Impressions from Microsoft Build – day 1

This year we are three developers representing Jayway at Microsofts Build conference. This blog-post summarizes just some of the stuff we’ve seen on day one of the conference.

Cross-Platform Mobile with Cordova and Ionic Framework

Demonstration of how to build a conference schedule application using Cordova and Ionic framework, which uses Angular 2 and Typescript.

Looks pretty powerful for applications that can be built using standard UI components available on all platforms. Quite similar to Xamarin Forms except with html/css/javascript instead of xaml/c#.

Also a brief demonstration of Code Push ( integration with Visual Studio Team services, which allows for live-pushing code updates onto Cordova client applications  in production by the means of a simple git push.

Instantly Releasing Updates to Your React Native Apps

Demonstration of Code Push ( command line interface, which allows live code updates in a React for Native cross-platform mobile application.

Command line interface gives control over staging/production releases, statistics on client deployments and automatic code rollback in case an updates crashes the app.

Universal App Model Overview: What’s New in the UWP App Model

An overview of changes coming to the packaging and deployment story of UWP apps in the Windows 10 Anniversary update. Investments are being made in the AppX packaging model to make it an option for more types of apps, and to position it as a successor of MSI installers. Not just for apps distributed through the Store, but also for sideloaded LOB apps.

The benefits are obvious. AppX has a cleaner install and uninstall experience. Registry access for apps distributed this way is being virtualized to a separate file to avoid bloating the main Registry.

Microsoft is working with MSI tool vendors to add support for outputting AppX packages too, and is providing a tool to convert MSI to AppX.

AppX packages can now be installed by double clicking on them, so you no longer need PowerShell.

These changes are a key enabler for the desktop bridge (Project Centennial) that was mentioned during the keynote. Packaging of desktop apps was demoed by creating packages for Paint.NET and even an old VB6 application.

The Future of Visual Studio

Information about new features coming in Visual Studio 2015 Update 2, released today, and in the next major release Visual Studio “15” which is out as a preview.

New much faster installation experience. The presenter did an update of her Visual Studio installation live in the presentation.

Visual Studio has or will get tools for any kind of app in any language on any platform.

Demo of awesome editor and debugging support for Python in Visual Studio.

Same thing for a c++ application for Linux. Remote deployment and debugging of Linux application directly in Visual Studio.

More code style rules have been added for C#, for example casing rules and whether to prefer “var” or explicit types.

Demo of a Bing developer assistant feature, where you can enter a “how do I” type of question and get code back as an answer without leaving Visual Studio.

For open source NuGet packages, VS 15 is able to pull down the source for you, enabling things like Find all references and Go to definition on referenced packages.

A Lap Around ChakraCore – Microsoft’s Open Source JavaScript Engine

ChakraCore outperforms competing JavaScript engines according to (selected) benchmarks.

By embedding ChakraCore in an OpenGL game, one can expose native game APIs to javascript and e.g. allow users to write mods to the game with plain JavaScript code.

By running Node.js in ChakraCore one can enable time travel debugging. This allows you to not only step forward from a breakpoint but also step back! One can also replay code execution based on log-files from production to debug problems.

What’s New in TypeScript?

Mostly just a demo of what TypeScript is.

React/JSX support in TypeScript so one can write react components as .tsx files with full typescript support.

Visual Studio Code supports debugging of TypeScript e.g. in Node applications.

TypeScript on JavaScript? Yes really. Using TypeScript type definitions you can get statement completion and proper intellisense in plain javascript code in Visual Studio Code. Useful e.g. when using any JavaScript library like jquery and lodash, which has TypeScript type definitions available.

Interesting new stuff coming along e.g. non-nullable types, control flow based type checking and type definitions packaged and served through npm.

Building a conversational bot: From 0-60

Introduction for creating bots. First some theories on how to make a bot great. It’s a lot about making the user feel empowered, and working in the right context.

After that the session had a bunch of live coding. Show bot the Bot Builder SDK  (in C#), where it’s really a quite simple service just responding to a message with a message. Using LUIS (Language understanding intelligent service …or something), you can process natural language. You can also use the bot connector from microsoft to act as an intergration point between your bot and services like Skype and Slack. Overall this sessions was great – tons of code.

Microsoft Edge: What’s Next for Microsoft’s New Browser and Web Platform

This was a really uninspiring session. Mostly about what web standards Edge supports / is going to support. It also contained some talks about how Edge is improving and growing, but the metrics selected felt hand picked to make Edge shine. The only really nice takeaway was the Edge is bringing proper JS-extensions, but that part was way too short. JS-extensions should come with the anniversary update though, so look forward to that.

Windows Store and Dev Center Overview: New Capabilities for Helping Developers Succeed

It’s good to see that improvements are being made to the Windows Store and Dev Center, which have been lagging behind competitors like the iTunes Connect on many areas.

Microsoft is finally releasing a universal app for keeping track of your store apps, so you don’t have to log on to the Dev Center website. They are also providing and SDK with APIs for fetching information about your store apps.

A new Dev Center Insider Program is being launched, and if you opt in to it you will get to try out new features earlier, just like with the Windows 10 Insider Program. Invitations will go out to all developer accounts shortly.

The ability to do prereleases to a subset of your customers that was available for Windows Phone 8 is coming back. You can also do percentage based releases, to gradually release an updated version. Say for example that you want to start by just releasing an update to 10% of your users.

Rolling releases will allow an app to be published at a certain local time around the globe. You might want to release an app at midnight a certain date, but of course midnight is a different point in time in different time zones.

Another new feature is that you will be able to reply to store reviews. Either privately through e-mail or publicly so it’s visible to everyone.

Microsoft is adding the Feedback hub in the Anniversary update of Windows 10, and making it available for use in our own apps. There are APIs to open a feedback form and let users supply feedback or issues, much like the Insiders hub available to Window Insider Program members.

The Skype Experience: The Future of Communications

An introduction to creating a skype bot. Very similar to the “conversational bot” talk. First it was a high level overview of the architecture.

Interesting projects


Push code updates to apps instantly for JavaScript based apps on all platforms = Cordova and React Native.

Microsoft Cognitive Services

APIs to add intelligence and natural interactions to apps across all platforms. E.g. build models to parse natural language queries into datamodels, recognize emotions from facial expressions, recognise content of images and describe it in plain text or convert speech to text and vice versa.

An example of what one can build is a service, that will describe the content of an image using plain text. It can recognize object, people, emotions, gender age etc. You can try it out here:

Bot Framework

Easily build and deploy bots that can communicate with users via a multitude of channels such as mail, sms, Skype, Slack, web or many other channels. Seems really powerfull when combined with cognitive services such as speech recognition and natural language understanding.


An open source project from Telerik to build cross-platform native applications using Javascript. Idea seems similar to React Native except it can use Angular instead of React.

Leave a Reply