We were testing iPhora Touch 2 with the attendees of MWLUG 2014 in the MWLUG portal and as usual, the interface had issues even after having outside users provide feedback before we rolled it out. When you are so focused on the creation process it hard to see issues until it is seen by fresh pair of eyes. So we are working on a new faster and improved interface for iPhora Touch. This is the evolution of any product.
Since we are using a JavaScript/JSON Restful API approach most of the changes will be on the front end and only minor changes to the back end which is great. In theory we can swap out different interfaces with little impact on the back end. This is where the newer AMD approach for Dojo comes into play. The use of AMD in Dojo allows us to modularize the UI components even more than what we had. Therefore, we have been upgrading all the iPhora widgets from Dojo 1.53 to Dojo 1.10.0 and at the same time upgrading from Bootstrap 2.32 to 3.2 and the latest version of Fontawesome. The View/Controller of the MVC framework becomes no more than a Dojo widget. This will be the subject of upcoming blogs.
The AMD approach with Dojo has many advantages, but it also has a number of things that I do not like about it. For example below is a simple set of code for reading a cookie and adding it to a DOM node and adding a class to a div tag using Dojo.
HTML
Dojo without AMD
Dojo with AMD it becomes
But like all good developers, I decided to create a new feature with my development tool to use a template to create my code. This is where my mustache function comes in handy. Below is an example of my code template that I have created.
As a result, all I have to do is have my compiler create the Dojo/Bootstrap widgets from the declaration tags, for example <ux:ipField id="hello"/> and assign the appropriate required modules to the assigned mustache tags and do a simple LotusScript find and replace.
However, one critical advantage of AMD is that you can create your own module and include or replace an existing module relatively easily. For example, we needed to replace the Dojo MVC JsonRest module "dojo/store/JsonRest" with our own "iphora/JsonRest" since it does not meet our requirement in how we are doing REST services. We copied the JsonRest.js code modified it and then placed into our iphora directory and now it is loading our module.
So we just point the path to our module. No other changes in the code are required.
Another disadvantage with the AMD module approach is that you need to be extremely careful if you are trying to create a compressed build of your own modules to reduce the number of http request during the loading process.
With Dojo without AMD, you can easily just concatenate all the modules into one single script file and load that script as part of the page instantiation process. However, Dojo with AMD requires you to create a build using an elaborate process that utilizes a package.json. This is similar to how npm in nodejs does their builds. More information can be found in this article.
http://dojotoolkit.org/documentation/tutorials/1.8/build/
Dojo with AMD is more powerful then ever but it just takes a little more work and can be fun!
Since we are using a JavaScript/JSON Restful API approach most of the changes will be on the front end and only minor changes to the back end which is great. In theory we can swap out different interfaces with little impact on the back end. This is where the newer AMD approach for Dojo comes into play. The use of AMD in Dojo allows us to modularize the UI components even more than what we had. Therefore, we have been upgrading all the iPhora widgets from Dojo 1.53 to Dojo 1.10.0 and at the same time upgrading from Bootstrap 2.32 to 3.2 and the latest version of Fontawesome. The View/Controller of the MVC framework becomes no more than a Dojo widget. This will be the subject of upcoming blogs.
The AMD approach with Dojo has many advantages, but it also has a number of things that I do not like about it. For example below is a simple set of code for reading a cookie and adding it to a DOM node and adding a class to a div tag using Dojo.
HTML
<div class="data-display"></div>
Dojo without AMD
dojo.require('dojo.cookie') var c = dojo.cookie('info'); var node=dojo.query('.data-display')[0]; node.innerHTML=c; dojo.addClass(node,'show');
Dojo with AMD it becomes
require([ "dojo/cookie", "dojo/query", "dojo/dom-class" ],function(cookie,query,domClass){ var c=cookie('info'); var node=query('.data-display')[0]; node.innerHTML=c; domClass.add(node,'show'); });As you notice the amount of code for Dojo with AMD can quickly add up and if you miss a module path/declaration you get an error that won't make sense. So you need to be careful in defining the path/declaration pair.
But like all good developers, I decided to create a new feature with my development tool to use a template to create my code. This is where my mustache function comes in handy. Below is an example of my code template that I have created.
As a result, all I have to do is have my compiler create the Dojo/Bootstrap widgets from the declaration tags, for example <ux:ipField id="hello"/> and assign the appropriate required modules to the assigned mustache tags and do a simple LotusScript find and replace.
However, one critical advantage of AMD is that you can create your own module and include or replace an existing module relatively easily. For example, we needed to replace the Dojo MVC JsonRest module "dojo/store/JsonRest" with our own "iphora/JsonRest" since it does not meet our requirement in how we are doing REST services. We copied the JsonRest.js code modified it and then placed into our iphora directory and now it is loading our module.
require([ "dojo/cookie", "dojo/query", "dojo/dom-class", "dojo/store/JsonRest" ],function(cookie,query,domClass,jsonRest){ var c=cookie.get('info'); var node=query('.data-display')[0]; node.innerHTML=c; domClass.add(node,'show'); }); require([ "dojo/cookie", "dojo/query", "dojo/dom-class", "iphora/JsonRest" ],function(cookie,query,domClass,jsonRest){ var c=cookie.get('info'); var node=query('.data-display')[0]; node.innerHTML=c; domClass.add(node,'show'); });
So we just point the path to our module. No other changes in the code are required.
Another disadvantage with the AMD module approach is that you need to be extremely careful if you are trying to create a compressed build of your own modules to reduce the number of http request during the loading process.
With Dojo without AMD, you can easily just concatenate all the modules into one single script file and load that script as part of the page instantiation process. However, Dojo with AMD requires you to create a build using an elaborate process that utilizes a package.json. This is similar to how npm in nodejs does their builds. More information can be found in this article.
http://dojotoolkit.org/documentation/tutorials/1.8/build/
Dojo with AMD is more powerful then ever but it just takes a little more work and can be fun!
Comments