Moving to Debian: Fun with NdisWrapper

Oct 20
2009

This post is meant to provide some practical guidance on a problem I think I solved that others may be experiencing.  The problem concerns use of the NdisWrapper on Linux.  If that is what you’re interested in, then please proceed directly to the section labeled Guidance.  However, I first want to take a step back and explain how I arrived at such a technical topic.

My first jobs were pretty technical.  I ran wire to connect a school systems administrative buildings to a network and then spent some time building custom PCs for a small business.  I’ve had a desire recently to get back into the technology with my hands.  For the past few years I’ve wanted to buy some parts from newegg.com and put together a machine.  I finally decided to do it.  I put together some parts for under $300 and the little box with a Celeron Conroe chip hums nicely.  The process of getting that hum going required some help and the exercise reminded me how nice it is to have friends who can help.

My buddy Al gave me some of the less exciting hardware parts and helped give me the confidence to get started.  One of the parts he gave me was a wireless PCI card, which I needed for this new machine as I don’t have any wire run in the house and I wasn’t interested in doing that right now.  I decided to put Debian on the machine after first trying CentOS — CentOS seemed great, but I’ll keep it in mind for a dedicated server.

I installed Debian Linux 5.03 (Lenny) from a DVD.  The Zonet ZEW1602A wireless network card Al gave me wasn’t automatically dedicated on the install.  I dug around a bit and discovered the NdisWrapper how-to guide.  I built the NdisWrapper module using the “from source” instructions in the guide. (That was after trying the default option first to no avail.)  I had to restart the machine to get the wireless network to show up after completing the steps, but the result was a working wireless connection!

I then realized that the main components of the stable distribution didn’t have the latest and greatest packages.  Having run Windows for the longest time, I wanted the latest software I could put my hands on.  I came across rickh’s how-to guide on setting up and maintaining a mixed testing/unstable system.  This seemed scary, but it was the plunge I needed to take.  I dove in and upgraded my stable (Lenny) to the unfamiliar testing (Squeeze) distribution (apt-get update; apt-get dist-upgrade).  After the numerous packages were installed I was left with a computer without a network connection.  Doh!

I must admit that I initially just contemplated reinstalling the stable Lenny release.  But then I began to feel like I was stuck only delaying the problem for the future when I would need to upgrade.  After looking at the error messages for a bit (network manager applet missing required components or something), I realized that my NdisWrapper module wasn’t working even though it appeared it still had the Zonet driver configured.

I went back and rebuilt the NdisWrapper this time using the Debian Package method from the how-to guide (i.e. via module-assistant).  Things worked great after another reboot.

I’m now up and running with the latest apps including skype, tweetdeck, Sun JDK, Flash, and Open Office all running over my wireless network.  I’m really enjoying this Debian experience and I hope this story helps someone.

Selling training: reflection on the MJ #lrnchat

Jun 26
2009

In the #lrnchat last night, there was discussion about “selling” training to management. I saw this morning that there is an active discussion in the ASTD National LinkedIn group responding to a request for help in presenting on the Return on Investment (ROI) of training. I thought I’d take a step back and reflect.

Training professionals assume that they need to sell training proposals using ROI calculations. While the ROI calculation may feel like a familiar financial concept to managers, it is not a recommended technique for analyzing and evaluating investment alternatives in managerial accounting. This is because it can lead to incorrect investment decisions due to the fact that it ignores the time value of money.

The recommended technique for analyzing investment alternatives is the discounting technique. Two different methods fall under this technique. The first is the net present value (NPV) and the second is the internal rate of return (IRR). The NPV is the Holy Grail as it is theoretically superior to other techniques, but the IRR is easier to compute and interpret.

The typical capital budgeting techniques include: discounting, ROI, payback, and urgency. Payback is perhaps the simplest calculation to interpret: When do I get my money back? However, like ROI, payback ignores the time value of money. The urgency technique simply requires an estimation of potential costs, which are usually the focus of training interventions (e.g., compliance training) — stop the financial bleeding. An important observation is that all the capital budgeting techniques except the urgency technique require forecasts of future cash flows.

Beyond the urgency perspective

If a manager requests an ROI, this is essentially a request for a quantified cost / benefit analysis. If you can’t convert the benefits into a quantitative form, then another representation for those benefits is probably required that accounts for the intangibles (i.e., stories, examples, etc.).

If you want to propose a new technology project that will require a capital outlay, then a NPV calculation using discounted cash flows is recommended. These cash flows could be increased sales (easiest to interpret) or more likely expressions of productivity improvement (as is typical of technology). Quantifying these productivity increases is important for capital decision making, but expressing these increases in ways that users understand is perhaps more important (i.e., WIIFM). How will the new technology not add to the work the user already does? How will it save them time? Ideally, it would save enough time that some of that time could be given back to users while some of it could be reinvested by management into greater increased earnings.

From ideas to implementation plans

All of this quantified talk is really only important assuming a solid understanding of the problem that includes clear distinctions of its symptoms. The training project or performance improvement technology should be a recommended solution to the identified problem or opportunity. This is the first step. The typical strength, weakness, opportunity, and threat (SWOT) analysis ideally comes next. If the idea seems worthwhile based on the SWOT, a cost/benefit analysis is necessary to ground the idea. If it is still worth pursuing and involves significant capital, then a net present value based on discounted cash flows should be calculated. That is my theoretical model of how an idea turns into plan that is ready to be implemented.

Blogroll

A few highly recommended friends...

Archives

All entries, chronologically...

Pages List

General info about this blog...