12.27.07
Posted in People at 4:16 pm by Todd
It isn’t often that I receive forwarded email that touches my heart. But I recently received the following story from my sister. And according to Snopes, this story is actually true — which makes the story that much better.
Our 14 year old dog, Abbey, died last month. The day after she died, my 4 year old daughter Meredith was crying and talking about how much she missed Abbey. She asked if we could write a letter to God so that when Abbey got to heaven, God would recognize her. I told her that I thought we could so she dictated these words:
Dear God,
Will you please take care of my dog? She died yesterday and is with you in heaven. I miss her very much. I am happy that you let me have her as my dog even though she got sick.
I hope you will play with her. She likes to play with balls and to swim. I am sending a picture of her so when you see her. You will know that she is my dog. I really miss her.
Love, Meredith.
We put the letter in an envelope with a picture of Abbey and Meredith and addressed it to God/Heaven. We put our return address on it. Then Meredith pasted several stamps on the front of the envelope because she said it would take lots of stamps to get the letter all the way to heaven. That afternoon she dropped it into the letter box at the post office. A few days later, she asked if God had gotten the letter yet. I told her that I thought He had.
Yesterday, there was a package wrapped in gold paper on our front porch addressed, “To Meredith , ” in an unfamiliar hand. Meredith opened it. Inside was a book by Mr. Rogers called, “When a Pet Dies.” Taped to the inside front cover was the letter we had written to God in its opened envelope. On the opposite page was the picture of Abbey & Meredith and this note:
Dear Meredith,
Abbey arrived safely in heaven.
Having the picture was a big help. I recognized Abbey right away.
Abbey isn’t sick anymore. Her spirit is here with me just like it stays in your heart. Abbey loved being your dog. Since we don’t need our bodies in heaven, I don’t have any pockets to keep your picture in, so I am sending it back to you in this little book for you to keep and have something to remember Abbey by.
Thank you for the beautiful letter and thank your mother for helping you write it and sending it to me. What a wonderful mother you have. I picked her especially for you.
I send my blessings every day and remember that I love you very much.
By the way, I’m easy to find. I am wherever there is love.
Love,
God
I hope you enjoyed that story as much as I did … and now go and wipe your eyes and get back to work!
Permalink
12.18.07
Posted in SAnToS at 9:51 am by Todd
I recently finished reading “Generative Programming for Embedded Software: An Industrial Experience Report” by Czarnecki, Bednasch, Unger, and Eisenecker. A very good paper that helps solidify what feature modeling is for me as well as introduce me to what they call generative programming (not a new concept to me).
The main ideas that I got from the paper are:
- Cardinalities are a requirement for real systems. The current notation is inadequate — based upon this case study.
- Attributes are not required but make the diagram less verbose (or more compact).
- Generative Programming can be very effective in the Embedded Systems industry.
- Feature Modeling tool support is adequate but could be better. They go on to describe their tool that should make feature modeling easier. They talk about how they used GME to prototype with as well as to generate the feature model that can be used with their configuration editor.
Some questions that I had while reading that were answered in the paper:
- What is a feature model? “A feature model describes the common and variable features of the products, their dependencies, and supplementary information such as feature descriptions, binding times, priorities, etc.”
- Do features map to components? They say not necessarily. “Features in a feature diagram need not correspond to concrete software modules, but may represent abstract properties such as performance requirements and aspectual properties affecting several software modules such as synchronization.”
Overall, this was a good paper that helped me better understand feature modeling. And I think it is a good paper for our group since we are working on a related area (sensor networks has a lot of similarities to the embedded systems industry).
Permalink
12.17.07
Posted in Manhattan, People at 3:09 pm by Todd
I recently heard about a new micro-brewery here in Manhattan (Kansas). For years, the only local brewery that we had was the Little Apple Brewery (located in Westloop). But now we have a new one, the Tall Grass Brewing Company.
The brewery is the dream of Jeff Gill, a friend of mine. Jeff and I are members of the Theta Xi Fraternity. We met when I was transferring from the University of Missouri-Rolla (Alpha Psi chapter) to K-State (Alpha Iota chapter) since he was the Rush Chairman. He convinced me to live in the house and introduced me to many people that are still my friends today. So thanks Jeff!
And I hope that your brewery is very successful. I just wish I drank beer so I could help your business!
Permalink
12.13.07
Posted in KSU, People at 4:02 pm by Todd
A friend of mine, Dr. Justin Kastner, has been hard at work creating a digital-presence for his research. It is called Frontier and was created along with Dr. Jason Ackleson of New Mexico State.
Justin has done such a great job of embracing technology to convey and enhance his message. He started by creating a great web presence (web site) and has enhanced it with podcasts and RSS feeds. And while the podcasts are not of professional quality, he is constantly improving and making them better. Quite a feat for a non-technical professor at K-State. So check out the website, checkout the podcasts (on their website and through iTunes), and checkout the RSS feeds. All great sources of information, even for those that aren’t experts.
Way to go Justin. Keep up the great work. And hopefully others at K-State will catch on and make use of technology to enhance their own message (at least 1 other professor on campus has done this).
Related Links:
Permalink
12.10.07
Posted in People at 10:41 am by Todd
Over the years I have come in contact with many people in many different walks of life. One of the really cool things is when those people go on to great things and are successful at their chosen path. Over the coming weeks I am going to talk a little bit about some of them. This is partly an exercise in ego (look at me, I know great people) but it is mostly about pointing out some really good people that have found success.
At this point, I am planning on writing about Noah Reagan, Adam Dolezal, Justin Kastner, Ben Dolezal, and Jeff Gill. And if I can, I will continue to do this on a regular basis as I catch-up with old friends, acquaintances, and colleagues.
Permalink
12.07.07
Posted in SAnToS at 10:17 am by Todd
I recently finished reading
“Staged Configuration Through Specialization and Multi-Level Configuration of Feature Models” by Czarnecki, Helsen, and Eisenecker. The ideas presented are still a little over my head but I think I got the basics. Here are the key concepts that I got from the paper:
- Staging is important when developing feature models. I believe this to be true since most software goes through an iterative design process and is shared among many designers. So software that supports designing should make it possible to share, specify, and finalize. Our group also sees it this way and it has influenced how we see development happening in Cadena.
- Cardinality of features is important. I think this is probably the case but I am not familiar enough to argue either side of this debate. But it certainly seems reasonable and necessary. The feature model that I am experimenting with seems to work best with cardinalities (sensor network domain).
While reading I had some questions that got answered so I figured I would write them up here:
- What is a configuration? A configuration is an instance of a feature model. They say “A configuration consists of the features that were selected according to the variability constraints defined by the feature diagram.”
- Why is staging done in two different ways (configuration
v specialization)? They don’t seem to discuss this so here is my take on it. Having the ability to evolve the feature model before instantiating it can be very important. This is specialization (“The specialization process is a transformation process that takes a feature diagram and yields another feature diagram”). Once instantiated, choices must be made which are much different than before instantiation. This is configuration (“The process of deriving a configuration from a feature diagram…”).
- What is a feature? “A feature is a system property that is relevant to some stakeholder and is used to capture commonalities or discriminate amount systems in a family.” My dumbed-down version, something that someone wants in an application. They also specifically mention that their definition is broader than the original definition (Feature-oriented domain analysis (FODA) feasability study) and that features do not necessarily map to software modules.
One annoying thing about their paper, they use the word configuration as a noun and as a verb. This gets to be confusing so I had to re-read some sections to make sure I “got it”.
Overall a good quality paper that introduces the notion of cardinalities as well as staged configuration and specialization for feature modeling.
Permalink
12.06.07
Posted in Business at 9:54 am by Todd
I don’t know what it is but I keep finding new sources of information (websites, blogs and podcasts) to take up my time. Today I started listening to a podcast named Tweak. I have only listened to the first two episodes but they are pretty good. Based on the first episodes, the podcast is about starting a business and finding “tweaks” to keep it going.
Permalink
12.05.07
Posted in SAnToS at 10:44 am by Todd
I started to experiment with the FeaturePlugin for Eclipse (fmp) and found it both useful and annoying.
To start with, it took me quite some time to get installed (several hours of effort). It requires an older version of Eclipse (3.0 or 3.2). I tried to get it working on my current laptop (bluecloud) but failed. I was getting an error (NS_InitEmbedding) when starting up Eclipse 3.2. So I fell back to my old laptop (whitecloud).
/opt/sun-jdk-1.5.0.13/bin/java: symbol lookup error:
/home/tcw/.eclipse/org.eclipse.platform_3.2.0/.../libswt-mozilla-gcc3-gtk-3236.so:
undefined symbol: NS_InitEmbedding
[Note: I replaced some text with ... in the error above to make it fit in my blog.]
I then had a hard-time installing since I forgot to install EMF before installing this plugin. I can’t blame the developers for this error but using an update site makes this error less likely (one of the reasons that Cadena is installed via update site).
Regardless, I finally got it installed and started my first feature model. I chose to model sensor networks (since that is what I am currently working on). I decided to think about it from a “sensor network developer” point of view and think about what capabilities a sensor network might have. So I figured the developer would want to define the types of sensing capabilities (light, temperature, radiation, motion), types of power (battery, ac/dc), and type of communication (wired, wireless). Here is the current model:
I still need to validate this model by passing it around the group as well as trying out some sample configurations (pursuer/evader and hydrology applications that KSUSNS is currently working on). For now, the FeaturePlugin was a little tough to get installed but it does work as advertised. It would be nice if they had provided an update site as well as a more rich UI (instead of the table-tree view from GEF it might look really nice as a more free-form tree).
Permalink
12.04.07
Posted in About Me at 10:48 am by Todd
Every morning I visit the same set of sites and I have been doing this for years (since I was working at Innovision). In many ways, this gets my brain warmed up for the day (kind of like stretching before exercise).
When I started this habit, I simply used a folder in my Firefox bookmarks and clicked on them one by one. Eventually Firefox added a cool feature, I call it “Open All in Tabs”. Basically, there is a menu item in each folder in the Personal Toolbar that allows you to open all links in that folder where each link is in it’s own tab.
Now there is a new way to do it, Morning Coffee. I will be trying out this application over the coming weeks to see if helps me in any way (I seem to have my morning browsing down to an art form).
For more on this you can see the Web Worker Daily page on it. Reading the comments on that page make me feel like less of a weirdo … I never thought others did this exact same thing.
Permalink