قرآن مجید کی تلاوت، نقل و حمل اور خط وطباعت میں طہارت کی شرط کا فقہی تجزیہ

The Islamic scholars have identified several rules which deal with the reciting and reading of the Quran and touching/handling the al-Mushaf: the script of the Quran. One of the most important prerequisites, as viewed by the classical Sunni schools of Islamic law, is taharat, which includes both the physical cleanliness—by ablution (wudu) or complete body wash (ghusl) or both—and the purification of thought from allkinds of disbelief (shirk). By extending the rule of taharat, the mainstream Sunni jurists categorically forbid all non-Muslims from touching the Quran; although they are allowed to touch and read its translation as well as to listen to its recitation. This paper analyzes the views of the mainstream jurists and argues for reappraisal of several aspects of the said condition of taharat for both Muslims and non-Muslims, relying upon those jurists whose views are though different from the mainstream but are more practicable and closer to the objectives of the Islamic Sharia today.

Evaluation of Hidden Markov Model for Malware Behavioral Classification

Malware is a growing threat to computer systems and networks around the world. Ever since the malware construction kits and metamorphic virus generators became easily available, creating and spreading obfuscated malware has become a simple matter. The cyber-security vendors receive thousands of new malware samples everyday for analysis. It has become a challenging task for the malware analysts to identify if a given malware sample is a variant of a known malware or belongs to a new breed altogether. Since making an accurate decision about the nature of an unknown malware sample is crucial for updating of signature databases and propagation of the update to their customers, therefore vendors of cyber-security products need accurate malware classi cation techniques for this purpose. The research community has been active for providing a solution to the above problem, and a number of diverse avenues have been explored such as machine learning, graph theory,nite state machines, etc. Furthermore, many syntactic and semantic aspects of computer programs have been tried out in search of the best aspect that could be used to distinguish between harmful and harmless computer programs, and to di erentiate malware belonging to di erent families. All the proposed approaches have merits and demerits of their own, and the search for a solution that maximizes the classi cation accuracy with minimal computational costs is continued. This dissertation formulates malware classi cation as a sequence classi cation problem, and evaluates a widely used sequence classi cation tool, Hidden Markov Model (HMM), for the task of malware classi cation. HMM has been a method of choice for a broad range of sequential pattern matching applications such as speech analysis, behavior modeling and handwriting recognition to name a few. The dissertation rst proposes and evaluates novel methods of malware classi cation by combining HMM and malware behavioral features, which are attributes frequently used to distinguish between normal and malicious programs and to di erentiate x among malware families. As an another major contribution, the dissertationlls a signi cant research gap by studying the role of an important HMM parameter, the number of hidden states, in malware classi cation applications. Based on observations from comprehensive experiments conducted on a large and diverse dataset consisting of malware behavioral reports, the dissertation concludes that although HMM shows encouraging results when used for malware classi cation tasks, its potential from a practical standpoint is fairly limited. The dissertation makes the third contribution by proposing to replace the HMM component of malware classi cation method with Markov Chain Model (MCM), and performing comparative evaluation between the two models. Results of the comparison prove that classi cation performance achieved by HMM can be attained much more e - ciently by MCM, and therefore MCM should be preferred over HMM for malware classi cation applications.