Skip to content. | Skip to navigation

solving complex IT problems!

Personal tools

You are here: Home / Blog / Agile work

Agile work

Software reviews – does high-quality software justify its cost?

Reviews take place throughout our entire software development process, starting with the analysis of requirements and the appropriate software architecture and ending with code and testing analyses. However, we also conduct audits of our customers software so that they receive an assessment for the further development and operation of their software.

What are the benefits of software reviews?

Software reviews lead to significantly fewer errors. According to a study of 12 thousand projects by Capers Jones [1], these are reduced by

  • in requirements reviews by 20–50%
  • in top-level design reviews by 30–60%
  • in detailed functional design reviews by 30–65%.
  • for detailed logical design reviews by 35–75%.

The study concludes:

Poor code quality is cheaper until the end of the code phase; after that, high quality is cheaper.
Costs across the development phases of requirement, design, coding, testing and operation

The study still assumes a classic waterfall model from requirements to software architecture, first to code, then to tests and operation. In agile software development, however, this process is repeated in many cycles, so that we spend most of our time developing software by changing or adding to the existing code base. Therefore, we need to understand not only the change requirements but also the existing code. Better code quality makes it easier for us to understand what our application does and how the changes make sense. If code is well divided into separate modules, I can get an overview much faster than with a large monolithic code base. Further, if the code is clearly named, I can more quickly understand what different parts of the code do without having to go into detail. The faster I understand the code and can implement the desired change, the less time I need to implement it. To make matters worse, the likelihood that I will make a mistake increases. More time is lost in finding and fixing such errors. This additional time is usually accounted for as technical debt.

Conversely, I may be able to find a quick way to provide a desired function, but one that goes beyond the existing module boundaries. Such a quick and dirty implementation, however, makes further development difficult in the weeks and months to come. Even in agile software development without adequate code quality, progress can only be fast at the beginning, the longer there is no review, the tougher the further development will be. In many discussions with experienced colleagues, the assessment was that regular reviews and refactorings lead to higher productivity after just a few weeks.

Which types of review solve which problems?

There are different ways to conduct a review. These depend on the time and the goals of the review.

The IEEE 1028 standard [2] defines the following five types of review:

With this static analysis technique we develop scenarios and do test runs, e.g. to find anomalies in the requirements and alternative implementation options. They help us to better understand the problem, but do not necessarily lead to a decision.
Technical reviews
We carry out these technical reviews, e.g. to evaluate alternative software architectures in discussions, to find errors or to solve technical problems and come to a (consensus) decision.
We use this formal review technique, for example, to quickly find contradictions in the requirements, incorrect module assignments, similar functions, etc. and to be able to eliminate these as early as possible. We often carry out such inspections during pair programming, which also provides less experienced developers with quick and practical training.
Often, before a customer’s software is put into operation, we conduct an evaluation of their software product with regard to criteria such as walk-through reports, software architecture, code and security analysis as well as test procedures.
Management reviews
We use this systematic evaluation of development or procurement processes to obtain an overview of the project’s progress and to compare it with any schedules.

Customer testimonials

Thank you very much for the good work. We are very happy with the result!

– Niklas Kohlgraf, Projektmanagement, pooliestudios GmbH

Get advice on software reviews from Cusy today:

Veit Schiele

Veit Schiele
Phone: +49 30 22430082

We will also be happy to call you!

Request now

[1]Capers Jones: Software Quality in 2002: A Survey of the State of the Art
[2]IEEE Standard for Software Reviews and Audits 1028–2008

Are Jupyter notebooks ready for production?

Are Jupyter notebooks ready for production?

Jupyter Notebook

In recent years, there has been a rapid increase in the use of Jupyter notebooks, s.a. Octoverse: Growth of Jupyter notebooks, 2016-2019. This is a Mathematica- inspired application that combines text, visualisation, and code in one document. Jupyter notebooks are widely used by our customers for prototyping, research analysis and machine learning. However, we have also seen that the growing popularity has also helped Jupyter notebooks be used in other areas of data analysis, and additional tools have been used to run extensive calculations with them.

However, Jupyter notebooks tend to be inappropriate for creating scalable, maintainable, and long-lasting production code. Although notebooks can be meaningfully versioned with a few tricks, automated tests can also run, but in complex projects, mixing code, comments and tests becomes an obstacle: Jupyter notebooks can not be sufficiently modularized. Although notebooks can be imported as modules, these options are extremely limited: the notebooks must first be fully loaded into memory and a new module must be created before each cell can run in it.

As a result, it came to the first notebook war, which was essentially a conflict between data scientists and software engineers.

How To Bridge The Gap?

Notebooks are rapidly gaining popularity among data scientists and becoming the de facto standard for rapid prototyping and exploratory analysis. Above all, however, Netflix has created an extensive ecosystem of additional tools and services, such as Genie and Metacat. These tools simplify complexity and support a broader audience of analysts, scientists and especially computer scientists. In general, each of these roles depends on different tools and languages. Superficially, the workflows seem different, if not complementary. However, at a more abstract level, these workflows have several overlapping tasks:

data exploration occurs early in a project

This may include displaying sample data, statistical profiling, and data visualization

Data preparation

iterative task

may include cleanup, standardising, transforming, denormalising, and aggregating data

Data validation

recurring task

may include displaying sample data, performing statistical profiling and aggregated analysis queries, and visualising data

Product creation

occurs late in a project

This may include providing code for production, training models, and scheduling workflows

A JupyterHub can already do a good job here to make these tasks as simple and manageable as possible. It is scalable and significantly reduces the number of tools.

To understand why Jupyter notebooks are so compelling for us, we highlight their core functionalities:

  • A messaging protocol for checking and executing language-independent code
  • An editable file format for writing and capturing code, code output, and markdown notes
  • A web-based user interface for interactive writing and code execution and data visualisation

Use Cases

Of our many applications, notebooks are today most commonly used for data access, parameterization, and workflow planning.

Data access

First we introduced notebooks to support data science workflows. As acceptance grew, we saw an opportunity to leverage the versatility and architecture of Jupyter notebooks for general data access. Mid-2018, we started to expand our notebooks from a niche product to a universal data platform.

From the user’s point of view, notebooks provide a convenient interface for iteratively executing code, searching and visualizing data – all on a single development platform. Because of this combination of versatility, performance, and ease of use, we have seen rapid adoption across many user groups of the platform.


Along with increasing acceptance, we have introduced additional features for other use cases. From this work notebooks became simply paramatable. This provided our users with a simple mechanism to define notebooks as reusable templates.

Workflow planning

As a further area of notebook ​​applications, we have discovered the planning of workflows. They have the following advantages, among others:

  • On the one hand, notebooks allow interactive work and rapid prototyping and on the other hand they can be put into production almost without any problems. For this the notebooks are modularized and marked as trustworthy.
  • Another advantage of notebooks are the different kernels, so that users can choose the right execution environment.
  • In addition, errors in notebooks are easier to understand because they are assigned to specific cells and the outputs can be stored.


In order to be able to use notebooks not only for rapid prototyping but also for long-term productivity, certain process events must be logged so that, for example, errors can be diagnosed more easily and the entire process can be monitored. IPython Notebboks can use the logging module of the standard Python library or loguru, see also Jupyter-Tutorial: Logging.


There have been a number of approaches to automate the testing of notebooks, such as nbval, but with ipytest writing notebook tests became much easier, see also Jupyter Tutorial: ipytest.


Over the last few years, we have been promoting close collaboration between Software Engineers and data scientists to achieve scalable, maintainable and production-ready code. Together, we have found solutions that can provide production-ready models for machine learning projects as well.

Technical debts

We usually develop the functionality requested by our customers in short cycles. Most of the time, we clean up our code even after several cycles when the requirements hardly change. Sometimes, however, code that has been developed quickly must also be put into production. Technical debt is a wonderful metaphor in this regard, introduced by Ward Cunningham to think about such problems. Similar to financial debt, technical debt can be used to bridge difficulties. And similar to financial debt, technical debt requires interest to be paid, namely in the form of extra effort to develop the software further.

Unlike financial debt, however, technical debt is very difficult to quantify. While we know that it hinders a team’s productivity in developing the software further, we lack the calculation basis on which to quantify this debt.

To describe technical debt, a distinction is usually made between deliberate and inadvertent technical debt, and between prudent or reckless debt. This results in the following quadrant:

  inadvertent deliberate
prudent «Now we know how we should do it!» sollten!»* «We have to deliver now and deal with the consequences »
reckless «WWhat is software design?» «We don’t have time Rfor software design!»

Read more