.net rocks! by Carl Franklin

How do you build quality software with LLMs? Carl and Richard talk to Den Delimarsky about the GitHub Spec Kit, which uses specifications to help LLMs generate code for you. Den discusses the iterative process of refining specifications to produce better code, and then being able to add your own code without disrupting the process. The conversation delves into this new style of software development, utilizing specifications to break down tasks sufficiently for LLMs to be successful, and explores the limitations that exist today.

What's coming for GitHub? Carl and Richard talk to April Yoho about the recent announcements from GitHub Universe and how they will roll out in 2025. The biggest topic, of course, is all the large language models coming to GitHub - there are a bunch of copilots! April talks about original GitHub Copilot, Copilot Workspace, and Copilot Chat - so many options! Now, you can choose your language model to move beyond OpenAI. And there are other changes at GitHub, including EU residency, new features in the enterprise cloud, and new instrumentation - 2025 looks awesome!

GitHub is about so much more than source control! While at Build, Carl and Richard chatted with Damian Brady about all the great things coming out of GitHub over the past few years, including Copilot and now Copilot Workspaces. The conversation digs into how large language models are changing how you write code and some idea of what things could be like in the future. Damian also digs into the more enterprise-centric features like Software Bill of Materials and effective sponsorship of projects. GitHub is the locus of development for many companies, and the tooling continues to expand to make things easier!

Do you understand how your APIs are being used? Carl and Richard talk to Anthony Alaribe about his experiences dealing with poorly documented APIs that need updates - but no breaking changes! Anthony tells a story about missing a use case for an API that cost a lot of money, which started him down the path to making APItoolkit.io. The toolkit allows you to see how your API is being used and any exceptions that are happening. It will also generate tests to validate that your new version won't cause problems! Check it out!

Azure and GitHub - better together? While at the Copenhagen Developer Festival, Carl and Richard talked to April Edwards for a special .NET Rocks Live. April talked about how Azure and GitHub work well together, discussing Azure DevOps and GitHub Actions on the CI/CD pipeline side and how other services can interact. Lots of laughter and great questions from the live audience!

How do you secure your existing applications from the security exploits out there today? Carl and Richard talk to Joylynn Kirui about the challenges developers face in making secure applications. Joylynn talks about understanding the threat landscape and staying up to date on the CVEs that can represent a zero-day vulnerability to your application. There are a ton of tools to help make you aware of the potential risks, check out all the links in the show notes. And check out Joylynn's webinar on shifting application security left at https://aka.ms/DevSecOpsDNR

Why should .NET developers build Power Apps? Carl and Richard talk to April Dunnam about the latest in Microsoft's Power Platform and why .NET developers should get involved. April talks about the fusion development methodology, where domain experts use the Power Platform tools alongside .NET developers using Visual Studio to build line-of-business apps. There are many opportunities in the cycle of building Power Apps where your skills as a .NET developer can make the process go faster, more reliable, and with more features! Also, check out April's upcoming workshops on building Power Apps online at https://aka.ms/PowerDNR and at the DevIntersection conference in Las Vegas in December!

How has the cloud changed data analytics? Carl and Richard chat with Vishwas Lele about his latest work taking a developer's view of data analytics - without upsetting the DBAs too much! Vishwas talks about how the cloud has changed bringing disparate data sources together for analytics. With the cloud's compute-on-demand, you don't need to do many transformations of data as it's loaded - but you can test it! This leads to a conversation of how CI/CD techniques can be applied to data to make for accurate data analytics - make your ingestion pipeline smart!

Bring the fight to the hackers with some clever code! Carl and Richard talk to Dana Epp about honeytokens - adding code and elements to your applications that are there only to attract bad actors. Dana talks about how hackers attack applications, looking for vulnerabilities. Often those attempts take weeks or even months and are hard to detect in regular logs. By adding code that would only run if an attacker was trying to exploit, you can raise a red flag to your security team early and take action before the attackers are successful. Have a listen, then chat (preferably over pizza) with your security folks about how you can help!

Are you using open-source libraries in your enterprise applications or products? What are the consequences? Carl and Richard talk to Rocky Lhotka about his work with organizations coming to grips with the use of open-source code in their projects. What happens if the project goes away? What if there is malware added to the code intentionally or accidentally? What about plagiarism of code? Diligence around the origin of code is important, whether it is open-source or not - and should be part of your build process!

What's new in Visual Studio 2022? Carl Franklin and Richard Campbell talk to Simon Calvert, Anthony Cangialosi, and Andy Sterland about the upcoming version of Studio. First on the docket is 64-bit: While Studio has been able to build 64-bit applications for more than a decade, the devenv.exe itself has remained a 32-bit application. The team discusses why it makes sense to move to 64-bit and what that means for development going forward. But there's much more to Studio 2022, including improvement to debugging and diagnostics across platforms, being able to collaborate with developers in real-time using Live Share, and getting smarter about how you navigate and explore existing code. Cool new things are coming!

Can you make open source projects in the Microsoft ecosystem and live to tell the tale? Carl and Richard talk to Aaron Stannard about his view of the Microsoft open-source ecosystem, including the challenges of building and maintaining Akka.NET. Aaron talks about Microsoft being good producers of open source but not good consumers of open source - they still have that not-invented-here viewpoint on things. The conversation turns to either making an open-source project that is popular but ultimately consumed by Microsoft, or profitable because it lives in the long tail that Microsoft won't explore. But could there be a better way?

Sure you can build .NET apps on Azure - but what about Amazon AWS? Carl and Richard talk to Robb Schiefer about his work building .NET Apps on AWS. Robb digs into the various reasons you might use AWS, while admitting that the two clouds are pretty comparable. But like Azure, AWS has good integration into Visual Studio, supports lots of versions of .NET and offers other cloud-based services that you need. AWS is an excellent option for .NET apps!

Are there Silver Bullets in Development? Carl and Richard chat with Mark Seemann about the 1986 Fred Brooks essay No Silver Bullet. Mark argues that there are silver bullets - but do we notice them? The conversation dives into technology like the Internet (maybe you've heard of it?) changing the way we do software development: Providing rapid access to resources, answering your questions, and simplifying deployment. Mark also talks about automated testing to shorten cycle times, how modern distributed source control can help you work faster and more!

Windows represents the single largest Git source control library in the world at 300GB - but what does it take to work on it? Carl and Richard talk to Ed Thomson and Jill Campbell about how Azure DevOps (formerly known as VSTS) functions under the load of 33,000 people working on the Windows project with 11 million work items. There are many things in Azure DevOps that can cope with that scale, but some aspects don't make sense to add directly, like moving millions of work items. For that, the team has built extensions available in the Visual Studio marketplace - check them out, maybe they can help you!

How should developers get started with Azure? Carl and Richard talk to Adam Cogan about his nine knights of Azure - in order, a set of services in Azure that developers should start with. Right off the bat, Adam suggests staying clear of the fancy cloud techs like containers and serverless. Start simple with Azure Web Apps, then start working on continuous integration and deployment. There are a ton of moving parts in Azure, and Adam has a great plan for getting your application and your skills gradually up to speed so that you can take them all on!

Microsoft acquires GitHub! What does this mean? Carl and Richard talk to Phil Haack about what's going to change and what isn't - starting with, at least for the foreseeable future, Phil is not a Microsoft employee! GitHub is going to be a wholly owned subsidiary of Microsoft, operated independently. But having Microsoft resources available does open some interesting doors - Phil talks about the vast amount of resources that Microsoft has to move quickly on cool features and projects around the open source world!

What does it take to implement a DevOps practice at Microsoft? While at Connect in New York, Carl and Richard talked to Lori Lamkin about the efforts of the VSTS team to deliver more value to customers faster. Lori talks about changing the tempo of deployment for TFS from 24 months to three weeks and the re-organization of the teams so that developers are responsible for features from conception to operations. The role of operations also evolves to one that focuses on monitoring, rather than deployment - it's a lot of change, but worth it!

Can everyone benefit from DevOps practices? While at Ignite in Orlando, Carl and Richard sat down with Donovan Brown and Damian Brady to talk about the recently formed League of Extraordinary DevOps Advocates and the effort to get both developers and operations folks taking advantage of Continuous Integration, Continuous Delivery, Instrumentation tooling and more. While the key to a good DevOps practice is the culture and process, having great tools makes everything easier, and today VSTS and TFS offer one-button delivery pipelines. It's never been easier for everyone to get onboard with a DevOps practice!

How does Microsoft make Visual Studio Team Services? With Visual Studio Team Services! While at Ignite in Orlando, Carl and Richard sat down with Dan Hellem and Rogan Ferguson to talk about their work in the VSTS team building and deploying features on a three-week cycle. The discussion dives into the agile practices that the different teams use and how they come together to get new versions out on a regular basis as part of the Software-as-a-Service offering of VSTS. Great conversation about keeping developers productive!

Git for everyone! While at Ignite in Orlando, Carl and Richard talked to Ed Thomson and Dmitry Lyalin about the latest updates to Visual Studio Team Services that includes deep integration with Git - not just GitHub, but Git the repository technology. The conversation digs into the commitment Microsoft has to contributing to open source projects including Git itself - including moving Windows into the world's largest Git repository, a whole 300GB of source code! Git has penetrated deeply into Microsoft, is it the right choice for your projects?

The whirlwind of DevOps is back! Carl and Richard talk to Donovan Brown about his on-going mission to get everyone into a DevOps practice - in fact, he talks about getting to a place where continuous delivery and monitoring are just the way you make software. Donovan talks about a session he has been giving where he creates four different delivery pipelines in an hour, with different programming languages, platforms, testing and deployment tools. And if he can make four in an hour, there's really no reason anyone can't build one pipeline in an afternoon. The conversation ranges over a spectrum of DevOps topics, including testing and monitoring - it's easier than ever to rub some DevOps on it!

How does Microsoft maintain mature applications? Carl and Richard talk to Dustin Metzgar about his work at Microsoft maintaining applications and libraries like Windows Workflow, older versions of ASP.NET and Entity Framework and more. These products are maintained for a long time, typically without adding features, but rather to make sure new operating systems still work with them, security is maintained and bugs are fixed. There's a ton of cool stories in this space, it's the ultimate brownfield project challenge - rarely, if ever, do folks who created a product continue with it throughout it's life time!

Are you adding feature toggles to your apps? Carl and Richard talk to Daniel Piessens about his approach to adding feature toggles, using frameworks to keep things organized. The conversation starts out talking about different kinds of features toggles, starting with the classic one that allows you to build features over time, but keep the code in the trunk, just not visible to the users until you're ready. In some cases, that feature toggle because permanent because it is a tool for ops to reduce load on a server at peak times. Toggles are also a strategy for A/B testing of different features, styling and advertising - you need to dig into this!

What does it take to make an application support multiple customers? Carl and Richard talk to Paul Stovell about his work making multi-tenant applications. As with most things, making multi-tenant apps is more complicated than it seems! Paul talks about making architectural decisions around separation between various customers - do they each get their own database? What about web server and/or app-pool? What about customizations and deployment. Do customers get new features immediately, or do they have the option to wait? How does the cloud impact your decision making? It's a complicated subject with a variety of trade-offs!

Here comes Windows Containers! While at Techorama in Belgium, Carl and Richard talk to Els Putzeys about the current technical preview of Windows Containers on Windows Server 2016. Els discusses how the Windows Containers can be built using templates so that you have a configuration-as-code capability, making repeated creation fast and easy. Now your documentation for building a system is code that actually creates it! We're still a few months away from Server 2016 being released, but it looks to be an important version of server for development!

MSDeploy can do it! Carl and Richard talk to Robert Schiefer about his experiences using MSDeploy to automate complex deployment solutions throughout the enterprise. Robert talks about the fact that MSDeploy is part of WebDeploy and includes a ton of features that are not always obvious - doing far more than just installing software, it can also modify security settings, update registry keys, even handle the deployment of databases using DACPAC. The documentation isn't great, but if you dig around, there's lots of capability there. If you're looking to save money, spend a little time understanding MSDeploy better!

Can you continuously deliver software on Azure? Carl and Richard talk to Jeffrey Palermo about his approach (and toolkit) to do continuous delivery - and Azure makes it far more feasible! The conversation starts out with some definitions, recognizing that the development community has been working its way toward faster delivery of value to customers for a long time - it's part of the Agile Manifesto. The question is purely how fast, or rather, how continuous is continuous? Jeff also notes that delivery doesn't mean deployment, but it does mean that you're at the place where you can immediately deploy when you're ready. Lots of tools and thinking on how to get this done, it isn't simple!

How do you manage your software releases? Carl and Richard talk to Kevin Mack and Brandon Rohrer about the various tools that come together to help automate software releases. The focus on this approach is a product formerly known as InRelease by InCycle, which was acquired by Microsoft back in 2013. Now it's known as Release Management Services for VSTS and deeply integrated with the Studio toolset. The conversation also dives into the challenges around databases when it comes to release management - how do you get your database to be part of your release? Lots to think about here to up your VSTS game!

Do you have a DevOps stack? Carl and Richard talk to Brian Randell about his experiences implementing tooling for DevOps practices using Visual Studio and other Microsoft tools. The conversation runs the gambit of elements in the DevOps stack, including continuous deployment, package management, configuration-as-a-code and instrumentation in production. The more automation the better when it comes to moving fast and reliably. While DevOps often focuses on the culture and process that goes into building software, great tools can make all of those things happen much more easily - you just have to figure out what stack is right for you!

Do you understand how to develop software with GitHub? Carl and Richard walk through the fundamentals of GitHub with Bill Wagner. Bill walks through the key differences between the distributed source control that is GitHub versus the more classic central control style source control systems. Often it is harder for experienced developers to deal with the reality of GitHub than someone new to source control! The conversation explores all the concepts of GitHub - repositories, cloning, forking, push and pull requests, merging and rebasing. There's a lot to know, but there are great tools for getting started. And at the end, a quick discussion about Humanitarian Toolbox and their awesome open source projects to help save lives - you can help!

Ci with SQL Server? Are you crazy? Carl and Richard talk to Ike Ellis about what it takes to get databases changes happening as smoothly as application changes. Really! Ike talks about the concept of database lifecycle management and how it is orthogonal to application lifecycle management. The tricky bit is managing the data! And to help with that, Ike discusses a great tool chain of source code management, testing and deployment tools that work with databases like SQL Server and integrate into your usual development lifecycle toolchain as well. This is some next level DevOps thinking!

Unifying your Git GUI experience! Carl and Richard talk to Amy Palamountain from GitHub about how GitHub Desktop is bringing together the separate GitHub for Mac and GitHub for Windows products. Amy explains that in the past, there were separate GUI clients for Mac and Windows, with separate feature sets and styles. Bringing the codebase together keeps the releases more consistent, but it also represents an interesting unified coding approach using technologies like Reactive Extensions and other cross-platform tooling. Check it out!

Ready for a GitHub Extension for Visual Studio 2015? Phil Haack is back and talking smack about the cool code being build at GitHub. The announcement about the product was made at Build, but this show dives deeper into the technology involved. This new tool goes well beyond Git integration, diving deeper into the social media aspects of GitHub and making it easier for cross-platform teams to collaborate at GitHub. You need Visual Studio 2015 to take advantage of this, but that's pretty easy to get - try it out!

So when do IT Pros become developers? Carl and Richard talk to Steve Evans about his operations folks writing code and dealing with all the challenges that everyone finds when they write, own and operate code. The code in question is primarily PowerShell, used to automate deployment. It needs testing, source control and everything else that code requires. Steve talks about how his operations folks have come to learn the same things that all developers need to learn, and for the most part, end up using the same tools. PowerShell may have its own IDE, but more than ever the IT Pros want to use Visual Studio for its richness of tools and integration with resources that they realize they need. Now all we have to do is convince Microsoft that just because you operate systems doesn't mean you don't program!

Carl and Richard talk to Greg Shackles about making continuous deployment work in the mobile apps space. The conversation starts out with the projects that Greg is currently working on, and the need to get them deployed routinely to iOS and Android devices - this means, of course, deployment to the AppStore, which only goes so fast. Greg talks about avoiding the AppStores by finding ways to make changes internally, rather than a full deploy. The discussion goes down his preferred toolchain and inserting as much automation as possible - not so much for speed as for repeatability! Check out Greg's talks and slides in the show notes on this subject.

Carl and Richard talk to Brian Randell about the latest on Visual Studio Online (VSO). As Brian explains, this isn't about putting Visual Studio itself into the cloud, but rather surrounding your copy of Visual Studio with services in the cloud that make development easier. The conversation goes down the list, starting with source control in general and TFS specifically. One huge advantage of VSO - you don't have to deploy TFS yourself! Beyond source control and task management there are deployment tools and testing, all of which can be run in the cloud, where you're only billed for what you use. This is modern development at it's finest!

While at the Belfast stop of the .NET Rocks Ireland Road Trip, Carl and Richard talked to the one-and-only Martin Woodward! Martin discusses the state of Team Foundation Server today and source control in general. The conversation ranges over the evolution of source control, the effects of the internet on creating distributed version control and how Git, SVC, HG and many other version control systems all get along. Great thinking from a guy who's spent years deeply immersed in source control!

Carl and Richard talk to Cory Fowler from Microsoft about how developers can put together a continuous delivery solution on Azure. The conversation starts out with a discussion about the core concepts of continuous delivery - integrating build and testing together with lots of automation to create a minimum number of manual steps as possible. Cory describes how Azure becomes the environment for development, QA and production, and how the MSDN subscription can help at each step of the way. Lots of tools are discussed, there are many ways to get to this level of automation.

Carl and Richard talk to Phil Haack about his latest work at GitHub. The conversation starts out talking about semantic versioning, just trying to establish a coherent set of documents about how versioning should work. Phil then talks about licensing around GitHub, discussing a new service at GitHub to make it easier to select a software license - running down the differences between MIT, Apache and GPL licenses, and more! The discussion then digs into GitHub for Windows, or rather Phil's efforts to make GitHub generally more friendly for .NET developers. Lots of cool new features!

Carl and Richard talk to Pablo Santos about Plastic SCM. Plastic SCM is a free ALM tool for managing source and tracking for up to 15 developers - after that you have to pay. The conversation starts out with the concept of task-driven development, letting many developers work simultaneously on a project while each focusing on individual tasks. The challenge is excessive forking of the code - how do you get everything merged together again? Pablo talks about the powerful merge system of PlasticSCM to help identify identical code by functionality, not just syntax. Other strengths include great cross-platform support, integration with multiple IDEs, and support for lots of third party tools. Well worth checking out!

Carl and Richard talk to Brian Summers and Dave Harrison of Columbia Sportswear about their work with Team Foundation Server. Brian and Dave discuss how TFS has allowed them to build a common work planning environment for almost all of the developers at Columbia, including .NET, Java and ABAP (SAP) programmers. The conversation dives into how the different teams are able to collaborate, how they've built a service bus based on Biztalk and the future of even more rapid application development with TFS 2012.

Carl and Richard talk to Martin Woodward about Microsoft's announcement of integrating Git into Visual Studio and TFS. Martin talks about Microsoft's thinking about why and how to add Git support to Visual Studio and TFS. Then the conversation digs into the different configuration options - the Git tools work even in Visual Studio Express! Martin compares the differences of using Git directly from Studio as well as via TFS. It's a cool time to be a Studio programmer!

While at the Los Angeles stop of the .NET Rocks! Visual Studio 2012 Launch Road Trip, Carl and Richard sat down with Phil Haack to talk about Git and GitHub. The conversation starts out with a primer on Git itself and its origins as a distributed source control for Linux. Then Phil discusses how GitHub created a more social environment for Git with pull requests and dialog around code. Phil also talks about how GitHub itself operates as a non-hierarchical business and what that means to his work day, his career and his prospects for a raise. A little NuGet sneaks in there too!

Carl and Richard talk to Microsoft tech fellow Brian Harry about Application Lifecycle Management with Visual Studio 2012 and Team Foundation Server. Brian digs into the thinking around the latest version of Microsoft's ALM offering, including the new tools for gathering requirements using storyboards, video, audio and annotations around applications. The conversation also digs into the core 'lather-rinse-repeat' cycle of developers getting work assignments, building and checking in their code. Then onto the amazing new features in testing and providing early look versions of applications for gathering feedback. Finally Brian talks about Team Foundation Services, TFS-in-the-Cloud so to speak, which while not directly tied to Studio 2012, does point to Microsoft's future focus for ALM. A great conversation with a guy at the center of everything ALM at Microsoft!

Carl and Richard talk to Paul Betts and Tim Clem of Git about the new GitHub for Windows. GitHub has always been accessible by Windows, but far more Linux friendly. GitHub for Windows addresses this with a standalone client for GitHub. No Visual Studio required!

Carl and Richard talk to Eric Sink about his latest work in distributed version control. Eric talks about his latest project called Veracity, an open source distributed version control system that makes it easy for widely distributed teams to share source. The conversation digs into the sometimes subtle variations between source control systems and why their users are loyal and vocal to their preferred products. Lots of digressions and explorations, along with an offer for free books - check out the links for how to get a copy!

Carl and Richard talk to Martin Woodward about Team Foundation Server being used for more than just .NET development. Martin used to work for Teamprise until the technology was acquired by Microsoft in 2009 and he made the transition from being an MVP to being a 'blue badge'. The conversation ranges over the new features of TFS 2010, how Java developers can also use TFS, the effects of different process templates in TFS, and more!

Holy cow, it's show 600! Carl and Richard look back on memorable moments from the last 100 shows.

Roy Osherove talks about his experience moving his software company's build servers into the cloud. Recorded June 17th in the Fishbowl at the Norwegian Developer's Conference.