Was sind die effektivsten Problemlösungsstrategien für die Softwareentwicklung?

Understand the business context and the problems that the solution is trying to solve before thinking of how to solve the problem

Understand the problem. Breakdown into High & Low level requirements. Document assumptions. Identify the design patterns that will be used to solve the problem. Break into decoupled modules & build iteratively. Put a testing strategy. And last but not least a documented acceptance criteria.

First off, understand the problem. Much of the time you are given a series of symptoms. Just because you stopped the runny nose doesn't mean the flu went away. Understanding the cause of the problem can take a little longer than just fixing a symptom but will prevent numerous other changes to 'fix' other symptoms. Various strategies can be used to find the root cause including talking to users, reviewing the application code and revisiting to ask clarifying questions. Once a root cause of the problem is identified you can create a plan to correct the effected parts. Solving root cause items can take longer than solving symptoms but it decreases the overall timeline, reduces bugs, improves stability, and overall user satisfaction.

Try to understand clearly the problem. In which conditions is it happenning? which persona have this problem? Once you have enough learning about the problem try to reproduce the same environment that cause the issue. Deeply involve into root cause investigation. Once you have enough knowledge on the system think about the possible solutions. Your approach may be different according to deadline or how emergent the situation is. So prepare short and long term solutions.

One of the most important things is to define the parameters of acceptance for a solution to be accepted as successful. Answers are different according to the domain, but the END customer(not always the paying one) will almost always be the one to build for. This requires a lot of detail orientation and the right questions. The answers to those questions help in deciding the right priorities for feature sets and the trade offs that may need to be considered later.

This looks straight forward at the beginning but when looked at the context of the entire system as a whole it is a step that needs to be done really well keeping in mind SOLID principles.

In software development, complex problems are best tackled by breaking them down into smaller, manageable subproblems. Techniques like abstraction, decomposition, modularization, and hierarchy help simplify the overall complexity, making it easier to handle. This approach enhances manageability, promotes reuse of components, and facilitates effective testing.

One phrase I often repeat to our teams is "Learn to bound your issues". Especially in complex systems, it can be easy to get overwhelmed by the multitude of places a problem could stem from. Even something as simple as writing some ideas down on a white board & crossing them off as you go can help to reveal the right answer. The TV show House did this for 8 seasons and I can honestly say that I very much give them credit for why I do it in my career today! 😎

Para poder desglosar un problema se puede realizar una matriz de riesgos e involucrados, donde podamos calificar cada riesgo implicado en el problema y entender el impacto a los involucrados.

Décomposer le problème en modules horizontaux et verticaux tout en encourageant la collaboration de l'utilisateur et en sollicitant son opinion est une approche très efficace pour obtenir une compréhension approfondie du problème et pour élaborer une solution cohérente et un bon schéma directeur.

Como desenvolvedora front-end, reconheço a importância de três etapas fundamentais para criar uma solução eficaz: projetar, implementar e testar. Projetar a Solução: Antes de começar a codificar, é essencial dedicar tempo ao planejamento e design. Isso envolve compreender completamente os requisitos do projeto, definir a arquitetura da aplicação e criar um layout que atenda às necessidades dos usuários.

Designing the solution blueprint from - Business architecture, - Use case - Technology/Solution architecture - Sequence/Flow/Consumer perspectives through various architecture diagram formats helps different stakeholders to visualize, how the said solution would resolve the problem from their respective view point.

After breaking down the problem, the next step is to design solutions for each subproblem by selecting appropriate data structures, algorithms, design patterns, and principles. Techniques like pseudocode, flowcharts, UML diagrams, and code reviews aid in this process. For example, if developing a sorting algorithm for a database, the design phase might involve choosing the most efficient sorting algorithm and representing it through pseudocode or a flowchart to ensure correctness and efficiency. This meticulous design ensures the resulting system is reliable, efficient, and easily maintainable.

Sobre el diseño ToBe que se realiza en la etapa de descubrimiento, y con las matrices de riesgos e involucrados, puedes diseñar una solución que te permita categorizar el valor de la misma según el impacto que tiene cada una de sus partes, pudiendo entregar valor desde el inicio de la solución, lo que ayuda a mitigar los riegos con más impacto desde el arranque.

This in my opinion is the most crucial step in the software development life cycle.Its like laying the foundation of a giant buiding.So doesn't matter how hard you work on the building construction and use quality materials and technology, if the foundation is not strong ultimately the building will collapse.In IT world, in the absence of a good solution, the application will not be steady and reliable no matter how good is the technology you used. A good solution not only ensures meeting of the business requirement but at the same time also considers robustness , reusability and minimum regression impacts.This further helps to reduce operational maintenance and cost.

This phase is normally spead over considerable amount of time or days. So one of the important step towards end of this phase is to look back and reflect upon what has been implemented. Developer need to check if there are any open points, assumption which are not clarified. Perform self review before actual review. Many times this crucial step is missed and easily avoidable bugs enter the development stream.

Implementation of solution(s) triggered from reviewed/approved design(s), where the approach of implementation from - execution (agile) - technical (best practices from language/framework/platform) standpoint - governance (connected and visible track of implementation task to the business problem) helps to keep it effective, plus it helps to evolve (fail fast) if any gaps/changes are identified.

Para su implementación, te recomiendo que realices ejercicios de escritorio donde pruebes la lógica de la solución antes de implementar, para que al momento de implementar tengas una primera prueba previa.

Amazon's approach to testing its AWS architecture offers a comprehensive multi-layered testing strategy that includes unit tests for individual components, integration tests to ensure components work together seamlessly tests to evaluate the entire ecosystem's performance under various conditions, use automated testing tools to simulate peak loads and assess the system's resilience and scalability, before launching new features or services Amazon employs chaos engineering practices injecting faults into the system to test and improve its resilience This rigorous testing framework ensures that services remain reliable and efficient, even under extreme operational conditions demonstrating the critical role of thorough testing in system design

Genera una matriz de pruebas por escenario desplegado con el detalle de la ejecución, estos datos te podrán ayudar a automatizar escenarios en caso de ser requerido, y tendrás evidencias concluyentes de que la solución funciona. Adicionalmente en el knowledge transfer es fácil identificar las partes probadas y los casos de uso.

Usability Testing: Usability testing evaluates the software's user interface (UI) and user experience (UX) to ensure that it is intuitive, easy to use, and meets user expectations. It involves gathering feedback from end-users surveys, interviews, to identify usability issues and areas for improvement. Compatibility Testing: Compatibility testing verifies that the software functions correctly across different devices, operating systems, browsers, and environments. It ensures that the software is compatible with a wide range of configurations and does not exhibit any unexpected behavior. Localization and Internationalization Testing: Localization testing verifies that the software is adapted to different languages, cultures, and regions.

Testar a Solução: Por fim, não podemos subestimar a importância dos testes. Realizar testes finais é essencial para garantir a qualidade e a segurança da solução desenvolvida. Isso inclui testes de usabilidade, de desempenho e de segurança, entre outros, para garantir uma experiência livre de problemas para os usuários finais.

Netflix's transition from a monolithic to a microservices architecture exemplifies evolving system design for scalability and agility. Facing growing user demand, Netflix reengineered its platform to microservices, allowing independent scaling and faster deployment of features. They utilized cloud infrastructure for resilience and global reach. This evolution facilitated Netflix's rapid growth, enabling it to handle millions of concurrent streams with high availability and low latency, showcasing the importance of continuous adaptation in system design.

Design thinking Design thinking principles are the guidelines to start thinking about and implementing problem solving in an iterative manner. User interviews, asking open ended questions, surveys will help in getting the problem right. Double diamond approach to design, prototype and iterative delivery are proven ways to approach solution.

I think this seven seven step article misleads the reader a bit into thinking System Design can be understood as a 7 step process. However, I think assumptions, design choices and risks need to be re-evaluated continuously over time.

Effective problem-solving strategies in software development, especially for big tech, include iterative development, modular design, and extensive testing. For instance, Google uses Site Reliability Engineering (SRE) to ensure their systems are scalable and reliable. They focus on automating operations tasks and measuring reliability through service-level objectives (SLOs), enabling rapid iteration and resilience. This approach helps in managing complex systems by breaking down problems, ensuring continuous improvement, and maintaining high service quality.

People don't realize how much software & data engineering can be automated -- giving lower cost & risk, shorter turnaround, & higher quality results. I've been automating software & data work for decades using XTRAN, the Expert System I created in 1984. It automates: o Analysis / assessment, e.g. code quality assessment, architectural analysis, impact analysis o Transformation / re-engineering, e.g. code quality remediation, modernization, repurposing o Translation, e.g. escape dying languages, move code to portable ones o Code generation from repository data We use XTRAN plus TemplaGen, a data wrangling utility created in XTRAN's rules language, to automate software & data work in all of those categories.

You need to manage the chaos. The biggest hurdles are other projects, unknown unknowns and things you forgot to think about. Once you head those off, the rest is relatively easy.