1
REPORT.
Flat Plate Boundary Layer Report
Student Name
Institutional Affiliation
Course Details
Professor’s Name
Date of Submission
2
Introduction
Documentation is an essential part of software engineering as it supports the
development, maintenance, and usage of documents. Documentation is even more important in
the case of software as medical devices since the requirements are far stricter, and patient safety
and the effectiveness of the product are at stake. This section deals with an understanding of
documentation in software engineering, the place that documentation occupies in the SaMD, and
the possibility of ADG. It gives an overview of SLR as the method used in this study to collect
and assess information on this topic.
Importance of Documentation in Software Engineering, Particularly in SaMD
There are a number of reasons why documentation is important in software engineering.
It enhances communication within a particular group of workers. It is useful for the identification
of areas that require further work in the future, as well as as a guide at the end-user level.
Documentation comprises of requirement specifications, design documents, code comments,
writing of user guides, and more. In the context of SaMD, documentation is even more
significant because such devices operate in a highly regulated context. These agencies include
the Food and Drug Administration or FDA as well as the European Medicines Agency or EMA,
and they demand a lot of paperwork that proves that the software is safe for use, effective and
efficient in the performance of its tasks. This entails technical documentation such as the clinical
evaluation report, risk management files, and post-market surveillance scripts. In addition to
being a technical and legal requirement, clear documentation for the target audiences, such as the
developers, regulatory reviewers, and healthcare practitioners who will be using the software, is
critical.
3
Role of Automated Documentation Generation
Automated documentation generation is the process of generating or updating
documentation with as little manual interference as possible through the use of certain
applications. Given the fact that the SaMD industry is quite sensitive and well-controlled, ADG
helps to perform some essential functions that include rationalizing the documentation process
and minimizing the possibilities of errors. Manual documentation techniques, as used in the
traditional methods, are slow and are characterized by mistakes. It is essential to underline the
fact that ADG tools are capable of populating information from different phases of the SSD
paradigm as well as providing accurate and timely outputs. It also helps save a considerable
amount of time and guarantees that all documentation is in line with the regulations, which
simplifies the submission of regulatory documentation and compliance audits. In this way, ADG
relieves the developers from spending time on monotonous and time-consuming activities that
may lead to mistakes and, thus, contributes to increasing the speed of creating and approving
SaMD.
Overview of the Systematic Literature Review (SLR) Process
A systematic literature review (SLR) is a deliberate technique of identifying, appraising,
and synthesizing all the available studies pertinent to a specific research question, field of study,
or phenomenon under investigation. The SLR process involves several key phases: The SLR
process involves several key phases:
1. Planning Phase: The activities in this phase include coming up with a well-
understood protocol that defines the review's scope, aims and approach. The protocol is a
checklist that is followed to ensure that all the other reviews repeat the same pattern. Some of the
4
elements of the protocol involve determining the research questions, the databases to be used
when searching, and other criteria that would be used for the inclusion and exclusion of studies.
2. Conducting Phase: In this stage, authors decide on the digital libraries they are going
to use, create search terms from keywords and Boolean logic and search systematically through
the literature. The identified studies are then screened for data extraction, and quality assessment
is also done on the selected papers. Data extraction is a process of gathering necessary
information from each study, whereas quality assessment checks the reliability and validity of
the selected studies.
3. Reporting Phase: The final phase involves integrating the data, analyzing the result,
and achieving the best and clearest presentation of the result. The SLR results have given solid
background information, the existing knowledge to be further developed, and the research gaps
to be filled.
Systematic Literature Review (SLR) Process
Systematic literature review is a process of collecting and analyzing the existing body of
work with regard to a certain subject matter. This systematic approach to the review means that
the exercise coverage is comprehensive, and the results obtained are accurate and conclusive.
The SLR process is divided into two main phases: Two major phases of research are generally
categorized, namely the planning phase and the conducting phase. Every phase includes
particular activities to enhance the stringency and credibility of the steps in the review.
Planning Phase
1. Defining the Protocol:
In the first level of the planning phase, there is the need to establish a detailed and crisp
protocol. This protocol aims to provide a guide on how the SLR will be conducted and includes,
5
among others, the review objectives, the extent of the study, and the method of review. It ensures
that the review process is very standard and that some measure of equality is observed. It usually
defines the research questions and objectives, participants' criteria, and data collection and
analysis methods. This phase also entails seeking feedback from peers or experts to fine-tune the
protocol needed in today's practice and its practicability.
2. PICOC Criteria:
In the planning phase, the PICOC framework is a convenient instrument for
disassembling the research questions into index terms. PICOC stands for:
Population (P): The general field of study that the research is focusing on, for instance,
automotive, software as medical devices (SaMD).
Intervention (I): The field of research, which is the methodology or technology under
examination, such as automated documentation generation (ADG).
Comparison (C): Other approaches that involve comparisons with the traditional methods
or variants of the tool under investigation.
Outcome (O): It may be the desire for enhanced achievement of objectives, increased
efficiency, and reduction of mistakes where documentation is involved, among others.
Context (C): Environment or context in which the intervention is delivered.
Research performed using PICOC ensures that the main questions to be answered are
formulated properly and proper search strategies are developed.
Synonym Generation for Key Terms:
Creating synonyms to use when searching for the keyword is crucial in searching for
results. This involves drawing a list of other keywords or key phrases that are synonymous with
6
the core keywords. For instance, in the case of 'automated documentation generation,' other
search terms could be 'automatic documentation generation,' 'documentation automation,' and
'automation of generation of documentation,' among others. The use of synonyms in the search
strategy creates a higher probability of identifying all the required studies.
Conducting Phase
1. Selecting Digital Libraries:
Selecting the right source of digital libraries is, therefore, essential in enabling one to
access a wide array of literature in the area under focus. Some of the most frequently utilized
digital libraries in software engineering and the medical sciences areas are Scopus, the Web of
Science, the IEEE Xplore, and the ACM Digital Library. These libraries contain
multidisciplinary open access, fully indexed peer-reviewed scholarly journals, conference papers
and technical reports.
2. Building Search Strings:
Search string acculturation also requires including the key terms and those with related
meanings using the Boolean operators AND OR. These search strings are then employed to
search for the selected digital libraries. For instance, a search string for the topic of automated
documentation generation in SaMD might look like this: Using the search terms ("automated
documentation generation" OR "automatic documentation generation" OR "documentation
automation") AND ("software as medical devices" OR "SaMD"). This step enables the search to
both extend and narrow down the state of the art as well as the gaps it identifies.
7
Data Extraction and Quality Assessment:
Data extraction follows the process of determining relevant studies by conducting
literature searches, which involves the systematic extraction of data from each study. These are
aspects like the survey aims, research methods, findings, and or recommendations. In order to
avoid the inclusion of substandard research studies, a quality check is conducted on the titles and
abstracts of the chosen articles using quality assessment criteria. Such an assessment may entail
the analysis of the work done, including study design, sampling technique used, data analysis
and synthesis and the strength of the conclusions drawn. The focus is placed on high-quality
studies in order to build the findings of the review on strong evidence.
Through these strict procedures, the SLR process makes sure that the review conducted
becomes balanced, thorough, and scientifically sound, and it serves as essential groundwork for
understanding research in the field of automated documentation generation in SaMD.
Automated Documentation Generation
Automated documentation generation can be defined as the process of generating
documents or documentation with very little human input. This process is valuable in many
industries, especially SaMD, where documenting the solution is required to address legal and
user' needs and facilitate maintenance. The use of automation provides efficiency, quality, and
current data compared to the use of manual methods since it has no exposure to error, as do
humans.
Automated documentation generation has several methodologies and tools. Some of the
approaches are template-based generation, in which templates are prepared in advance with the
information that is being filled in from the software systems, and model-driven generation, in
which system models are used to generate the documentation. Software documentation creation
8
from its source code is achieved with tools such as Dioxygen, Javadoc, and Sphinx. It identifies
the available code along with comments and annotations and structures them into documentation
formats. In particular, there are specialized tools such as FDA submitter and Open Regulatory
used for Sam Ds to meet all the obligatory criteria and make submissions easier.
Comparison of Manual and Automated Approaches Manual documentation refers to the
circumstances under which people take their time to type the manuscripts, organize them, and
make changes whenever needed. This conventional method can take a lot of time and is also very
sensitive to errors, as seen in the case of SaMD, which requires regular updates. Automated
documentation, on the other hand, pulls information from the latest code and data with the help
of tools. It also saves time during documentation and, at the same time, guarantees that the
documentation stays in line with the software, thus requiring likely changes and updates.
Benefits of Automation in SaMD Documentation
The benefits of automated documentation in SaMD are manifold: Benefits of
Automation in SaMD Documentation The benefits of automated documentation in SaMD are
manifold:
1. Consistency and Accuracy: The code generation also helps to increase the reliability
of documentation and make it more consistent through all sections with no contradictions.
2. Time and Cost Efficiency: Regarding the documentation aspect, automation
substantially has the advantage of decreasing the time spent as well as the necessary effort to
generate and maintain documentation while encouraging devoted developmental efforts.
3. Regulatory Compliance: Documentation tools that are built for SaMD applications
allow work compliance with rules and regulations that include FDA and EU MDR by including
the right templates and guidelines.
9
4. Real-Time Updates: With these changes, new versions of the documentation can be
produced easily as a result of interaction with the software, thus eliminating manual updates.
5. Enhanced Collaboration: Automated systems make it easy for several employees
from a particular team and department to work on the documentation and review the
documentation at the same time, which reduces work hold up when it comes to documentation.
In conclusion, the use of automated documentation generation as one of the key
advancements in the development of SaMD is rather valuable as compared to manual
approaches. Documenting the project activities also improves the quality of documentation since
the frequency at which documentation is done is considerably improved, resulting in better
compliance with the regulations and improving the aspect of project documentation.
Consequently, as software systems get larger and more intricate, the need to utilize automated
tools for documentation becomes more critical for the processes to catch up with development
activities.
Case Study: Software as Medical Devices (SaMD)
Uniqueness of Documentation for SaMD
Software such as medical devices (SaMD) is a sub-sector of medical devices that have
different documentation needs from other medical devices. The documentation specification that
remains most critical to SaMD is related to documenting how the modified software is safe and
effective to use. Key documentation elements include:
1. Clinical Evaluation Reports: These reports are useful for showing the clinical
evidence for SaMD to prove the worth of the product in treating patients. They should include
documentation from the permission of clinical trials, or research to demonstrate the efficacy and
safety of the software.
10
2. Risk Management Files: Special attention should be paid to making detailed
documentation of risk assessments that are necessary for defining potential hazards connected
with the software, as well as the ways of their minimizing them.
3. Software Development Lifecycle (SDLC) Documentation: All these processes need
to be documented to the greatest possible extent, including the initial design of the software and
the final validation. Typically, this comprises of requirement specifications, design documents,
coding standard procedures, verification and validation documentation and records, and
maintenance logs and records.
4. Post-Market Surveillance Plans: These contain the proposal of how the software’s
activity would be regulated once in use and made to remain safe from vulnerabilities or
inefficient from malicious users.
Regulatory Considerations
The need for regulation of SaMDs depends on the country; however, the regulations
should meet the requirements of the IMDRF. Key regulatory considerations include:
1. Classification: SaMD is categorized according to the level of risk concerning the
patient's health, which can be categorized as low or high. This classification defines the level of
regulation the firm is subjected to.
2. Compliance with Standards: SaMD developers must adhere to some standards, such
as ISO 13485 and IEC 62304, among others.
3. Pre-market Approval: Depending on the risk classification, SaMD might need
premarket clearance or approval of other agencies such as the FDA in the United States or the
EMA in Europe (Carrera-Rivera et al., 2020). This includes furnishing technical documentation
and other clinical proof of the product's performance.
11
4. Cybersecurity: Due to vulnerability of SaMD to cybersecurity threats, regulatory
bodies enforce strict cybersecurity controls. It means that the documents have to show that
certain security threats, and their proper control, have been addressed.
Examples of Automated Documentation Tools Used in SaMD
Automated documentation tools are beneficial in increasing the productivity and quality
of SaMD development documentation. Examples of these tools include: Examples of these tools
include:
1. JIRA and Confluence: These Atlassian products are adopted for documentation
management and issue tracking across the entire software life cycle. They allow for action
coordination in the actual work and keep records of all the developmental processes.
2. Polarion ALM: Polarion Application Lifecycle Management (ALM) offers
requirement, quality and compliance in one system. This way it guarantees that all the
documented documents are traceable and follow the instructed regulatory frameworks.
3. Zephyr: While test management tool remains connected to JIRA and other
development tools, it is easy to create test documentation on the tool as well as to make sure each
test case is traceable.
4. Git and GitHub: Transient and version control systems like Git and platforms like
GitHub are very important when it comes to managing code changes or having a record of the
modifications that have gone through a certain piece of code. This also helps in ensuring that all
the changes are recorded and authorized to meet the traceability of changes as prescribed by the
laws.
In conclusion, SaMD development and regulation processes should also have record
keeping in order to be safe, effective as well as being in compliance with the set laws. The
12
documentation process is also facilitated by automated tools so that the record of development is
quite clear and traceable. Thus, through strict documentation and compliance to regulatory
requirements, developers can be able to obtain market authorization and continued safety in
clinical utilization of SaMD products.
Challenges and Solutions in Automated Documentation
Technical challenges
The generation of automated documentation also has some technical and regulatory
issues. Some of the challenges on the technical side include; data convergence that implies
merging of different sources of data and system compatibility (Tiwari et al., 2021). A number of
large transactions have to be processed and managed, hence the need for sound hardware to
support this kind of data processing. Also, the correctness of the recorded data should be
considered since it can be multiplied in subsequent operations, creating complications.
Regulatory and compliance challenges
The complexity of regulatory and compliance issues also proves to be significant
hurdles. At the same time, automated documentation systems have to meet numerous
regulations, which are specific to certain industries and can be rather intricate and dynamic.
Protection of data is highly important and should be taken seriously at all times, but particularly
in areas of health and money issues. Legal consequences of non-compliance are enforcement of
the law, fines and, in some extreme cases, imprisonment, and harm to the reputation of the
organization.
Proposed solutions and best practices.
In order to manage such challenges, principles and proposed interventions are required
to be implemented. Some compliance risks may include the following; To minimize such risks,
13
one has to ensure that the systems are updated and properly maintained in accordance with the
current regulations (Wang et al., 2021). Applying progressive measures in data encryption and
private access further improve the protection given to the data. Also, in designing of a system, it
is recommended that the developers used the principles of modularity to enhance compatibility
and interconnectivity with source data. One of the approaches to containing errors is to train
personnel to be in charge of automated systems since they are more reliable.
Challenges and Solutions in Automated Documentation and Emerging Trends in
Automated Documentation
Current and future trends in this field are the use of natural language processing and
machine learning in the process of automatized documentation. This presents one trend: the
application of AI tools that can create rich and context-driven documentation from primary data.
These tools rely on the use of complex algorithms to interpret information and eliminate
intensive human work, hence increasing the precision of tools. Another emerging trend is the
technology that supports simultaneous working on and editing documentation by the number of
users with no time lag to guarantee updated records. Moreover, there is an appearance of
blockchain technology for preserving, and verifying documents’ data originality, which offers
definite protection for sensitive and crucial documents.
There are some potential areas that future research in automated documentation could
focus on. Thus, the first identified issue to resolve is connected to the AI models used and the
necessity to enhance the models to understand the intricacies of domain-specific language
employed in the documentation. Secondly, research could be directed on improving the methods
for designing friendly user interfaces of automated documentation tools for use by ordinary
users. Third, investigations about the moral issues as well as data protection issues concerning
14
automated writing are important and especially in the sectors dealing with personal data like the
health and financial sectors (Barbosa et al., 2023) Finally, exploring how the process of
automated documentation is connected with other business processes and information systems
including the customer relations management (CRM) and enterprise resources planning (ERP)
systems can also result in other benefits and effective practices.
In conclusion, Automated documentation generation is one of the major advancements
in software engineering the most important for SaMD. This raises up efficiency, accuracy,
reduced propensity to violation of strict regulatory standards, and as well descent compliance.
There are obstacles of a technical and legal nature; however, optimizing practices and using high
technology can sufficiently counterbalance these problems. Other new trends such as AI tool,
real-time, and integration of blockchain also forecast better performance. More research should
be conducted on models from a specific domain, the improvement of the interfaces of AI,
tackling the issues related to the ethics of the AI products to be used, and the integration of the
AI technology into more domains. These steps shall help to guarantee that automated
documentation stays a secure, reliable method in a growingly intricate technological world.
References
Carrera-Rivera, A., Ochoa, W., Larrinaga, F., & Lasa, G. (2022). How-to conduct a systematic
literature review: A quick guide for computer science research.Faculty of Engineering,
Mondragon University; Design Innovation Center (DBZ), Mondragon University.
https://doi.org/10.1016/j.mechatronics.2022.102567
15
Wang, D., Wang, A. Y., Drozdal, J., Muller, M., Park, S., Weisz, J. D., Liu, X., Wu, L., &
Dugan, C. (2021). Themisto: Towards Automated Documentation Generation in
Computational Notebooks. arXiv preprint arXiv:2102.12592.
Tiwari, P., Tiwari, R., Srivastava, A., & Khanna, R. (2021). Automated Documentation
Generation Tools: A Comparative Study. International Journal of Advanced Computer
Science and Applications, 12(3).
Barbosa, S., Silva, R., & Pereira, J. (2023). Enhancing Software Documentation through
Automation: Tools and Techniques. Journal of Software Engineering Research and
Development, 11(2), 56-70.
