Rainbows over Space Virgins camp.

I was purple and wizardly. Purple capes, a purple staff with purple el-wire, purple goggles, and a purple bike with purple fairy wings.

I did know what I was trying to do: simply have both my TypePad hosted blog and my Google Apps hosted e-mail work with maimedleech.com at the same time.

Of course my DNS provider let me put my DNS records into this state without a bit of a warning. And, of course, they must have had the “don’t mess with this unless you know what you’re doing” like message on the form I used to screw things up. But, of course, I didn’t know that I didn’t know what I was doing. After all, I had used MX records and CNAME records before.

This problem was, of course, a common topic in the DNS provider’s support forums. And a problem a web search would quickly identify. But, of course, I didn’t know I should look before I made the change.

And, of course, I didn’t know my e-mail was bouncing. Because not everything was bouncing. Turns out some e-mail systems, hotmail for example, delivered mail just fine.

What I ended up doing to fix things is switching my blog from TypePad to WordPress. When you setup domain mapping on WordPress it takes over your DNS name services AND it supports MX records for Google Apps. Just Google Apps, not anything configurable. But I know it works and I know I’m tired of messing with this for now. I also know it won’t last, that I’ll want to run a home server or something.

Well, at least I did known that you don’t know what you don’t know.

I’m using an ATMEGA32 CPU and a XL-MaxSonar-AE0 sensor. I have some LEDs hooked up to show some diagnostic output and a trimmer to control the voltage level that triggers that output.

When I hook my scope (an inexpensive PC USB Oscilloscope DiSco) up to the analog envelope output of the sensor I get output like this:

The initial pulse occurs at the start of the sensors cycle. The pulse that starts at the green line is generated when the sound emitted by the sensor that bounces off the ceiling (I have the sensor pointed up from my table) makes it back to the sensor.

The output changes to this when I hold a 3/8″ dowel over the sensor:

When I hold a 6″ board over the sensor at the same height, this is the signal I get:

The difference in the voltage level between small objects and large objects is the key to how I want to use the sensor. The idea is that as the area in the view of the sensor fills up with people the voltage level will go up. As it goes up I will trigger more of the lights in the meter to come on.

But, I’ve had so much fun with the projects I built for Burning Life I wanted to do something similar for Burning Man. My current idea is a “Group Hug Meter”. It would use some ultrasonic sensors to read the size and density of a group of people and light up a bar graph as more people join.

As building stuff like this is new to me, especially for as challenging environment as Burning Man,  any advice on the structure design would be very helpful.

I’ll have to do some custom programmable micro controller development to make this work like I want it to. I’ve never done electronics like this but I’m sure I can count on some help from some of my friends who are into that sort of thing. The first challenge will be to see if the sensors work something like I hope they do.

The idea is the use three sensors with increasing beam angles. I’ll read the “analog envelope” from each sensor and compute the area under the graph where the distance is less than 10′ and compare it to the area under the graph where the distance is greater than 10′. If smallest beam angle sensor is “full”, then go on to read the next larger. When it is “full” read the largest.

I plan to update my blog with more details as I make progress…. Wish me luck!

I'm on the "Global Retail Catalog Systems" platform team. We process all the data Amazon gets from vendors about the products Amazon sells and use it to generate the catalog entries you see on the Amazon web site. Amazon sells a lot of stuff and it is changing all the time, so this can be quite a challenge at times.

Why leave Microsoft? Basically, just had had enough of that particular kind of BS.

In fact programmability is so important to OpenSim that there are at least five distinct ways to do it: DotNetEngine, XEngine, MRM Scripts, MRM Loader, and Region Modules. Since I just added MRM Loader as a new project in the OpenSim forge, I thought I would explain how I think each of these relate to each other and why OpenSim probably needs at least four of the five.

This Venn diagram captures some of the key aspects of each option:

First, you'll note that I put DotNetEngine and XEgine together. They are both focused on implementing scripting in a way that is compatible with the Linden Scripting Language (LSL) used by SecondLife. I don't understand why there are two engines with such similar functionality. Nor is it clear to me which one is better for any particular scenario. There is some history to this issue. (I don't want to restart endless debate, but does OpenSim really need both of these?)

In any case, LSL compatible scripting is important in that they allow scripts and knowledge to be directly ported from SecondLife into OpenSim. That helps increase OpenSim adoption.

However, both DotNetEngine and XEngine suffer from their LSL roots. LSL wasn't terribly well thought out, the APIs don't follow any conventions consistently, and it bears little resemblance to any widely used programming language or environment.

That is where MRM Scripts come in. MRM (which stands for Mini Region Module) makes no attempt to be compatible with LSL. It provides a well designed object based API and attempts to move beyond LSL's limitations in other ways.

The LSL based engines and MRM script are all focused on "in world" scripting. This is great for simple things, novice content creators, and quick tests or experiments, for example.

But the scripting approach has its limitations. First, you typically edit the script code using the feature poor editor built into the client or you repeatedly cut and paste between editors. Second, you have no ability to organize your code into independent reusable components. And third, you can't debug your code. All of these combine together to make developing complex functionality very frustrating, especially for experienced developers.

MRM Loader takes the same API as MRM Scripts, but makes it available in a professional programming environment. You get to use your favorite CLR language in your favorite editor so you can use features like API auto complete, code navigation, and refractoring support. You build your code into a standard CLR assembly, and you can reference other assemblies as necessary. Finally, you can step through your code using your standard debugger.

Region modules, like MRM Loader, give you a professional programming environment. Unlike MRM, they also give you full access to all of OpenSim's internals (in fact most of OpenSim is implemented using region modules). This makes them extremely powerful but also more complex. Region modules are also more likely to break as OpenSim's implemention evolves, while the MRM API is expected to be more stable. In the end, region modules are well suited for extending OpenSim itself but not well suited for adding behavior to virtual world content.

So we have the following reasons for being:

• DotNetEngine or XEngine – LSL based compatibility
• MRM Script – "modern" in world scripting
• MRM Loader – professional content development
• Region Module – extending OpenSim

Each has a place in the OpenSim programmability picture. MRM Loader isn't intended to replace any of these, but to round out OpenSim's programmability story in a way that will appeal to the professional content creator. I'll go into more detail on how I expect MRM Loader accomplish this in a future post.

At least we got moved, not thrown out. I do realize that things could have easily been much worse, and appreciate efforts made on our behalf. My close call allows me to empathize a little what others have been going though, both inside and outside Microsoft.

But, I still need to sort out the impact this change is having on me and those around me.

Windows Mobile has had it pretty rough recently. I used to carry a Windows Mobile phone and now I carry an iPhone. I spend time thinking about how and why that happened. I understand that having a viable mobile platform is strategic to Microsoft.

The challenge is certainly there, but this isn't the battle I picked. Can I get excited about it? Can I do an honest day's work without being excited about it? Since I'm being forced out of my comfort zone, what other opportunities are available or to be created? What impact will following any of these paths have on my family, my career, and myself?

I've spent a week now trying to wrap my head around this change. I've written and re-written this post many times. There just aren't tangible answers to questions like these.

Guess I'll go to work.

I would love be involved with OpenSim, an open source virtual world server project. I want to fix bugs and add features so that I can do cool stuff with Eye Ree on osgrid.org.

But I work for Microsoft and Microsoft has rules and procedures for using or contributing to an open source project. Unfortunately they seem to apply even to what I do on my own time and even when my job doesn’t currently have anything to do with the open source projects I would work on.

I will work the process, but it takes time and involves both lawyer and executive types. These people are responsible for protecting Microsoft’s IP. My interests and desires aren’t really part of their equations. It is much easier, and safer, for them to say “no” than to say “yes.” So in the end I don’t expect a positive outcome.

I could find a new job, or hide my activity, or hope nobody notices or cares. But I’ve got a family and responsibilities. I’ve invested a lot in my career at Microsoft. Is being able to hack on OpenSim worth the disruption or risk? I suppose not, but should I really have to make a choice between a job and a hobby just because both involve software?

I can go over all the arguments on all the sides of the big picture issues involved here, and I am becoming more involved with internal efforts to foster change. But none of that lets me hack away on OpenSim without reservation.

As I said, I’m frustrated.

UPDATE: A few weeks ago the lawyers said this qualified as moonlighting and is covered by that section of the HR rule book. So, since I don’t see how my helping out with OpenSim violates those rules, I did this.

One of the things I still need to figure out is how I can best put a public region or two online while being able to run custom builds. As this is just a hobby, the solution has to be economical. Still, I want to be able to support some large builds with a number simultaneous visitors.

Most likely I'll get a static IP FIOS connection and buy/scrounge a decent server machine. However, hosting in the cloud has some advantages, if I can find a cheep enough option.

Stuff like that needs to “just work”. We have a very long way to go….

An iframe tag:

Created with Admarket’s flickrSLiDR.

An object tag:

Created with Admarket’s flickrSLiDR.

HTML Badge:

www.flickr.com Eye Ree’s Vivid Purple photoset

Flash Badge:

An embed tag:

An SWFObject:

An objec tag: