Occam's Machete

Last time we created a simple HTTP proxy, that supports CORS , so we can access the Xero API directly from a browser. This time we'll be navigating the Xero OAuth labyrinth to make sure we have the required API access. We'll start by registering our website with Xero and follow by implementing the RequestToken-Login-AccessToken process. The OAuth Labyrinth Our test web page will contain a single button to initiate the process. Once clicked, the browser will jump to a Xero page whereby a user can allow access. Once we allow access, the Xero page will jump back to our test page and our test page will show a simple message. Behind the scenes, the process is a little more involved. The steps are shown below. Xero Setup Now, before any of this can happen, we must create a Xero account and register our "App" by supplying an App name, application URL, and callback domain. The App name is arbitrary, however, the Application URL must point to our previously create...

RPC frameworks ( Java RMI , WCF , DCOM , CORBA , SOAP , etc) allow us to call functions over a network as if the function existed locally. Generally, the framework creates a local "proxy" (or "stub") version of the function that hides the networking details for us. Sadly, this isn't usually the case if the client is JavaScript, so here I present a simple method to generate a JavaScript client proxy for C# MVC / Web API functions. What's the point of all this? Well, the idea is that you can add, delete, or change Web API methods and begin using them immediately, in almost the same form, from your browser. An Example Let's say you have the following C# Web API function: public int Some_Fn(int a, int b) { return a+b; } If, like me, you prefer RPC-style calls because you're less than impressed with REST then you'd like your JavaScript calls to look something like this: Some_Fn(a, b, function(res) { ... }); Th...

The Xero API doesn't support CORS so it's impossible to access it from a Web browser. Or is it? One solution is to use an existing CORS proxy service . However, for something so simple we could just build our own and avoid the additional dependency . In this post, I'll present a simple Java proxy, followed by another post with example Javascript code to navigate the OAuth 1 labyrinth, and lastly some sample API calls. It's worth pointing out that if you were hoping for an easy plug-and-play project, you'll be out of luck. I'm a firm believer in building your own stuff , if possible. The opposite almost always leads to bloated, slow, brittle, over-engineered rubbish. Enough talk. Let's get started. Our Options The generally expected architecture for making Xero calls is via a backend server (see Figure 1) using something like C#, Java, Node.js, etc. That's all fine and dandy if you love MVC, server-side rendering, listening to Vanilla Ice on your Walk...

In this post I'll be providing a simple example of an Android Activity with a form-like layout that's built dynamically and with no XML. The official Android developer guides will have you building XML layouts for anything View or Activity related but I'm not a fan of XML. I consider it a failed technology. A lot of promise, nice idea, but in reality not really worth the effort. Kinda like Communism. Don't believe me? Then why the push away from SOAP to alternatives such as JSON? Remember XSLT? Hardly mentioned anymore. XML databases? Again, yesterday's flavour. XML configuration files? Not so popular now that we have attributes and annotations. XML's bloated and complicated text-based structure means it takes far more processing power then should be needed. Especially when there are far more efficient binary-based alternatives. Original Image Source: SiegfriedSassoonGraveMells(GrahamAllard)May2006 " by Graham Allard. Licensed under CC BY-SA 2.0 v...

Whilst perusing images of Asian interior decorating I became enthralled with the interplay of light and shadow in one particular photo of a bathroom with gold tiling. I've yet to achieve the exact effect with HTML but I did manage to produce the much simpler shadows. You can see the effect here . I've only tested the page on Chrome for Windows or Android. No guarantees if you're using another browser or operating system. If you're not able to see the effect then the idea is that an imaginary light source, in this case the mouse, moves across the page and accordingly changes the shadows cast by various elements. Elements nearby will cast sharp and dark shadows. Elements far away will cast wide and lighter shadows. The positions of the shadows also differ depending on their relative position to the light source. Although the effect is simple, I believe that placed within the context of a more dynamic and artistic page, it has the potential to add a nice touch of reali...

Most Android documentation regarding background processing and updates to the UI will have you creating separate threads but this is not strictly necessary. Some of you with experience on other platforms might recall "timer" style mechanisms that allow one to regularly execute code on the main thread and, as long as it's kept brief, is usually imperceptible to the user. In JavaScript we have the "setTimeout" and "setInterval" methods with which you can specify a function to be executed once in the future or repeatedly. In Windows we can use the "SetTimer" function to periodically send timer events to the application's message queue or we can also use a "Timer" class if using the .Net platform. This is possible because most modern GUIs operate using an Event Driven model whereby the main process sits around doing nothing until the user interacts with the UI. Mouse, keyboard, or touch "messages" are then popped on...

Following on from the previous post of my series " Build Your Own ORM ", we will be adding an "insert" function that, when given any object, will create an appropriate SQL Insert command and insert the object's data into the appropriate table. Some Changes First Before we begin let's go over a couple of refactorings that will make the existing code easier to work with. We will be adding two functions. One to create comma delimited strings and the second to build an OdbcCommand object along with associate SQL parameters. The Append_Str function takes a string with potentially existing comma delimited data like "one,two,three" and appends a given string inserting a comma if appropriate. Alternative approaches are available if you don't mind additional code dependencies and/or want something faster or more efficient. 1: public static string Append_Str(string orig_str, string new_str) 2: { 3: string res = orig_str; 4: 5:...