{"id":3222,"date":"2020-10-22T08:00:54","date_gmt":"2020-10-22T06:00:54","guid":{"rendered":"https:\/\/se-trends.de\/?p=3222"},"modified":"2021-01-13T09:30:42","modified_gmt":"2021-01-13T08:30:42","slug":"natural-language-requirements","status":"publish","type":"post","link":"https:\/\/www.se-trends.de\/en\/naturlichsprachige-anforderungen\/","title":{"rendered":"3 approaches to processing natural language requirements automatically"},"content":{"rendered":"

Model-based system descriptions are useful, but almost all developments start with requirements in human language. This will not change any time soon, because systems are developed for people. After all, we think in natural language (and images).<\/p>\n\n\n\n

However, natural language cannot be processed automatically without further ado, but this would be useful, especially with increasing complexity. There are already commercial solutions and, of course, research activities. In the following, I will present three solutions for the automated processing of requirements in natural language. The results are improved requirements, traceability or even derived models.<\/p>\n\n\n\n\n\n\n\n

A survey and a request<\/h2>\n\n\n\n

I would be very interested to know whether the automated processing of requirements is of interest to you. I am currently developing a new product for this purpose, Semiant<\/a>. The survey can be found on the product page<\/a>.<\/p>\n\n\n\n

I am also currently supervising a master's thesis on this topic. We are looking for companies to apply these concepts. Do you need help with your requirements that could possibly be simplified using an automated system? Then please contact me directly by e-mail<\/a> to.<\/p>\n\n\n\n

Thank you for taking part in the survey (right column)!<\/p><\/blockquote>\n\n\n\n

1. improve the quality of requirements<\/h2>\n\n\n\n

15 years ago, the DESIRe tool from the Hood Group<\/a> in the free download. DESIRe can be used with Word or DOORS and analyzes the requirements for weak words and other terms. However, this means that DESIRe does not automatically improve the requirements, but merely points out opportunities for improvement.<\/p>\n\n\n\n

Recently IBM has been trying to solve the problem with artificial intelligence (AI)<\/a> to solve the problem, which is known there as Watson<\/a> is marketed. Last year, Watson was integrated into DOORS, the result is in this video<\/a> to see. However, Watson does little more than DESIRe, which I find somewhat disappointing. Hubert Spie\u00df from IBM has set up a demo page where you can see the Try out requirements analysis with Watson yourself<\/a> can.<\/p>\n\n\n\n

The young company Qualicen has developed powerful tools for the automated analysis of natural language requirements. This Blog article provides an overview of four relevant use cases<\/a>. Qualicen is of the opinion that it is difficult to use these tools without appropriate support. Qualicen therefore only uses them in the context of consulting or training projects.<\/p>\n\n\n\n

Conclusion: Yes, there is automated assistance to improve the quality of requirements. But the improvement itself still has to be carried out by humans.<\/p>\n\n\n\n

2. recognize relationships<\/h2>\n\n\n\n

The company takes a completely different approach Relatics<\/a>. Their tool performs a semantic analysis, taking into account documents, emails and much more. Based on this analysis, the system creates a traceability, as shown schematically in the following image:<\/p>\n\n\n\n

\"\"