Joseph E. Lang

Effective masses for nonparabolic bands in p‐type silicon

The nonparabolic nature of the valence bands of silicon gives rise to an explicit temperature dependence in the density of states effective mass as defined through the carrier concentration as well as an energy dependence in the mass defined through the density of states itself. Both masses are important to those researchers interested in fitting electronic transport data. In the past the temperature dependent effective mass of Barber has been most commonly used. The approximations that have gone into the Barber mass are somewhat oversimplified in that he only approximated the actual band structure. We present a more complete calculation, and compare it with the results of Barber for nondegenerate p‐type silicon. There is a significant difference, a factor of about 1.4, between our results and his in the temperature range 300–600 K. The difference is quite noticeable in calculating electronic transport quantities such as carrier concentration or conductivity mobility.

Temperature dependent density of states effective mass in nonparabolic p‐type silicon

Previously, an exact calculation was made for the temperature dependent density of states effective mass in p‐type silicon. This calculation was made for the nondegenerate regime and included the full nonspherical‐nonparabolic nature of the valence band structure. For those researchers interested in using this mass for the analysis of their transport data we offer a polynomial fitted expression. The values generated are within 1% of the theoretical values.

Characterization of the intrinsic double acceptor in undoped p‐type gallium arsenide

A detailed study of the dominant double acceptor found in p‐type, undoped GaAs grown by the liquid‐encapsulated Czochralski technique has been performed using temperature‐dependent Hall effect, infrared absorption, spectral photoconductivity, and photoluminescence. Improved fitting techniques are presented for analyzing the temperature‐dependent Hall‐effect data. A dependence of the activation energy of the neutral state of the double acceptor as determined by Hall effect on the concentration of the defect has been observed and is discussed in terms of prevalent theories. The ionization energy of this level as determined by all three optical techniques is constant and independent of any concentration. Additional data on the correlation between the concentration of the double acceptor and the boron concentration in the material is presented. A model for this correlation is presented that is consistent with the isolated gallium antisite model for the double acceptor. Aluminum and indium alloyed samples were...

Scheduled supervised laboratories in CS1: a comparative analysis

Retention in our Computer Science I course has been around 50%. In order to improve retention, we started using closed labs in the course. In this paper, we will compare two offerings of our Computer Science I course with notfor-credit closed labs, against two most recent offerings of the course with open labs. We found that closed labs did not help improve retention or project completion rates in the course. However, we observed a qualitative improvement in student learning in the closed lab sections. Our observations suggest that closed labs may be helpful primarily in the first third of the course. We plan to continue to collect and evaluate data to verify these results.

Ada in CS1

As the use of Ada in commercial programming increases, it becomes more important to make an attempt to introduce it into the curriculum as early as possible. We have taught CS1 successfully using Ada by strictly concentrating on a subset of the language and through the use of a student-oriented package which enables students to begin writing programs in Ada after one lecture. A laboratory of personal computers was found to be adequate for CS1 use and students were able to write up to 2 programs per week in the course.

Object-oriented programming and design patterns

When teaching object-oriented concepts, design patterns are usually left until the end or left out entirely. In this paper, design patterns are related to the rules of object-oriented design, strongly suggesting that these two concepts should be taught together as beneficial for the understanding of both.

Teaching the abstract data type in CS2

Teaching the abstract data type in CS2 is made difficult by the fact that the topic is intertwined with issues of language support, dynamic data structures and implementation techniques for dynamic data structures. When we switched to Ada to teach CS2, details of the language support for data abstraction caused us to restructure the CS2 course. By pushing the topic of the abstract data type toward the beginning of the course, we have found that it is covered more successfully.

Grammars for FFP

In Backuss original paper on FP and FFP [1], the syntax of these language is described informally. Though a formal syntax for Berkeley FP has appeared in print [2], no formal representation of FFP syntax has yet appeared. It is desirable to have such a formal representation for implementation purposes, and such a representation may also be useful for specifying where the syntax of FFP ends and its semantics begins. In this paper, we will write down formal grammars for FFP based on the syntax as given by Backus [1] and Williams [3] and will discuss their semantic implications. A third grammar similar to that for FP is presented and it may be used for implementing the language.

An interpreter for Backus's formal functional programming language

Backuss formal functional programming (FFP) language [1] is an extension of his functional programming (FP) language [1] that allows the user to define functional forms. Like FP, it has the property that there are to be no side effects during execution. Unlike FP, it has the advantage that there is no distinction between data and functions: data may be applied (interpreted as a function) to data and functions may operate on functions (interpreted as data). This extends the expressive power of FP by giving the user the ability to create and modify functions and functional forms at will. Furthermore, FFP contains a predefined function APPLY which turns a sequence of two elements into a function application. Such a function would not be possible in FP because of the fundamental distinction between functions and data in FP.