Efficient interaction between PHP (Hypertext Preprocessor) applications and MySQL databases is essential for the performance of modern web systems. This study systematically compares the two most widely used PHP APIs for working with MySQL databases—MySQLi (MySQL Improved extension) and PDO (PHP Data Objects)—under identical experimental conditions. The analysis covers execution time, memory consumption, and the stability and variability of results across different types of SQL (Structured Query Language) queries (simple queries, complex JOIN, GROUP BY/HAVING). A specialized benchmarking tool was developed to collect detailed metrics over several hundred repetitions and to enable graphical and statistical evaluation. Across the full benchmark suite, MySQLi exhibits the lowest mean wall-clock execution time on average (≈15% overall). However, under higher query complexity and in certain connection-handling regimes, PDO prepared statement modes provide competitive latency with improved predictability. These results should be interpreted as context-aware rankings for the tested single-host environment and workload design, and as a reusable benchmarking framework intended for replication under alternative deployment models. Statistical analysis (Kruskal–Wallis and Mann–Whitney tests) confirms significant differences between the methods, while Box-plots and histograms visualize deviations and the presence of outliers. Unlike earlier studies, this work provides a controlled and replicable benchmarking environment that tests both MySQLi and PDO across multiple API modes and isolates the impact of native versus emulated prepared statements. It also evaluates performance under complex-query workloads that reflect typical reporting and analytics patterns on the ClassicModels schema. To our knowledge, no previous study has analyzed these factors jointly or provided a reusable tool enabling transparent comparison across PHP–MySQL access layers. The findings provide empirical evidence and practical guidelines for choosing the optimal API depending on the application scenario, as well as a tool that can be applied for further testing in various web environments.

Precise time interval generation is a cornerstone of modern measurement, automation, and distributed control systems, particularly within Internet of Things (IoT) architectures. This paper presents the design, implementation, and evaluation of a low-cost and high-precision time interval generator based on Complementary Metal-Oxide Semiconductor (CMOS) logic counters (Integrated Circuit (IC) IC 7493 and IC 4017) and inverter-based crystal oscillators (IC 74LS04). The proposed system enables frequency division from 1 MHz down to 1 Hz through a cascade of binary and Johnson counters, enhanced with digitally controlled multiplexers for output signal selection. Unlike conventional timing systems relying on expensive Field-Programmable Gate Array (FPGA) or Global Navigation Satellite System (GNSS)-based synchronization, this approach offers a robust, locally controlled reference clock suitable for IoT nodes without network access. The hardware is integrated with Arduino and ESP32 microcontrollers via General-Purpose Input/Output (GPIO) level interfacing, supporting real-time timestamping, deterministic task execution, and microsecond-level synchronization. The system was validated through Python-based simulations incorporating Gaussian jitter models, as well as real-time experimental measurements using Arduino’s micros() function. Results demonstrated stable pulse generation with timing deviations consistently below ±3 µs across various frequency modes. A comparative analysis confirms the advantages of this CMOS-based timing solution over Real-Time Clock (RTC), Network Time Protocol (NTP), and Global Positioning System (GPS)-based methods in terms of local autonomy, cost, and integration simplicity. This work provides a practical and scalable time reference architecture for educational, industrial, and distributed applications, establishing a new bridge between classical digital circuit design and modern Internet of Things (IoT) timing requirements.

Numerical simulations and checks of face recognition software on given image databases represent a type of empirical research. Face recognition software works on the principle of comparing a photo of the person’s face with the photos in the database. The operation of face recognition software can be divided into three stages. The first stage is face detection, the second stage is face tracking and the third stage is face recognition. For this purpose, software solutions have been developed, with different work techniques. However, it is characteristic that regardless of the different techniques, each expresses its effect with a probability expressed in percentages. Simply put, for now, no software solution can be said to be 100% effective. For now, no computer solution can be compared to the human ability to recognize and identify a person.

Security of information systems has become a key component in modern technologies, given the constant threats that come from cyber attacks. This paper explores the ways in which the security of information systems can be improved through the implementation of security measures in the program code. The focus is on one of the most trusted programming languages, Python, which is known for its flexibility and wide application in security operations. Key techniques such as encryption, vulnerability detection and protection against malware attacks are analyzed , along with relevant code examples.

U ovom radu bavimo se nuđenjem usluga koje bi inteligentna javna uprava mogla nuditi svim korisnicima određenih usluga. Ovdje predlažemo primjenu elektroničkih i mobilnih usluga u javnoj upravi te korištenje nekih novih koncepata koji mogu podržati t...

This paper explores the application of programming code in computer network forensics, with a focus on methods for collecting, analyzing and preserving network data. We discuss various techniques and tools for detecting network threats, monitoring network traffic, and identifying anomalies that may indicate security incidents. The goal is to show how programming code can improve the efficiency of forensic investigations and enable more accurate analyses.