techencoder

LINQ classes generated by the VS designer (or sqlmetal.exe) are marked partial by default.

[System.Data.Linq.Mapping.DatabaseAttribute(Name="MyData")]
public partial class MyDataDataContext : System.Data.Linq.DataContext
{
    // blah ... blah ...
}

public partial class MyDataDataContext
{
    // My additional methods, constructors, etc.
}

As I’ve shown above, this allows you to extend a class without modifying the original code. The additional class definitions just live somewhere else, typically in another file, and the compiler integrates them as needed. That’s great for keeping your code out of the machine generated file, but you still need to be aware of what the machine generated code is doing before you extend it! One thing to check is the default (sqlmetal generated) constructor when you are planning to overload it with your own. In some cases, the default constructor will initialize dependencies, so if you write an overload, make sure it calls the default constructor too:

public partial class MyDataDataContext
{
    public MyDataContext() : this() // We must call the designer generated constructor because it contains initialization code.
    {
        // blah ... blah ...
    }
}

Notice how I used the constructor initialization keyword : this().
From the C# Specification Section 10.10.1

Don't forget to call the default constructor!

Happy Coding!

I tried to leave feedback on Soma’s most recent post, but apparently the comment system thought I was long winded.  After being redirected to the home page a few times (without any indication if the comment submission was successful),  I reentered most of my thoughts in multiple short segments.  Well, the end result was not exactly what I wanted, and I do own a blog, so here they are in full with a little extra elaboration. 

So what is so great about stackoverflow?

• It’s CC-wiki (And now a data dump torrent)
• It’s fast. (But they only use a few servers)
• No sub-forums.  (Just tag your questions and everyone (including Google) will find them just fine)
• Reputation earns privileges. (The site gains more administrators the more it is used)
• True OpenID support. (Not a provider-only half implementation)
• No stupid signatures at the bottom of every post.
• No “please mark this as the accepted answer” posts.  (You’d be modded down for that)
• Meta-discussion and clarifications have a clear home. (comments)
• Good answers bubble up, bad ones are pushed down.
• Other (trusted) people can clean up your post.  (Its a thing of good if u speak new to englissh)
• and lastly …

All programming questions are fair game. Today you can ask “how to setup Tomcat 5.5 for ArcIMS 9.2″ and tomorrow you can ask “what’s the big deal about Dependency Injection” after answering a few questions about LINQ and C#.  No one will ever say “Excuse me, you are on the wrong forum and I’m moving (or closing, or ignoring, or down modding) your question. “

Maybe Microsoft shouldn’t try to keep up with stackoverflow.

Come to think of it, maybe Microsoft **shouldn’t** try to keep up with stackoverflow.  I’d much rather have those forums dry up and move to an open site like stackoverflow.  Now that would be awesome!

What do you think?

Silverlight 2 deployment on IIS 5.1 or 6 requires the addition of 2 MIME types in the web server’s configuration:

• .xap application/x-silverlight-app
• .xaml application/xaml+xml

Adding them to IIS is easy:

1. Right click on the website and select “Properties”
   
2. Choose the HTTP Headers tab.
   
   
3. Click the “File Types…” button.
   (It is labeled “Mime Types…” in IIS v6.0)
   
   
4. Add the types.
   

In part 1 of this series, I explained that exponential progress in computing has been going on for a long time.  Gordon Moore saw the trend over 45 years ago, and industry experts believe it will continue for at least another decade or even much longer than that.

I also made the case that this trend is not limited to computing. Progress in other areas has driven the worldwide average standard of living much higher today than just 50 or 100 years ago.  Humanity is better off [year after year] in just about every measurable material way.

Modern technology in the 3rd world. Used with permission by MarkKelley.

We are at a significant point in history. Great advances in computing, nanotech, biotech, robotics, energy, artificial intelligence and communications are occurring every day.  Most do not make the mainstream news, but their cumulative effect will raise our standard of living more rapidly than ever before.  I do not predict a completely Utopian future (we will discuss the increased dangers in a subsequent post), but I do believe most people will have the chance to live in far greater prosperity if we avoid the biggest risks.

Beginning with this post, I want to point out real world examples of accelerating change.  I think you will see that Moore’s Law is not just an isolated trend confined to the world of geeks.  It is something that will affect your life on multiple fronts, regardless of who you are.

Human genome printed, used with permission by JohnJobby

The fusion of health care and information technology – i.e. biotechnology – is experiencing exponential growth.

The fusion of health care and information technology – i.e. biotechnology – is one of many industries experiencing rapid and (by some measures) exponential growth.

Biotech could be the most promising and dangerous technology for humanity, but how soon will such things as personal DNA analysis and real time biometric monitoring become commonplace?

Consider the Human Genome Project (1990-2003):   One billion DNA base pairs were sequenced in the first 4 years.  Another billion base pairs were sequenced in the next 4 months and 1.5 billion base pairs were sequenced in the last 4 weeks. [source]

By the end of this year, you can have your entire genome sequenced for just $5000.00 and results will be ready in a few days.  By 2014, genome sequencing is projected to cost less than a nice pair of jeans and results will be ready in about 8 hours.

years > months > weeks > days > hours

See the trend?

The first genome cost 2.7 billion US dollars.  This price dropped to $2,000,000 by 2007 and $1000 is coming in 2009 or 2010.

$,$$$,$$$,$$$ billions > $,$$$,$$$ millions > $,$$$ thousands > $$$ hundreds

See the trend?

Does the promise of a biotech revolution sound like science fiction?

Personal computers (that calculate billions of instructions in one second) and personal GPS devices (that pull data from satellites 12,000 miles above the earth) would have sounded like science fiction only a few decades ago.  Does the promise of a biotech revolution sound like science fiction today?

I’m excited about the rapid price drop in genome sequencing, and I’m especially looking forward to the breakthroughs in medicine that DNA analysis will bring.  Once millions of people can afford this technology, we will have vast amounts of data to find new treatments on everything that ails us.  But what about the risks?  How will we deal with this information getting into the wrong hands?  How will we protect our privacy and how will we prevent genetic discrimination?  These issues have not been widely discussed, but you can bet the $100 genome will get lots of people talking.

UPDATE: August 10, 2009 — Another benchmark has been reached : http://arstechnica.com/science/news/2009/08/human-genome-completed-using-one-machine-for-four-weeks.ars

A typical IBM-PS1 could store 713 million bits and process 8.5 million instructions per second.

My family brought home our very own IBM PS/1 in 1992.  This small box could store 713 million bits and calculate 8.5 million instructions per second.  It was 10x faster than enormous supercomputers built in the 1960′s, and no computer this compact and powerful was available in the 1970′s and early 80′s.  Like any other “appliance”, we thought this machine would last at least a decade.  After all, the sales associate said it was “the first and last computer we would ever need”.

Half a dozen computers later, I wonder what ever happened to that old PS/1. We really didn’t keep it or any other computer around for more than a few years. Something always pushed us into getting a replacement.  Something made them useless enough to warrant another big purchase on another new system.  Was it too much dust in the keyboards? Faulty hardware? Break-ins from envious neighbors? No. The answer has everything to do with Mr. Gordon Moore.

Moore’s prediction has held true for 45 years with no signs of decline for at least another 20 years.

The original prediction (also called Moore’s Law) was strictly about integrated circuits, but we now use the term to describe exponential growth in all areas of technology:  Everything from Internet bandwidth to the efficiency of solar panels, DNA sequencing to location services, robotics, and nanotech have followed an exponential growth trend. Many of these technologies are still in their infancy, but that does not mean they are 75 or 100 years from making a profound impact on your life.  Exponential growth is explosive once the doubling factor is noticeably large.

Consider this graph.  It shows the growth of processing power for general purpose computers since the early 1970′s.

It appears that little or no progress was made for the first 30 years, but in the last 10 years we jumped from almost nothing to over 70 thousand million instructions per second.

Now look at the very same data on another graph.
The y-axis is set to display exponential data in a linear fashion using log base10.

This view shows that the first 30 years did indeed have steady growth. We doubled from 1 to 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, … to over 70 thousand — in about 35 years.

There are 3 important points we should notice about this:

1. The gains we have made in computing are not something new.
   – We have made steady progress for at least 35 years (and actually much longer that that).
2. This growth is a much bigger deal today than it was 10, 20 or 30 years ago.
   – Every 2 years we gain as much as all previous years combined.
   – Every doubling from this day forward brings an increase of at least 70 billion instructions per second.
3. Progress in computing is not isolated.
   – These are real gains that have a real impact on our standard of living.

Where might this lead us in the next 20, 30 or 40 years?

Where might this trend lead us in the next 20, 30 or 40 years? It is a difficult question that most people have not even begun to ask, but that is exactly what I will address in subsequent articles.

I will give you a new perspective that is based on solid data and reasoning that most people completely miss.  You will see (among many things) why expecting oil demand to double by 2050 is pretty unrealistic.  You will see that most predictions about 2050 are completely wrong just because people don’t foresee the second half of the chessboard.  It is a simple mistake.  We tend to expect the next 20 years to change about as much as the last 20, but in truth it will change exponentially more.

Stay tuned for some very interesting commentary on this accelerating future, and in the meantime, check out a video that shows exactly what I mean.