We point out some relevant issues that are related to the computing-with-words (CWW) paradigm and argue for an urgent need for a new, nontraditional look at the area, since the traditional approach has resulted in very valuable theoretical research results. However, there is no proper exposure and recognition in other areas to which CWW belongs and can really contribute, notably natural-language processing (NLP), in general, and natural-language understanding (NLU) and natural-language generation (NLG), in particular. First, we present crucial elements of CWW, in particular Zadeh's protoforms, and indicate their power and stress a need to develop new tools to handle more modalities. We argue that CWW also has a high implementation potential and present our approach to linguistic data(base) summaries, which is a very intuitive and human-consistent natural-language-based knowledge-discovery tool. Special emphasis is on the use of Zadeh's protoform (prototypical form) as a general form of a linguistic data summary. We present an extension of our interactive approach, which is based on fuzzy logic and fuzzy database queries, to implement such linguistic summaries. In the main part of the paper, we discuss a close relation between linguistic summarization in the sense considered and some basic ideas and solutions in NLG, thus analyzing possible common elements and an opportunity to use developed tools, as well as some inherent differences and difficulties. Notably, we indicate a close relation of linguistic summaries that are considered to be some type of an extended template-based, and even a simple phrase-based, NLG system and emphasize a possibility to use software that is available in these areas. An important conclusion is also an urgent need to develop new protoforms, thus going beyond the classical ones of Zadeh. For illustration, we present an implementation for a sales database in a computer retailer, thereby showing the power of linguistic summaries, as well - - as an urgent need for new types of protoforms. Although we use linguistic summaries throughout, our discussion is also valid for CWW in general. We hope that this paper-which presents our personal view and perspective that result from our long-time involvement in both theoretical work in broadly perceived CWW and real-world implementations-will trigger a discussion and research efforts to help find a way out of a strange situation in which, on one hand, one can clearly see that CWW is related to words (language) and computing and, hence, should be part of broadly perceived mainstream computational linguistics, which lack tools to handle imprecision. These tools can be provided by CWW. Yet, CWW is practically unknown to these communities and is not mentioned or cited, and---reciprocally---even the top people in CWW do not refer to the results that are obtained in these areas. We hope that our paper, for the benefit of both the areas, will help bridge this gap that results from a wrong and dangerous fragmentation of break science.