Design Patterns” book is our new practical guide for people who design, prototype and build all sorts of forms for digital services, products and websites. The eBook is free for Smashing Members.

Check the table of contents ↬

I’m going to set a scene that almost all tech companies experience on a regular basis:

Marketer Mary has an idea for a new product. Mary thinks this product will be a huge winner and goes to tech to have them build it.

Developer Diane hears Mary’s idea and tells her she will start working on it. As Diane continues to build out functionality, she realizes that the product will actually be much more efficient if she combines some of the parts and cuts out others. Diane finishes building the product and sends it over to Creative Chris.

Chris meets with Mary to discuss the product’s UI, where Mary explains her ideas. Chris gets to work. He starts to incorporate Mary’s ideas but quickly realizes that he has a much better and more interesting UI concept in mind, and eventually completes the product that way.

Once the project is complete, Mary is upset. The product has neither the functionality she asked for, nor does the UI allow the user to navigate the product properly.

All three departments go back to the drawing board to redo the product, and a significant amount of time and money has been wasted.

This common situation is not one person’s fault. Mary should have explained to Diane and Chris her clear end goal of the product and let them create it with the goal in mind. Diane should have understood that Mary was asking for specific functionalities for a reason and even though her product may have been more efficient, it’s useless if it doesn’t accomplish the desired end goal. Chris should have understood that even though Mary’s UI concept may not have been the most aesthetically pleasing, she had that UI for a reason. They should have worked together to create a concept that works for everyone.

There are a few ideas out there on how to attack common product development miscommunication problems, however, many have found the most successful solution is to create storyboards throughout the product development process.

A storyboard is a collection of cells, either in a linear progression or mapped out from a central idea that tells a story. Each cell can contain an image, a title, and a description that provides specific information to the reader about certain aspects of the story. Storyboards are meant to be simple representations of a larger concept, and force both its creators and readers to break down large complex topics into simple step by step subsections.

The Journey Of Entrepreneur Erin

The fictional, but based on real events journey of an entrepreneur using storyboards to help her and her team design, build, and iterate a her product.

Erin left her job at the big company she worked for and decided to pursue her dreams and start her own company. She already had an idea — SoLoMoFoo. SoLoMoFoo is an application to alert employees when free food is available in common areas — like conference rooms, shared office kitchens, or private offices. At her old job, she had noticed far too often that free food goes to waste due to lack of awareness and employees become disgruntled after hearing about free food that they had just missed. She decided that this problem needed a solution and that she was going to build it! First, she needed to figure out who exactly her target users would be.

A useful way to discover potential target users is to create personas. Erin decided she was going to map out a few of her target users and buyers and record some of their unique characteristics.

Creating Personas
Creating personas can help you step out of yourself. It can help you to recognize that different people have different needs and expectations, and it can also help you to identify with the user you’re designing for.

— Rikke Dam, Co-Founder of Interaction Design Foundation
Personas for SoLoMoFoo (Large preview)

Erin knows that SoLoMoFoo will solve a problem that exists (the lack of awareness about free available food) — but who does this problem exist for? Who will be using her product? Before Erin starts to create storyboards she first has to build her personas. Generally, companies will have to focus on two different types of personas — user personas and buyer personas.

1. User Personas

These are fictional depictions of quintessential users fitting a certain criteria. Most products will try to limit the number of two or a maximum of three key user personas, and then focus the majority of marketing efforts on attracting these users. In SoLoMoFoo’s case, there are two key user personas that Erin has identified:

  • Baking Ben
    Ben is often bringing free food into the workplace to share with coworkers. He feels a little weird about emailing the entire office every time he brings in cupcakes so he would love an app that alerts his coworkers for him.
  • Hangry Hank
    Hank is constantly missing free food and is upset because of it. He feels less productive when he’s hungry and would be very interested in an app that alerts him anytime food is available.
2. Buyer Personas

Often times the intended user of your product is not the same as the intended buyer. In SoLoMoFoo’s business model, an entire company will purchase the SoLoMoFoo app and have their employees download it. This way everyone in the office will both be able to send alerts when they have free food, and receive alerts when they’re looking for food. Erin has decided that the most likely purchaser of SoLoMoFoo will be the HR department.

  • HR Hailey
    Hailey is the HR manager and has purchasing power. She is constantly looking for ways to improve employees’ morale and engagement. She is incentivized by her superiors to inspire energy and teamwork amongst the employees and has a budget to spend on applications or tools that will help her do this.

Creating these personas will help you step inside the shoes of both your users, and your buyers (if they are different). It helps you take a step back from your product and see it through the eyes of the people you are designing for.

To start creating the personas you need, there are several online resources you can use. For example, you can use a persona worksheet template like one of these:

HubSpot persona template (Large preview) Xtensio persona template (Credit to Xtensio) (Large preview)

Or use a persona creation tool like this one from HubSpot.

After personas have been created, the next step is to discover how these personas would come in contact with the problem you're solving for, your product as a solution, and how the product would ultimately benefit their lives. A great way to step into your personas’ shoes is to create journey maps.

Journey Mapping
Customer journey mapping helps you to visualize your customer’s experience from the customer’s point of view, across all the different touchpoints they have with your brand as they seek to achieve a specific goal or goals.

— Tandem Seven Experts

Now that Erin has identified SoLoMoFoo’s key user and buyer personas it’s time for her to discover how these personas may come across her product, how they would use it, and what potential obstacles they may face throughout the process. A great way to do this is to create journey maps in the form of storyboards. Creating these journey mapping storyboards forces the product designer to walk in the user or buyers shoes and experience their product in a step-by-step manner.

Customer journey map for SoLoMoFoo (Large preview)

A journey map storyboard can generally be broken down in six key components:

  1. Problem Experienced
    You have decided to create your product for a reason. You believe that your target users are experiencing a problem that they want solved. What is the problem that your product solves for?
  2. Solution Search
    After your persona has experienced the problem, you believe they will go searching for a solution. What methods will they use to search? These can be potential marketing channels to consider for your go-to-market strategy.
  3. Product Discovery
    During their search, your personas will come across your product and decide to start using it. How will they know this is the product for them? How will they get started? What barriers to entry might they face?
  4. Product Experienced
    The persona will now use the product and experience the intended objective. How do they use it? Can they use it immediately or are there other steps they need to take?
  5. Problem Alleviated
    After the intended product objective is attained the users problem is alleviated. Is this the same problem you were attempting to solve for at the beginning of your storyboard? What other potential problems might stem from your solution?
  6. Beneficial Outcome
    Now that persona’s problem is alleviate, why is their life better? What benefit did solving their problem bring to them and how will this improve their situation?

Need an example? Take a look at the illustration below:

Customer journey map template (Large preview)

In the case of Entrepreneur Erin, she has to consider how the lack of access to free food would affect HR Hailey, how she would research possible solutions, how she would come across SoLoMoFoo, how the SoLoMoFoo platform would be implemented at her place of work, and the potential benefits and timeline for those benefits that the SoLoMoFoo program would bring.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Improve Animated GIF Performance With HTML5 video Improve Animated GIF Performance With HTML5 video Ayo Isaiah 2018-11-05T14:30:14+01:00 2018-11-05T21:23:43+00:00

Animated GIFs have a lot going for them; they’re easy to make and work well enough in literally all browsers. But the GIF format was not originally intended for animation. The original design of the GIF format was to provide a way to compress multiple images inside a single file using a lossless compression algorithm (called LZW compression) which meant they could be downloaded in a reasonably short space of time, even on slow connections.

Later, basic animation capabilities were added which allowed the various images (frames) in the file to be painted with time delays. By default, the series of frames that constitute the animation was displayed only once, stopping after the last frame was shown. Netscape Navigator 2.0 was the first browser to added the ability for animated GIFs to loop, which lead to the rise of animated GIFs as we know them today.

As an animation platform, the GIF format is incredibly limited. Each frame in the animation is restricted to a palette of just 256 colors, and over the years, advances in compression technology has made leading to several improvements the way animations and video files are compressed and used. Unlike proper video formats, the GIF format does not take advantage of any of the new technology meaning that even a few seconds of content can lead to tremendously large file sizes since a lot of repetitive information is stored.

Even if you try to tweak the quality and length of a GIF with a tool like Gifsicle, it can be difficult to cut it down to a reasonable file size. This is the reason why GIF heavy websites like Giphy, Imgur and the likes do not use the actual GIF format, but rather convert it to HTML5 video and serve those to users instead. As the Pinterest Engineering team found, converting animated GIFs to video can decrease load times and improve playback smoothness leading to a more pleasant user experience.

Hence, we’re going to look at some techniques that enable us use HTML5 video as a drop in replacement for animated GIFs. We’ll learn how to convert animated GIFs to video files and examine how to properly embed these video files on the web so that they act just like a GIF would. Finally, we’ll consider a few potential drawbacks that you need to ponder before using this solution.

Convert Animated GIFs To Video

The first step is to convert GIF files to a video format. MP4 is the most widely supported format in browsers with almost 94% of all browsers enjoying support, so that’s a safe default.

94% of all browsers support the MP4 format (Large preview)

Another option is the WebM format which offers high quality videos, often comparable to an MP4, but usually at a reduced file size. However, at this time, browser support is not as widespread so you can’t just go replacing MP4 files with their WebM equivalents.

Internet Explorer and Safari are notable browsers without WebM support (Large preview)

However, because the <video> tag supports multiple <source> files, we can serve WebM videos to browsers that support them while falling back to MP4 everywhere else.

Let’s go ahead and convert an animated GIF to both MP4 and WebM. There are several online tools that can help you do this, but many of them use ffmpeg under the hood so we’ll skip the middle man and just use that instead. ffmpeg is a free and open source command line tool that is designed for the processing of video and audio files. It can also be used to convert an animated GIF to video formats.

To find out if you have ffmpeg on your machine, fire up a terminal and run the ffmpeg command. This should display some diagnostic information, otherwise, you’ll need to install it. Installation instructions for Windows, macOS and Linux can be found on this page. Since we’ll be converting to is WebM, you need to make sure that whatever ffmpeg build you install is compiled with libvpx.

Getting workflow just right ain’t an easy task. So are proper estimates. Or alignment among different departments. That’s why we’ve set up “this-is-how-I-work”-sessions — with smart cookies sharing what works well for them. A part of the Smashing Membership, of course.

Explore Smashing Membership ↬

To follow along with the commands that are included in this article, you can use any animated GIF file lying around on your computer or grab this one which is just over 28MB. Let’s begin by converting a GIF to MP4 in the next section.

Convert GIF To MP4

Open up a terminal instance and navigate to the directory where the test gif is located then run the command below to convert it to an MP4 video file:

ffmpeg -i animated.gif video.mp4

This should output a new video file in the current directory after a few seconds depending on the size of the GIF file you’re converting. The -i flag specifies the path to the input GIF file and the output file is specified afterwards (video.mp4 in this instance). Running this command on my 28MB GIF produces an MP4 file that is just 536KB in size, a 98% reduction in file size with roughly the same visual quality.

But we can go even further than that. ffmpeg has so many options that you can use to regulate the video output even further. One way is to employ an encoding method known as Constant Rate Factor (CRF) to trim the size of the MP4 output even further. Here’s the command you need to run:

ffmpeg -i animated.gif -b:v 0 -crf 25 video.mp4

As you can see, there are a couple of new flags in above command compared to the previous one. -b:v is normally used to limit the output bitrate, but when using CRF mode, it must be set to 0. The -crf flag controls the quality of the video output. It accepts a value between 0 and 51; the lower the value, the higher the video quality and file size.

Running the above command on the test GIF, trims down the video output to just 386KB with no discernable difference in quality. If you want to trim the size even further, you could increase the CRF value. Just keep in mind that higher values will lower the quality of the video file.

Convert GIF To WebM

You can convert your GIF file to WebM by running the command below in the terminal:

ffmpeg -i animated.gif -c vp9 -b:v 0 -crf 41 video.webm

This command is almost the same as the previous one, with the exception of a new -c flag which is used to specify the codec that should be used for this conversion. We are using the vp9 codec which succeeds the vp8 codec.

In addition, I’ve adjusted the CRF value to 41 in this case since CRF values don’t necessarily yield the same quality across video formats. This particular value results in a WebM file that is 16KB smaller than the MP4 with roughly the same visual quality.

Now that we know how to convert animated GIFs to video files, let’s look at how we can imitate their behavior in the browser with the HTML5 <video> tag.

Replace Animated GIFs With Video In The Browser

Making a video act like a GIF on a webpage is not as easy as dropping the file in an <img> tag, but it’s not so difficult either. The major qualities of animated GIFs to keep in mind are as follows:

  • They play automatically
  • They loop continuously
  • They are silent

While you get these qualities by default with GIF files, we can cause a video file to act the exact same way using a handful of attributes. Here’s how you’ll embed a video file to behave like a GIF:

<video autoplay loop muted playsinline src="video.mp4"></video>

This markup instructs the browser to automatically start the video, loop it continuously, play no sound, and play inline without displaying any video controls. This gives the same experience as an animated GIF but with better performance.

To specify more that once source for a video, you can use the <source> element within the <video> tag like this:

<video autoplay loop muted playsinline>
    <source src="video.webm" type="video/webm">
    <source src="video.mp4" type="video/mp4">
</video>

This tells the browser to choose from the provided video files depending on format support. In this case, the WebM video will be downloaded and played if it’s supported, otherwise the MP4 file is used instead.

To make this more robust for older browsers which do not support HTML5 video, you could add some HTML content linking to the original GIF file as a fallback.

<video autoplay loop muted playsinline>
    <source src="video.webm" type="video/webm">
    <source src="video.mp4" type="video/mp4">
      
    Your browser does not support HTML5 video.       
    <a href="/animated.gif">Click here to view original GIF</a>
</video>

Or you could just add the GIF file directly in an <img> tag:

<video autoplay loop muted playsinline>
    <source src="video.webm" type="video/webm">
    <source src="video.mp4" type="video/mp4">
    <img src="animated.gif">
</video>

Now that we’ve examined how to emulate animated GIFs in the browser with HTML5 video, let’s consider a few potential drawbacks to doing so in the next section.

Potential Drawbacks

There are a couple of drawbacks you need to consider before adopting HTML5 video as a GIF replacement. It’s clearly not as convenient as simply uploading a GIF to a page and watch is just work everywhere. You need to encode it first, and it may be difficult to implement an automated solution that works well in all scenarios.

The safest thing would be to convert each GIF manually and check the result of the output to ensure a good balance between visual quality and file size. But on large projects, this may not be practical. In that case, it may be better to look to a service like Cloudinary to do the heavy lifting for you.

Another problem is that unlike images, browsers do not preload video content. Because video files can be of any length, they’re often skipped until the main thread is ready to parse their content. This could delay the loading of a video file by several hundreds of milliseconds.

Additionally, there are quite a few restrictions on autoplaying videos especially on mobile. The muted attribute is actually required for videos to autoplay in Chrome for Android and iOS Safari even if the video does not contain an audio track, and where autoplay is disallowed, the user will only see a blank space where the video should have been. An example is Data Saver mode in Chrome for Android where autoplaying videos will not work even if you set up everything correctly.

To account for any of these scenarios, you should consider setting a placeholder image for the video using the poster attribute so that the video area is still populated with meaningful content if the video does not autoplay for some reason. Also consider using the controls attribute which allows the user to initiate playback even if video autoplay is disallowed.

Wrap Up

By replacing animated GIFs with HTML5 video, we can provide awesome GIF-like experiences without the performance and quality drawbacks associated with GIF files. Doing away with animated GIFs is worth serious consideration especially if your site is GIF-heavy.

There are websites already doing this:

Taking the time to convert the GIF files on your site to video can lead to a massive improvement in page load times. Provided your website is not too complex, it is fairly easy to implement and you can be up and running within a very short amount of time.

(ra,dm)
Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Separate tags by commas
To access this feature, please upgrade your account.
Start your free month
Free Preview