This paper presents a comprehensive survey of advancements and challenges in Handwritten Text Recognition (HTR), focusing on the French language and related datasets. The study introduces a hybrid form archive written in French: the Belfort civil registers of births. The digitization of these historical documents is challenging due to their unique characteristics such as writing style variations, overlapped characters and words, and marginal annotations. The objective of this survey is to summarize research on handwritten text documents and provide research directions toward effectively transcribing this French dataset. The survey classifies HTR systems based on techniques employed, datasets used, publication years, and the level of recognition. It also presents an analysis of the systems' accuracies, highlighting the best-performing approach. The paper showcases the performance of some HTR commercial systems and summarizes publicly available HTR datasets, especially those identified as benchmark datasets in the International Conference on Document Analysis and Recognition (ICDAR) and the International Conference on Frontiers in Handwriting Recognition (ICFHR) competitions. The paper presents updated state-of-the-art research in HTR and highlights new directions in the research field. The study discusses the challenges in HTR, including document layout, reading order, hybrid formats, marginal mentions, diverse text styles, skewness, and degradation. The paper also presents state-of-the-art architecture stages and performance accuracy reported on the French dataset (RIMES), along with an accuracy comparison of HTR commercial systems in recognizing English and French languages. The evaluation metrics used are the Word Error Rate (WER) and the Character Error Rate (CER). The results show that the best performance on isolated word recognition processes based on recurrent neural networks recorded an error rate of nearly 10%, whereas when combining the recognizers, the system achieved an error rate of 5%. The paper also highlights the importance of Belfort civil registers of births in advancing the field of handwritten text recognition due to its impediments in document layouts, reading orders, hybrid formats, marginal mentions, diverse text styles, skewness, and degradation.This paper presents a comprehensive survey of advancements and challenges in Handwritten Text Recognition (HTR), focusing on the French language and related datasets. The study introduces a hybrid form archive written in French: the Belfort civil registers of births. The digitization of these historical documents is challenging due to their unique characteristics such as writing style variations, overlapped characters and words, and marginal annotations. The objective of this survey is to summarize research on handwritten text documents and provide research directions toward effectively transcribing this French dataset. The survey classifies HTR systems based on techniques employed, datasets used, publication years, and the level of recognition. It also presents an analysis of the systems' accuracies, highlighting the best-performing approach. The paper showcases the performance of some HTR commercial systems and summarizes publicly available HTR datasets, especially those identified as benchmark datasets in the International Conference on Document Analysis and Recognition (ICDAR) and the International Conference on Frontiers in Handwriting Recognition (ICFHR) competitions. The paper presents updated state-of-the-art research in HTR and highlights new directions in the research field. The study discusses the challenges in HTR, including document layout, reading order, hybrid formats, marginal mentions, diverse text styles, skewness, and degradation. The paper also presents state-of-the-art architecture stages and performance accuracy reported on the French dataset (RIMES), along with an accuracy comparison of HTR commercial systems in recognizing English and French languages. The evaluation metrics used are the Word Error Rate (WER) and the Character Error Rate (CER). The results show that the best performance on isolated word recognition processes based on recurrent neural networks recorded an error rate of nearly 10%, whereas when combining the recognizers, the system achieved an error rate of 5%. The paper also highlights the importance of Belfort civil registers of births in advancing the field of handwritten text recognition due to its impediments in document layouts, reading orders, hybrid formats, marginal mentions, diverse text styles, skewness, and degradation.