Just recently I had the privilege of doing a formal analysis of the code base of the client using CodeScene. I spent two days analysing the code base of around 35 active developers. The company wanted to see if some of the difficulties they had could be better understood by looking at the code and more importantly by looking at the interaction patterns with the code. My peers were thrilled at the level of understanding they could acquire in such a short timespan! Not only does it give hope in showing that it's only a small fraction of the technical debt that actually counts, including which part! But it also gives valuable insights into concerns around the code, like team organisation, developer turnover and even product management difficulties. I've been using CodeScene for more than a year now. I'm a technical coach and as I get to see a lot of different teams. CodeScene really improves my understanding each team's situation and thus I can be of better use to them. Tools are only tools numbers are only numbers but when a tool is making the invisible visible then it is really, really useful for making more informed decisions!

Otherwise almost invisible information is pulled up right into our face The incredible insight it gives into the pain points of the project in a very short time. That it focuses on only the most costly problems It's focus is not only code, but also team dynamics, delivery dynamics and architecture. I don't know of other tools that look at Change Coupling Very visual

The UX is sometimes confusing, this is probably where there's most room for improvement.

CodeScene as a company has been highly responsive to my input over the whole course of my engagement with them. My suggestions based on real-world usage contributed to improvement in the products. Throughout my entire engagement, the CodeScene team has stayed highly engaged to make sure I was getting the most out of the product. Opening the tool for the first time was a truly exciting experience, and the more I learned, the more excited and empowered I got.

CodeScene gave me a view of the code base in a way that no other tool - or person - could. It helped me understand where our quality issues really came from, and to redirect my team to focus on remediating those areas. This not only improved quality but helped make sure we were focused on the most refactoring value.

There aren't a lot of downsides to this tool. The subject matter requires a lot of knowledge that takes time to attain. This is to say you have to understand the theory behind the insights to get the most value out of them. However, I'd argue that any engineering manager -should- understand this theory if they want to be effective at managing any code base. It shows up all the meaningful metrics I'd always hoped for and never knew existed.

CodeScene is one of the best visual tools to identify pull requests that leads to technical debts.

The most important thing about Codescene is that it allows the software development team to remove technical debts from the source code that can severely impact the performance of the software system later.

It was indeed quite difficult to remove all the technical debts all at once and when we use this tool our release frequency decreased.

Above all else, I appreciated that CodeScene would provide data to check my instincts about where and when to apply more effort to improve a design. As a developer, CodeScene’s hot spot feature helped me feel more confident (or helped me find another candidate) when I identified an area to improve. As a team manager, the metrics and graphs in CodeScene gave our team a way to monitor progress as we worked on a several week effort to make small, incremental improvements to a tricky area of code.

CodeScene gave us a new insight into a problematic area of our code - team structure. We identified a hotspot that continued to grow as we added new features. The Team Dynamics views highlighted an issue that kept the hotspot growing; several teams contributed equally to the code - a case of a missing shared mental model and lack of clear stewardship. The information generated by CodeScene helped us create a plan to improve the code and the visuals helped us tell a compelling story to our whole team. We used CodeScene on new projects to help us come up to speed on a new project quickly. We identified the hotspots to prioritize conversation topics with the original authors as we transitioned the code from one team to another. As we started making changes to the project, I appreciated the GitHub Pull Request integration that provided an extra check to see if we had missed changing some files that had historically changed together and might have caused a bug in production.

We used the hosted solution (codescene.io) which tended to lag the on-premises version and get the newest features later. We were unable to use some of the delivery-focused features as we used Pivotal Tracker and later Clubhouse, neither of which were supported at the time. The UI was sometimes hard to get a summary of the information we needed, though that has improved with the Hotspots Code Health view.

CodeScene prioritises the biggest bottlenecks in our projects. It finds the code improvements that give the biggest improvement in the development and maintenance of our codebases. We use CodeScene to XRAY our codebases quickly and improve the structure of our code and sometimes even the structure of our teams. We use it as a guide for work conversations and code / project reviews and to assess (external) codebases. The super easy setup, leading to detailed insights, led to explosive growth of the service within our company. We didn't have to put a lot of effort into the adoption; CodeScene sells itself.

Running an analysis with minimal configuration (the repository path and a branch) already gives a lot of insight. By fine-tuning the configuration and including external information from Gitlab and Jira (among others) the CodeScene analysis will provide even more insight. I love the ability to quickly zoom in and out from architecture level to code level and back again.

Although most developers really like the CodeScene perspective on their codebase, it's sometimes hard to follow up with improvements. With the integration into build pipelines, the code supervision functionality, the reports, monitors and the actionable improvement advice, this issue is mostly solved.

CodeScene has helped us find ways to improve the productivity of our development teams and measure our success. Ultimately improving how quickly we can deliver value to the business.

The recent addition of delta analysis and hooks into our CI/CD tooling has got to be my favourite feature. It has allowed us to incorporate analysis results into our software development life cycle. This has really improved the way insight from the tool is fed back into our teams as continuous, bit sized chucks of information. This helps keep teams engaged and proactively looking for further insight from the tools many other features.

Integrating with other CI/CD tooling can be tricky and has previously required lots of further setup. Some of the analyses do attempt to mine data from free text, so they require some more complex setup to attempt to capture conventions used by the teams and their git tooling. However, the CodeScene team have always been very responsive to feedback and continuing to improve the product to make these things easier.

I have utilized CodeScene both for item advancement (on-going investigation) and oddball examinations for projects we have been on, and the outcomes consistently disclose to us something. In one case it assisted us with anticipating what part of our venture planned to defer the item discharge!! With CodeScene it took us a couple of moments to perceive how clear it was, however it was not clear to the designers and the Project Manager without CodeScene. Not at all like numerous other static examination devices, CodeScene doesn't need a protracted arrangement or reconciliation measure - you simply point it at your Git archive and it gives quick bits of knowledge. Their help group is additionally great at answering, and the author, is an uncommon individual. I'm colossally enamored with the instrument, and the organization and their estimating model behind it urge me to utilize it considerably more. I would suggest that everybody chipping away at a huge undertaking (>10 engineers) run their code through CodeScene

Very simple to arrangement and get results - simply point it at your Git repos! Bits of knowledge are extraordinary and consistently give a point of view that I have not seen anyplace before For creating microservices, this is one of the just out of the container approaches for distinguishing compositional conditions (through fleeting coupling)

The device could do with some development on the worker side, yet as it is very straightforward and deployable in a Docker holder, this isn't actually an issue

The highest risks in the code base lead directly to quality problems in production. Finding these is made considerably easier with CS. In addition, refactoring away from a monolithic design is eased considerably with this this tool. We use the tool in both weekly analysis mode and as part of the CI pull requests to identify risky commits.

The visualisation and prioritisation of code helps teams understand the magnitude of the task before them and guides them on how to choose where to start.

Sometimes the user interface is not the most intuitive. Also, pervious versions lacked tools for easy imports for large organisations.

I have used CodeScene both for product development (on-going analysis) and one off analyses for projects we have been on, and the results always tell us something. In one case it helped us predict which component of our project was going to delay the product release!! With CodeScene it took us a few seconds to see how obvious it was, but it was not obvious to the developers and the Project Manager without CodeScene. Unlike many other static analysis tools, CodeScene does not require a lengthy setup or integration process - you just point it at your Git repository and it provides immediate insights. Their support team is also really good at replying, and the founder, is an exceptional person. I am hugely fond of the tool, and the company and their pricing model behind it encourage me to use it even more. I would recommend that everyone working on a large project (>10 developers) run their code through CodeScene.

Super easy to setup and get results - just point it at your Git repos! Insights are unique and always provide a perspective that I have not seen anywhere before For developing microservices, this is one of the only out of the box approaches for identifying architectural dependencies (through temporal coupling)

The tool could do with some maturity on the server side, but as it is quite simple and deployable in a Docker container, this is not really an issue

I really like the Delivery Performance functionalities that allows me to rapidly visualize and act on the most important lead time metrics such as Lead Time for Changes or Planned vs Unplanned work. This insight and the possibility to quickly identify areas in our code with highest technical debt and complexity, areas that leads to quality issues, is a killer feature

The possibility to connect the history of your code with business metrics that allows you to drive and push for increased speed as this is increasingly more important going forward.

The user interface could be more intuitive.

I am really satisfied with CodeScene becuase it is more than just a static analysis tool, it provide and overview of the real problemens and what to focus on in am easy-to-use package.

The whole idea of using historical data, together with the different features CodeScene offers, makes it much easier to guide your improvement work and improve the quality of your code base, including your test code. By starting on a high level you get a good overview of the most problematic areas and the you can drill down and look at the details and see where your real problems are. Combined with trend analysis, code biomarkers and other features like analyzing social interaction between teams, i.e. are they all over the place or in their own modules, makes it a very versatile tool when working with quality. When I was working on my Ph.D. thesis I wanted to develop a tool that could visualize the problematic areas in an intuitive way. Now, twenty years later it actually is here!

It does not get any better than the data you are using. If you have not commented your pull requests or check-ins correctly, to be able to trace e.g. defects in certain areas, then it will of course be difficult to make any conclusions.