The art of simplicity

In my blog post yesterday I mentioned the usage of the F# XML Type Provider to easily parse and process XML files. However the file I wanted to process(a NuGet nuspec file) had a lot of optional parts meaning that I had to provide a lot of samples to guarantee that the Type Provider understands my XML structure correctly. However on the NuGet GitHub site, I found an XSD file . Wouldn’t it be a lot easier if I could use an XSD file instead? The XML Type Provider doesn’t support this out-of-the-box. Luckily someone created an FSharp.Data.Xsd package that augments the XML Type Provider with schema support. The Schema parameter can be used (instead of Sample ) to specify an XML schema. The value of the parameter can be either the name of a schema file or plain text:

I’m a big fan of the F# type providers. They make my life working with JSON, XML,… a lot easier and less error-prone. Last week however I had a problem when parsing a .nuspec NuGet file. I used the following NuSpec file as the sample file: In this sample, you see in the metadata tags that some dependencies are available. However when I tried to load another specfile without any dependencies the F# XML type provider complained that it couldn’t find any dependency elements inside the metadata tag. So how can I tell the XML type provider that the dependencies are optional? One way to do this is extend the sample file with another metadata tag without dependencies and telling the XML type provider that a list of samples is provided(using  SampleIsList=true ).

Last week a colleagued shared the following error message: NHibernate.Hql.Ast.ANTLR.QuerySyntaxException: Exception of type 'Antlr.Runtime.NoViableAltException' was thrown. The error happened in an ASP.NET Web API controller using the built-in OData functionality in combination with NHibernate. Further investigation showed that the problem occured when one of the following OData filters were used: startswith a combination of ‘and’ and/or ‘or’ conditions Some query url examples that are causing trouble: Products?$filter=startswith(ProductName,’ABC’) Orders?$filter=year(OrderData eq 2016 and OrderType eq 'FastShipping' NHibernate is unable to parse the expression tree generated by ASP.NET Web API. Luckily someone already created a solution for this problem; https://github.com/Pathoschild/webapi.nhibernate-odata/blob/master/README.md Install the NuGet package and apply the attribute to your queryable Web API controller action: [Query...

If you have tests, you can refactor. If you don't, you can only rewrite. Tests are the best tool you have to fight technical debt. Source unknown

I see a lot of organisations struggle to keep good software engineers. Although they pay their engineers well they see their best people leave and they cannot figure out why?! Here are some of the malfunctions I noticed: Respect: Treat your software engineers as overgrown children that play around with “toys”. Don’t involve them in any kind of decision making but micro manage them instead. Involvement: Keep your software engineers as far away from business as possible. It’s not their job to understand business problems, their job is writing code.Don’t involve them in any kind of the non-technical decision-making processess. Throw requirements over the wall to engineering and expect them to implement it. Hardware and Tooling: Provide your software engineers with outdated systems with unsufficient resources(writing code cannot require a powerful machine, right?). Deny any tool that can improve productivity and makes the life of your engineers easier.  Accountabi...

JSON is becoming more and more the new XML. So similar to XML Schema Definitions(XSD) we see schema definitions pop up for JSON data. One tool that can help you generate a schema definition is JSONschema.net . From the website : JsonSchema.net is a tool that automatically generates JSON schema from JSON according to the IETF JSON Schema Internet Draft Version 4. JSON Schema will be automatically generated in three formats: editable, code view, and string. Using this tool, you provide a valid JSON object, click Generate Schema and done! The generated schema can be viewed, edited(temporarily disabled) and copied:

For a long time, the Visual Studio debugger allowed to debug both backend .NET code and client-side JavaScript running in the browser. Unfortunately the only browser were this was supported was Internet Explorer. With the upcoming release of Visual Studio 2017, this will finally change. From now on you can debug your JavaScript and TypeScript running in Chrome  from inside Visual Studio. You only need to select Chrome as your browser inside Visual Studio, hit F5 and a debugger will be attached. Great feature! More information: https://blogs.msdn.microsoft.com/webdev/2016/11/21/client-side-debugging-of-asp-net-projects-in-google-chrome/

A colleague mailed me the following error message when using TypeScript and the ‘import’ syntax to load module: error TS1147: Import declarations in a namespace cannot reference a module. Here is the related code. The error was thrown on the import statement: We found the solution quickly. After moving the ‘import’ statements outside the module, the error disappeared:

Yesterday I blogged about Azure Functions and how the Visual Studio tooling can make your life easier. But the guys at Amazon are not sleeping and I noticed that they added support for C# as well. Note that the AWS Lambda functions in C# are using the .NET Core 1.0 runtime. Similar to Microsoft, you get full tooling support thanks to the AWS Toolkit for Visual Studio . Install the AWS Toolkit for Visual Studio. Open Visual Studio and create a new project. Choose the AWS Lambda Project template and click OK. Next step is to select a Blueprint. We choose the Simple S3 Function blueprint. You get a new project containing the following files aws-lambda-tools-defaults.json: . This file contains default values that the blueprint has set to help prepopulate some of the fields in the deployment wizard. Function.cs: Inside this cs file you can add your own logic. project.json: this is the default project.json for .NET core application...

I blogged about Azure Functions before. It is one of the incarnations of FaaS(Functions as a Service), comparable to AWS Lambda on the Amazon Cloud stack. But Microsoft wouldn’t be Microsoft if they didn’t had plans to improve the development experience. And yes, now we have (a first preview of )Visual Studio Tools for Azure Functions. After downloading the preview , you get the ability to create a function project in Visual Studio, run and test them locally and publish them to Azure. Once you have the tools installed, you can open Visual Studio and choose the Azure Functions project template. Click OK . Inside the created project, you’ll find 2 important files appsettings.json: this file is used to store configuration data for the azure function host.json: contains the global configuration options affecting all functions(more info at https://github.com/Azure/azure-webjobs-sdk-script/wiki/host.json ) Next step is to add one or more azure funct...

On regular intervals, the people from ThoughtWorks release a Technology Radar update . Through this radar they share their thoughts on the technology and trends that are coming and going. On multiple levels(Techniques, Tools, Platforms, Languages & Frameworks) they share if you should adopt, try, assess or move away(hold) from a technology and/or trend. In the november 2016 update I noticed something interesting in the Languages & Frameworks. Where 2015 was the year of Angular.js, I noticed that no Angular could be found in the Adopt/Trial/Assess section. Instead you could find web frameworks and libraries like Ember.js, React.js and even Aurelia. AngularJS could be found in the ‘hold’ section: I’m wondering why they didn’t mention anything about Angular 2…

After publishing my blog post yesterday about type definition files I got a problem when I tried to do the same thing for Modernizr: Visual Studio finds my @types files but complains that it is not a valid module. In fact this makes sense. When using the import syntax, Typescript is looking for a 'module' i.e. a commonjs, amd, systemjs or umd module. Modernizr is added to the global namespace and is not available as a module. To get intellisense, I have to use the ///<reference path=””/> syntax

TypeScript has the concept of Type Definition files to provide TypeScript metadata for libraries that are not written in TypeScript. These TypeScripts files are maintained by the community and available at the DefinitelyTyped repo on GitHub. Before TypeScript 2.0 these type definition files were published on NuGet(Microsofts own package manager). As the TypeScript community kept growing and people outside the Microsoft ecosystem started using it, Microsoft decided to switch to npm as the package source for all type definitions. So starting from TypeScript 2.0 you require no tools apart from npm. Let’s try this! Open a web application in Visual Studio Right click on your project and choose Quick Install Package… from the context menu Note: if you don’t have this option, go install the Package Installer extension first The Quick Install Package window is loaded. Choose NPM from the list of package managers and specify the library for which you want to lo...

One way to block access to a specific folder in your ASP.NET MVC website is by combining the <location> with an <authorization> section inside your web.config: In fact this is not the best approach as it is possible that this configuration is not applied when the ASP.NET pipeline is not invoked. A better approach is to block the access at the IIS level by using the following configuration inside your web.config:

Did you ever hear about the CRAP cycle , the Create/Repair/Abandon/rePlace cycle? You build an application. Over time you accumulate some technical debt. The application becomes harder to maintain. Developers start avoiding and working around certain aspects of the code. Maintenance becomes more and more expensive. Developers complain. New features become harder and harder to write and cost more. Business complain. The application becomes too complex to maintain, we abandon it and start replacing it. Only this time “we are doing it right!”. And of course we make the same mistakes. And the loop starts again… How can we break this circle? What can we do to avoid it? Or is it an unbreakable law of software? I don’t think it has to be. The problem is that most of the time architecture, code quality and a common set of guidelines are only applied at the beginning of a project. Although most applications are built using an iterative approach, the time spent in guarding the qualit...