Scott McMahan

Contents

Introduction

About This Site

SoftBase Systems

• Overview
• IBM Stack Setup
• Web-Based VTOC Report
• Java-based Test Framework
• 64-bit Software Development
• Websphere MQ
• .NET Research Project
• Hibernate Persistence Research
• Web Interface For Mainframe App
• XMI Backup Solution
• WORKOUT Test Suite
• Precompiler Re-launch
• Web Product Delivery
• Secure Web Site
• Automated Product Bundling
• Company Backup System
• Mainframe Systems Programming
• Flex-ES Systems Programming
• Mainframe Development/Testing
• Data Masking Tool
• Internet Presence/Administration
• Checkpoint-Restart Web Monitor
• Real-time Data Management Tool
• Java Checkpoint-Restart Tool
• Cross-Platform Checkpoint-Restart API
• Web-Enabled Customer Management
• In-house Phone Report Tool
• In-house License Tool

• Overview
• Channel Intelligence and Miva 5
• Quote Automation System
• Online Used Car Search
• User Validation Module

• Computer Answers

• Education

• Automating Windows With Perl
• Unix Administration
• Improvement
• RPM Planning With Software
• Languages
• Databases
• OS Environments
• Development Tools
• Web Related
• Low-level Programming
• The B.D. McKay Papers
• LaTeX Typesetting
• Media Upload Client
• Web Services For Media
• Open Source Contributions
• History of Computing
• Subversion (SVN)

• Building Computers
• Bare Metal Setup
• Handyman

SoftBase Systems, Inc (1994 - present)

.NET Research Project

After not having developed for the Windows platform since the Visual C++ 6.0 and Windows 2000 days (except I wanted to learn Mono and C#, but never found time), I was given a 64-bit Windows workstation, a copy of Visual Studio, a copy of DB2, a lot of systems programming books, a C# book, and asked to do a research project that involved understanding how various aspects of Windows worked and what hook points were available. I can't get into exactly what this research was about, because it is a company secret and I can't blab about it on my web site.

I learned how the Windows kernel and APIs related to the .NET runtime, and where hook points were available to call native code. I learned about data access paths like ADO.NET, and what was available for object-relational mapping and persistence (mainly Hibernate and LINQ). I also looked into the future to see how the new Entity Framework was changing how databases were accessed. All of this research gave me an in-depth understanding of how all the parts of modern Windows came together and played different roles to support the development of business applications.

Since I last looked at Windows as a development platform, a lot had changed (Borland was no longer a player with their C++ compiler, which I had always liked better than Visual C++ in the past), but a lot had not changed (Richter and Hart still had their systems programming books, although vastly updated; Richter's books are still a treasure chest of information for systems programmers). The introduction of the .NET runtime presented a whole new abstraction layer and set of system services. The sheer bulk of all the technologies that are now available to developers on Windows means that a general knowledge of how it all works is harder than ever to maintain. Windows is now getting so complex that specialization is almost unavoidable.

I also investigated the feasibility of porting an already quite portable C code base to 64-bit Windows. This code was ported to 64-bit Solaris sometime in the early 2000s, and was recompiled on 64-bit Linux with almost no changes. The code had once run on 32-bit Windows, but had not been recompiled there in a considerable number of years. Ironically, the switch from Borland C++ to Visual C++ seemed to be the biggest issue, not portability to 64-bit Windows.

Live site statistics