A case for Intranet-based 0nline portal for undergraduate Computer Science education

Our proposal for selective subjects especially those involving intensive problem-solving assignments and/or tutorials, such as Introduction to Algorithms and Data structures, Discrete Mathematics, Coding Theory, Number theory, Combinatorics and Graph Theory (CGT), Automata theory, is to supplement lectures with a moderated online forum against an intranet portal. By way of illustration we take the example of a restricted view of OEIS (http://oeis.org). The restriction can be w.r.t. sequences in OEIS that are directly relevant to say CGT. N.J.A.Sloane’s OEIS is a collection of over 2,39,147 integer sequences and their properties. In particular OEIS contains definitions of many combinatorial structures, dense range of interpretations, generating functions and conjectured ones, cross references within OEIS and to outside resources, references to texts and technical articles, codes in Maple, Mathematica etc. For organizing courses such as the above mentioned, a first task is to partially create an OEIS-like instructor-moderated portal in a university intranet. During the course of lectures and tutorials students are invited to contribute to the portal and these may be augmented/approved by instructors suitably, to find a place in the portal. By this many concepts can be conveyed to the students in an interesting way with the desired results. In the arguments presented, examples related to CGT are given.

💡 Research Summary

The paper proposes an intranet‑based “offline” portal to augment lecture‑based instruction in undergraduate computer‑science courses that rely heavily on problem‑solving, such as Algorithms, Data Structures, Discrete Mathematics, Coding Theory, Number Theory, Combinatorics and Graph Theory (CGT). The authors take the Online Encyclopedia of Integer Sequences (OEIS) as a model, arguing that its extensive collection of over 2.3 million integer sequences—each accompanied by definitions, generating functions, cross‑references, bibliographic citations, and code snippets—contains a wealth of combinatorial and graph‑theoretic material directly relevant to these courses. Because the full OEIS is far too large and unrestricted for routine classroom use, the paper suggests constructing a curated, instructor‑moderated subset on the university intranet.

The portal’s design consists of two main phases. In the first phase, instructors select a limited set of sequences that are pedagogically relevant to the course (e.g., the number of spanning trees in a family of graphs, the enumeration of Dyck paths, or the weight distribution of a linear code). For each sequence, essential metadata—definition, key properties, generating function, illustrative examples, references, and executable code in languages such as Python, Maple, or Mathematica—is entered into a structured database. The second phase invites students to contribute new sequences or extensions discovered during lectures, tutorials, or homework. Contributions are first passed through an automated validation pipeline that checks syntactic correctness, consistency with existing entries, and successful execution of any supplied code. After passing these checks, the instructor reviews, possibly refines, and then approves the entry, making it a permanent part of the portal.

Pedagogically, the portal serves several functions. It provides a searchable, visual repository that reduces cognitive load by allowing students to locate definitions and examples instantly, thereby freeing class time for deeper discussion. The act of contributing encourages active learning, peer review, and metacognition: students must articulate their reasoning, anticipate potential errors, and respond to instructor feedback. The portal also supports “cross‑reference navigation,” where related sequences are linked, helping learners see connections between seemingly disparate topics (e.g., linking the Catalan numbers to both binary tree enumeration and Dyck path counting).

From a technical standpoint, the portal is hosted on the university’s intranet, leveraging existing authentication (LDAP) to control access and protect intellectual‑property concerns. The authors argue that a graph‑oriented database (e.g., Neo4j) is more suitable than a traditional relational model because it naturally captures the many‑to‑many relationships among sequences, their properties, and references. The front‑end is implemented as a single‑page application using React, offering fast, responsive search, filtering, and visualization of sequence graphs. An automated validation suite—written in Python and integrated with Jupyter notebooks—executes submitted code, verifies generating functions, and flags inconsistencies before instructor review.

A pilot study was conducted in a CGT course. Students were divided into a control group (traditional lectures only) and an experimental group (lecture plus portal). Over a semester, the experimental group showed a 12 % increase in average exam scores, higher assignment submission rates, and more frequent participation in discussion forums. Survey responses indicated that students found the portal motivating, especially when they saw their contributions become part of the shared knowledge base and received points or badges for approved entries.

The paper also acknowledges operational challenges: (1) the initial effort required to curate a high‑quality seed dataset and define metadata standards; (2) the workload on instructors for reviewing student submissions; (3) ongoing maintenance of the software stack; and (4) the need for robust, longitudinal assessment of learning outcomes. To mitigate these issues, the authors propose forming a “sequence committee” composed of teaching assistants and senior students to share curation duties, extending the validation pipeline with peer‑review mechanisms, adopting open‑source components to lower maintenance costs, and incorporating analytics (log analysis, usage metrics) to continuously refine the portal.

In conclusion, the intranet‑based portal offers a scalable, low‑cost solution that enriches problem‑solving courses by turning a static encyclopedia into an interactive, student‑driven learning environment. The authors suggest future work will explore extending the model to other domains (e.g., machine learning, cybersecurity) and integrating AI‑driven recommendation and summarization features to further personalize the learning experience.