Another one of my talks was selected for Techbash: Alternative Device Interfaces and Machine Learning.
In this presentation, we will look at the how users interface with machines without the use of touch. These different types of interaction have their benefits and pitfalls. To showcase the power of these user interactions we will explore: Voice commands with mobile applications, Speech Recognition, and Computer Vision. After this presentation, attendees will have the knowledge to create applications that can utilize voice, video, and machine learning.
Users use voice (Alexa, Cortana, Google Now) or video as a mode of interaction with applications. More than a fad, this is a natural interface for users and is becoming more and more common with the ever-decreasing size of hardware.
Different types of interaction have their benefits and pitfalls. To showcase the power of these user interactions we will explore: Voice commands with two app types: UWP and Xamarin Forms (iOS and Android). Speech Recognition with Cognitive Services: Verifying the speaker with Speaker Recognition API. Computer Vision with Cognitive Services: Verifying a user with Face API.
By utilizing UWP, Xamarin, and Cognitive services; a device with the ultimate in customization for user interactions will be created. Come and see how!
This year I will be presenting Enable IoT with Edge Computing and Machine Learning at TechBash. Here is the outline:
Being able to run compute cycles on local hardware is a practice predating silicon circuits. Mobile and Web technology has pushed computation away from local hardware and onto remote servers. As prices in the cloud have decreased, more and more of the remote servers have moved there. This technology cycle is coming full circle with pushing the computation that would be done in the cloud down to the client. The catalyst for the cycle completing is latency and cost. Running computations on local hardware softens the load in the cloud and reduces overall cost and architectural complexity.
The difference now is how the computational logic is sent to the device. As of now, we rely on app stores and browsers to deliver the logic the client will use. Delivery mechanisms are evolving into writing code once and having the ability to run that logic in the cloud and push that logic to the client through your application and have that logic run on the device. In this presentation, we will look at how to accomplish this with existing Azure technologies and how to prepare for upcoming technologies to run these workloads.
This year I will be at Music City Code presenting Configure, Control, and Manage IoT with Mobile.
Configure, Control, and Manage IoT with Mobile
The internet of things allows for communication with devices through various means (without touch, mouse, keyboard, or a screen). Mobile devices give users a dynamic interactive experience with these devices by communicating over several different wireless protocols or through the cloud. In this presentation, we will see how to use Xamarin to create a cross platform mobile application to control devices of all shapes and sizes. After this presentation, attendees should be able to create a basic mobile application and have that application communicate with peripherals over Bluetooth and the cloud.
This presentation is to showcase creating mobile applications with Xamarin and how those applications can interact with both off the shelf and with custom hardware. First, we will create a Xamarin Forms application; for iOS, Android, and Windows; that will interact with both Microsoft Azure and Bluetooth Low Energy to create an interactive experience with the hardware and the cloud. To get a better understanding, we will discuss mobile communication with the cloud and hardware to get a picture of how mobile can act as a bridge between the two.
About four days ago Lamar and I went to Mixed3D to get some scans done for his birthday. They have an interesting booth filled with Raspberry Pis that generate a 3D model. They also offer to print those 3D models in full color using sandstone (which we took them up on). From the scans we picked 2 that best fit the theme of the day and were most usable in Lamar’s Unity projects.
Also, Adam got one done to use around the office
After nearly a year of hard work, the Home Control Flex application has finally reached a new release point. There have been major improvements around the use of Xamarin Forms and the use of mobile features. There were major changes around framework dependencies and utilization of navigation pages.
The major problems in the previous version was poor usage of navigation pages, dependencies on old frameworks, and lack of sharing of global resources. Adding all of these failures together resulted in an unstable application that crashed on multiple pages. Fixes to those crashes were a slow roll out of shims and hacks to keep the previous decisions working.
The largest problem was the poor usage of navigation pages. For some reason, to implement a tabbed page where the tabs were at the bottom on Android, the previous developers decided to use a ContentPage, and make the tab pages within the ContentPage ContentViews and swap those views out whenever a tab was changed. This caused almost every major problem that could not be resolved in the app moving forward. To fix it, the BottomNavigationBarXF Nuget packages was used. The base renderer was overridden to implement some custom functionality but overall it was a clean integration or at least as clean as such a big overhaul to the navigation system can handle.
Since the pages were being swapped out whenever a tab was changed, the previous developers mush have decided that instead of needing navigation pages, they would just continue to change the view out and have their own navigation stack. Without using NavigationPage within their app, the page lifecycle was completely off and the were object disposed exceptions that were being thrown by the Forms framework due to the fact that the views lifecycle was not correctly managed. Xamarin Forms couldn’t track whether a view was to be reused or not and would collect on disappeared views that were going to come back later. Once NavigationPage was used this was no longer a problem.
This app was riddled with copy and paste code reuse. Every page shared the same Style declaration with the same name (which was the style for that page). Every page had a declaration of a Converter for inverting a boolean. All these “shared” resources were moved to the App.XAML for reuse by every page within the application.
After fixing the above issues, changing a variety of pages within the app, and adding a load of new features, the app should finally be a stable release with market effects that was expected out of its first release. I hope to continue to improve on the line of applications from Telular including this app.