Thursday, December 23, 2010

Information Diet Planning - Part 1


Goal Setting

I’ve let it slip that one personal goal for 2011 is a more regimented information diet. I’m becoming increasingly convinced that the concepts we apply to diets on the body can be applied with some parallel applications in the world of information.

Calorie Counting

The first thing that applies to physical diets is the concept of tracking intake. Our bodies can utilize up to a certain amount of food after which, no matter how good it is, the food is going to be stored up as fat. Is tracking intake something that can be applied to information? The number of sites visited in a day, the number of browser tabs open, the amount of time getting pumped with information via some form of media: podcast, radio, screencast, television or otherwise? Is there a point where that additional reading does no good and takes away from what might have been retained?


The second basic thing in the world of physical diets is some concomitant form of exercise. Some of this might be for burning away calories (cardio) but sometimes it’s about gaining mass or turning “fat” into “muscle.” I wonder what exercise looks like in the world of information. Steve Yegge had an old article about practice for programmers and I suspect that exercise involves designating time specifically for the mental effort related to processing information efficiently.

“The great engineers I know are as good as they are because they practice all the time. People in great physical shape only get that way by working out regularly, and they need to keep it up, or they get out of shape. The same goes for programming and engineering.”

Media and Delivery

There are different “food groups” associated with a healthy diet. Some types of food, like “fatty carbs” are very difficult to incorporate into any meaningful diet but others, like fruit, are a staple of most sensible dieting efforts drawn up for a healthier lifestyle. I wrote down my main sources of intake and looking at the list I would consider some blog entries analogous to fatty carbs or sugar whereas other forms of intake such as technical books to be a more substantive form of media for input and processing.

Confession: I’m an architecture geek1. It started in earnest when I moved to South Dakota – it was the first time I’d lived in a rural area. I missed the built environment I was used to in cities. Along the way I’ve managed to have 14 different architecture blogs in my RSS Reader. Especially since I’m not an architect, this is excessive. I usually enjoy a story here and there but I leave a lot unread. This is a case where it’s not my curiosity that needs to go away, it’s the delivery format. I’d be better served rereading Steen Eiler Rasmussen’s Experiencing Architecture or Bjarke Ingalls’s Yes Is More than getting distracted by blog entries from the web.


I’m not sure there is any equivalent in the world of dieting but the last thing I’ve been thinking about related to my information diet is designating time for processing. Earlier this year I finally made some headway in understanding Getting Things Done and though I can’t say I’ve implemented everything David Allen recommends two things have stuck: his recommendation of using a calendar and the idea that you process information rather than letting it idle in an indeterminate state.

It’s occurred to me that in recent years information access for personal and professional development is trivial. Wikipedia is a great resource for documenting general knowledge. For professional work there are websites, link aggregators, and vendors eagerly providing a glut of information to learn from.

In this environment, what does it mean to “process” an information resource? For example, at work this week we watched Bart De Smet give a talk on a language feature of C# called LINQ. I’ve used LINQ quite a bit but his talk covered some direction in the technology that is new (Rx). In a more general sense, he referenced the Wikipedia entry on Monads – another resource which I would like to process.

I’m not sure what processing should look like but it seems like what it means to process an information resource is less of an issue than the discipline it takes to designate time for it. Perhaps that ties this notion back to exercise.


I’ll be using the next couple of weeks to think more about what an Information Diet should look like and a practical structure to use. I’m posting in part because I would love any input on tactics and also because committing this to the permanence of my blog means the commitment is formalized. Four areas of focus will be:

  • Calorie Counting – tracking input
  • Exercise – deliberate practice
  • Media and Delivery – determining which formats and timing work best for information
  • Processing – what steps are involved in making meaning out of information


1A running joke with my wife is that I will quit my job, enroll in an architecture program just so I can be a critic. She takes great joy in laughing about this.

Monday, December 20, 2010

Language, Programming, Quirks, Conviction; Derek Sivers at RailsConf


Imagine that you are standing on a street, and you are in America, and a Japanese guy comes up to you and says “Excuse me… what is the name of this block?”

And you say “well, I don’t understand… there’s Oak street and there’s Elm street, this is 27th street and that’s 26th…”

And he says: “Yes but what is the name of this block?”

You say: “I don’t understand what you mean; blocks don’t have names, streets have names. Blocks are just the unnamed spaces in between the streets.”

He looks a little disappointed and leaves.

So now imagine that you are standing in Tokyo one day and you’re a little lost and you turn to somebody next to you and you say “Excuse me, what is the name of this street?”

And they say, “That is block 17 and that is block 16.”

And you say “Yeah, but what is the name of this street?”

And they say, “Streets don’t have names; that’s just the space between blocks 16 and 17.”

A great talk from Derek Sivers given at RailsConf earlier this year available on IT Conversations.


Sunday, December 19, 2010

The Future Is Balkanized


“I like HTML5, but I think you're duct taping a horn on a horse and hoping it will become a unicorn.”

Shawn Wildermuth hit the nail on the head with a recent analysis entitled “The Next Application Platform? All of them... ” It would be interesting to see how many more hits the article would have gotten with a title like “XYZ is dead” or “The Death of XYZ,” perhaps something along the lines of “The Death of the One Stop platform.”

A basic summary is that he points to a future in which developers tangle themselves with the following development targets:

  • An HTML5 solution for the web
  • Plugins to extend HTML5 as necessary
  • Desktop/Browser apps for in house/well known customers
  • Apps for mobile/tablets in Objective-C, Java, and Silverlight

The sentiments are corroborated by my own experience, most recently even with a potential customer who wanted to have an iPad friendly application (native) but also wanted to support web based users who had no such device. If I could translate the request, it could simply be that they want things in the best possible experience for each of the disparate platforms for which they had users. Of course the smart phone was no exception.

In a world like this, what could the future portend but that we will have multiple platforms, no singular delivery, and a constantly increasing need to build things quickly, portably?

I’ve cited the article quite a bit but over the years I’ve always gone back to Jonathan Edwards article on Beautiful Code:

“I wish someone had instead warned me that programming is a desperate losing battle against the unconquerable complexity of code, and the treachery of requirements.”

Although Edwards was talking more specifically about code, all of my experience points towards this truth in the realms of the disparate platforms, specifically that getting something to work in a “real world” where people use different operating systems, browsers, displays, and connectivity paradigms. Over the years there have been different tactics in trying to solve the problem: web standards, ActiveX controls, Java, plugins, and of course the latest new craze, HTML5. It’s not that these tactics (and the technologies related to them) are bad: I still advocate web standards to a point, I still have my day job programming Silverlight. It’s just that the pragmatist in me thinks this is the Arab-Israeli conflict in software: no perfect solution no matter how badly its desired by those in each of the platform camps.

The big insight for me as a developer is this line:

This means as a developer you'll need to expand what you know and learn more platforms. Is that bad or good?  Both. It means more work for all of us, but it does mean you'll need to less focus on silo's of platforms and use your knowledge across these platforms.


Monday, December 13, 2010

Advent Calendars for Developers


Since my days trying to hack up some Perl1 I was charmed by the idea of a developer’s advent calendar: a day by day lead up to Christmas with articles pertaining to a specific topic. I’m otherwise a scrooge during the holiday time; not given to commercialism, Christmas music, and other seasonal rituals.

But I am up for the both the Perl Advent Calendar and the RJBS (Perl) Advent Calendar. On a similar vein, Sys Advent, a sys admin advent calendar should be interesting tracking too.

Others have riffed on the idea, most notably web developers. 24ways is a seasonal advent project that I’ve tuned into each year – I usually have at least one or two of the posts stick.

Finally, there is an entrepreneurial advent calendar, 24waystostart. A lot of good things in there, I feel the onset of holiday good cheer.

I’m disappointed not to find a .NET Advent calendar (or even a Python one!). If you know of a good advent calendar, developer or otherwise, feel free to share. A few good articles can make the season pleasurable, even for a grinch like me.


1Could it have been that long ago? Funny to see those old posts, but that's part of why I blog

Thursday, December 02, 2010

Serialization’s “Pit of Success” with C# and JSON


Although it’s been mentioned before here, it bears repeating that before you find yourself making a reference to the DataContractJsonSerializer, you ought to consider and indeed are more likely better off with the NewtonSoft library Json.NET. Whereas the DataContractJsonSerializer relies on giving you the flexibility of creating types and using the DataContract and DataMember attributes on them for a lot of granular control, this is burdensome with types that should naturally find themselves serializable and JSON friendly. For example, let’s say I have a Dictionary<int, string>; the intent of the user shouldn’t be too complicated in the serialization process. Or, if I have a class Foo with a couple of properties, and I want to serialize an IEnumerable of said class, again, the intent doesn’t require the extra use of attributes and a complexity overhead.

This is not to say that the DataContractJsonSerializer does not have its place; if you are leveraging WCF and choosing JSON as your serialization format, it is probably the best route to go. That said, however, I have fallen more than once into a “pit of success” with Json.NET after getting burdened with trying to use the DataContractJsonSerializer.

Json.NET is still quite active, the last release during the latter half of this year. Check it out.


Friday, August 27, 2010

Design Related: Branding with Debbie Millman


I always struggled with a singular identity. Maybe that had something to do with going back and forth from East Africa and the United States, maybe it is because my father had a lot of books and it gave me a curious nature. Whatever it is, I am the kind of person who, when I find myself in a bookstore, picks up a magazine on a topic I know nothing about and enjoys trying to discover the nuances of a not yet explored world of information. I’ve done it with things like pens, model trains, and antiques.

When it comes to design, as much as it may seem by some to be orthogonal to software development, I’ve had the growing understanding that the two go hand in hand. Design, to me, is not picking nice colors for gradients or making icons. Design is planning, it’s understanding a problem and solving it. Although as a software developer my visual design has often been an afterthought (something I’ve fought with increasing intensity over time) and yet when I write a piece of software, put together a user interface, or model data structures, I am planning and problem solving. I am designing, even if I’m making a poor effort of things.

The more aware I’ve become of this, the more I’ve sought to have a baseline understanding of design which is why I found myself at the South Dakota AIGA hearing Debbie Millman give a talk on the history of branding, which was originally prompted by her thinking about the phenomenon of MySpace. I’d long been aware of Debbie; I listened to her Design Matters show when it originally became available online and look forward to a new podcast via Design Observer.

In speeding us to the present, the question of why there may be 300 national brands of cereal or 100 different types of bottled water in a store, Millman began 50,000 years ago with the “big brain bang” where scientists believe we developed our Triune brain. The 3 parts of the brain Millman talked about, Reptilllian, Limbic, and Neocortex are a basis for how she believes that branding works today – if I could shortcut to the conclusion, the reason why we put ourselves in groups and find affiliation important is because this need is hardwired. In one humorous aside, Millman said of the part of the brain that fears the unknown:

“You can’t meditate it away. It’s just there.”

The talk bolstered this point along many fronts, some anthropology, some historical/statistics and other scientific studies and anecdotes. She then divided the history of branding, starting from the US Trademark Registration (landmark legislation that is the legal foundation for branding) in 1876, into 5 different “waves:”

Wave 1: 1875 – 1920

Brands as a guarantor of consistency (Think Campbells soup, Quaker Oats).

Wave 2: 1920 – 1965

Brands as a guarantor of quality. (Think Mortin’s Salt, Pepsi).

Wave 3: 1965 – 1985

Brands as self expressive statements, telegraphing what others should think of us. (Think Nike, Levis, Volskwagen).

Wave 4: 1985 – 2000

Brands as an experience. (Think Starbucks.)

Wave 5: 2000 – present

Limbic brands. (Think Social Networking). In reference to the earlier portion of the talk on brain development, this is the part of our brains which serves us emotion and sense of group. In our present, fractured world, where traditional models of group like family are being uprooted by modernity, we turn to brands to fill that portion of our lives, which is hardwired.

I found the talk riveting, in large part because it aimed at a higher level thinking of how branding works. Afterward I had a chance to talk to Debbie and it sounds like there are new Design Matters shows in the making. I’ll look forward to hearing them and learning more.


Wednesday, August 11, 2010

Failure in two parts, IronRuby


I recently ran across a two part series from Dave Winer related to failure. Essentially Winer asserts the following:

“The only way to succeed in my opinion is if you cannot visualize failure.”

In the first post he describes a personal experience where failure was imminent and he pushed on despite it because he realized his back was to the wall.  In the second he keeps the theme by emphasizing the special level of determination necessary for success. Many of the blog posts I find my way to via Hacker News and other outlets have similar exhortations but Winer holds more gravitas for me because of his success as a founder and longevity as one of the alpha dog programmer/internet personalities.

Both pieces were inspired by the recent news that Google was no longer pursuing Wave – the kind of capitulation that may have inspired Winer’s title: “Embracing failure is a good way to fail.” As an aside, Dare Obasanjo has some thoughtful points on why Wave couldn’t grow.

Not long after I discovered that IronRuby’s future as a Microsoft project was in flux. It seemed to me similar to Wave in the sense that I remember the fanfare when all the Iron* (IronRuby, IronPython, etc… ) projects were beginning and the excitement of people in the programming community only to perk my head up now and find that the project seems no longer to have the good graces of its company. What a contrast.

I speak as an outsider but I suspected that as time passed the level of interest in these projects began to erode: everything from key people leaving teams to diminishing excitement from people on the outside. Most of my interest was directed toward IronPython since it was the most mature dynamic language implementation in the beginning (the whole concept of the Dynamic Language Runtime (DLR) came from Jim Hugunin’s early port of Python). This was both fortunate and unfortunate; I was eager to learn Ruby but since I can’t learn languages very fast I decided to maintain my focus on Python in exclusion of other DLR language implementations. All I can do is suspect but perhaps there were others like me who decided rather than trying to learn different languages simultaneously it would be better to focus on one.

Now as I think upon my encounters with the DLR, I realize that I’m a part of the “failure,” if one can call it that, of Microsoft in valuing this and other dynamic language implementations. Although I was excited early on, even spending some time creating projects and coming up with ideas that leveraged the dynamic languages on the DLR, I got bogged down by how hard it was without a lot of documentation and the difficulty of trying to convince people around me that what I was doing was worth it. (Insert the glazed eyes and “why not just use C#? It has intellisense…” conversation). In the interests of a more full disclosure, a personal tragedy from last year took a lot of wind out of my sails and I didn’t do any programming my spare time for a long stretch of time.

But of course I neither think that IronRuby is a “failure” nor do I think my lack of initiative over the last few years is the way things need to remain. Now seems an even better time to become more active in what community IronRuby has since the absence of Microsoft will create space for people like me, on the outside. I will probably return to the drawing board with a few projects and ideas in IronPython but Bruce Tate’s latest book has inspired me to think that perhaps trying to tackle multiple languages at the same time isn’t so outlandish a goal after all. The important thing I’m going to try to bring with me is the frame of mind I quoted above from Winer: being unable to imagine that IronRuby can go away or lose value as an idea or a platform. 


Thursday, July 29, 2010

Management Perspectives from a Groundling and Dan Pink


Over the weekend I took the time to watch Stan Lee’s Mutants, Monsters and Marvels, a film that consists entirely of an interview between film maker Kevin Smith and Stan Lee. As a kid who grew up loving comics, I loved the unveiling of the world behind the worlds I spent as much time as I could with my imagination: Spiderman, Fantastic Four, Hulk, Iron Man, and the list goes on though I’ll stop before daunting anyone without appreciation for such comic book geekdome.

One of the interesting segments dealt with Stan Lee’s approach as the top editor at Marvel. Although the characters we are familiar with are his creations and he would collaborate with artists on the first few releases of books, subsequently other teams of artists and writers would work on their own releases of the characters he created. When Kevin asked how he managed this he waved his hand and said that he found the best way to manage creatives was to leave them alone: that’s how you would get their best work.

Stan Lee seems to be a man of intuition (in the interview he’s matter of fact and terse, less a man of “Big Ideas” and more of “Common Sense” when it comes to managing) but I think he was getting at the management theory that is being articulating these days by Dan Pink. Pink’s ideas have to do with the connection of motivation and better outcomes that are concomitant with employee autonomy. A summary of his ideas as well as their rational basis is in this entertaining animated video from RSA:

Although I do not have any formal disagreements with Pink’s ideas (they are spot on in my book) I think what is telling is his background and where he worked. Some top notch schools and the upper echelons of government imbue a sense of personal worth and require the freedom and tactics he recommends. The same is true with Stan Lee: his management style worked precisely because of the environment he was in and the nature of his work; in comics being creative and motivated as well as finding ways to help your employees in that regard are the name of the game.

What I would add, from the groundling floor, is that in many places it’s not that these theories don’t make sense or aren’t highly empirical (Pink glosses over some behavioral economics although if you really want to dive into the field Dan Ariely is a better start); it’s that different organizations have different goals. As much of a thought bludgeon as it is, the haunting truth is that some places are perfectly happy trading out creativity, autonomy, and the requisite unpredictability associated with them.


Tuesday, July 27, 2010

XAML Iconography


I have been having a lot of fun lately using Adobe Illustrator and symbol icons to build icons in XAML. Although the typical icon libraries provide very compact image files, the advantage of XAML comes not just in space, but in the ability to manipulate the graphic that you are leveraging.

Here is a brief example of what I mean:

1. There is no shortage of good symbol fonts (or fonts in general, for that matter). I have bought fonts at both and and had great experiences. The symbol font I will use in this case is Energetics.


2. In Illustrator, I create a document and insert the character of choice making it large enough that it will be easily visible when I import it to Blend.


3. Using the File –> Import option from Blend, I have two Adobe related options; in this case we obviously leverage the Import Adobe Illustrator option


4. Voila! We now have our same symbol in Blend but digging in it’s a XAML Path rather than an image. Although it has been imported, however, you will notice a few extra pieces of XAML that you don’t need holding the Path as its container. I usually copy the Path, delete the extras and then paste it back in.


Now that the symbol is in XAML, not only is it not taking up the kind of space that the original image would have, it is also easier to work with if you need to manipulate it programmatically.  Here’s my weight lifter in XAML:

<Path Data="F1M562.59375,264.9375C562.59375,283.96875,560.390625,301.390625,556.015625,317.234375C551.625,
264.9375L562.59375,264.9375 M535.9375,182.734375L512.875,182.734375C508.765625,187.75,505.140625,194.59375,502.015625,
212.203125,525.078125,203.421875C528.203125,194.640625,531.828125,187.75,535.9375,182.734375L535.9375,182.734375 M447.75,
434.421875,302.703125L447.75,262.375 M312.234375,237.59375L309.84375,223.921875L285.40625,224.9375C280.84375,231.671875,
300,253.578125,304.625,250.671875C309.234375,247.765625,311.765625,243.40625,312.234375,237.59375L312.234375,237.59375 M375.296875,
319.578125,364.53125,315.359375C371.703125,311.140625,375.296875,305.90625,375.296875,299.625L375.296875,299.625 M328.125,
348.15625C323.21875,346.453125,328.125,343.78125,328.125,340.125L328.125,340.125 M259.078125,268.6875L189.015625,271.265625L187.65625,
259.078125,268.6875L259.078125,268.6875 M112.453125,200.5L92.796875,200.5C84.703125,210.53125,79.125,226.03125,76.046875,
246.984375L95.703125,246.984375C98.78125,226.03125,104.359375,210.53125,112.453125,200.5L112.453125,200.5 M360.59375,518.71875C360.25,
337.53125,518.71875L360.59375,518.71875" Stretch="Fill" Margin="57,81,0,0" UseLayoutRounding="False" HorizontalAlignment="Left"
Height="131.781" VerticalAlignment="Top" Width="187.405">
<LinearGradientBrush EndPoint="0.5,1" StartPoint="0.5,0">
<GradientStop Color="#FF2D2D2D" Offset="0"/>
<GradientStop Color="#FF3D3C3C" Offset="1"/>



Saturday, July 24, 2010

Excel Interop in C# .NET without COM


Tonight I ran across the excellent albeit unimaginatively called excellibrary project on Google Code and leveraged it in one of my personal projects. It’s nice to be able to get the Excel file format without having to do it in either CSV or with COM Interop.

Reference: Stackoverflow Question / Answer


Thursday, July 22, 2010

TortoiseSVN Extensions Missing on Windows 7


Ran into this issue recently after upgrading my TortoiseSVN shell extensions. To fix, simply run the installer again using the Repair option.

Reference: StackOverflow Question / Answer.