Specialization is for Insects

A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.

I can only do about half of these myself, but I’ve always agreed with the sentiment. Back in college my advisors all explained why it is important in one’s career to specialize. But I’ve always chosen breadth over depth. My ideal has never been the specialist, but rather, the synergist – someone who can combine multiple skills and talents to come up with creative and unique solutions to problems.

Some would argue that when one tries to do many things, it becomes impossible to do any of them well. This is not true. What is true, is that if you do not specialize, it becomes impossible to be the very best at something. I know this, because for a time in my career I did specialize in a specific area of technology and became one of the best at it anywhere. Now, I can safely say that I am not the best at anything. But it does not mean I don’t do a lot of things very well.

What might Heinlein add to his list if he were writing for today’s technical society?

A human being should also be able start and run a company, speak in front of a group (any size), write a book, film and edit a video, plan an event, write and place an ad, teach a class, read or write a financial statement, manage investments, work with spreadsheets, write production quality code (on several platforms and in multiple languages), handle a TV or radio interview, mentor a kid, build a computer, wire a network, play an instrument, configure a firewall, architect a complex software system.

Ok, that might not be his list, but it is at least part of mine. It turns out that if you spend time doing a lot of different things over enough years, you can actually become surprisingly good at them. You can even find connections – ways different knowledge sets overlap – that a specialist might never see.

So if you are still early in your career, I encourage you to create your own list. Don’t limit yourself to the confines of what you are taught in school or at work, or what you think you are supposed to be learning. And please don’t limit yourself just to technology – there’s a lot more going on out there.

And if you’re visiting this site to gain some insight into my background, suffice to say – it’s a long story. I like to think that were it possible for me to meet Heinlein today, he would, after some conversation, judge me to be a human being. Coming from him, I would consider it the highest of compliments.

In addition to my technical career, I’ve spent over 20 years volunteering as a youth group advisor with a group that emphasizes youth leadership.

I really believe that in today’s world, academic success is not enough to achieve success in a career or in life. It takes more. Call them social skills or communication skills, or initiative or the ability to plan and execute a task… All of them fall into the category of leadership skills.

Over the years I’ve learned a thing or two about how to teach these skills, and have just published a new book on the subject. Like my technical books, it is a practical “how to” guide – not an abstract theoretical tome. Unlike my more technical books, it’s an easy read and, I hope, entertaining as well as useful.

Whether you are a coach, scoutmaster or youth group advisor, a teacher, or even a parent of a teen – I’m confident that this book will offer both insight and practical strategies to help you become more effective at teaching leadership skills, and better able to help teens prepare for an increasingly complex future.

You can read more about it at TeenLeadershipBook.com.

I wanted a shopping cart that had really good extensibility – one that could connect to our licensing server not just to allocate keys, but to perform custom operations like allowing purchase of additional installations for existing keys.

What I found was largely disappointing. Not that there weren’t some great packages out there. If I wanted to set up a large online store with many products, it was clear that setting up my own Amazon.com equivalent would not be hard at all. There are any number of powerful online stores with numerous features available. But none of them had exactly the right features, and none had the extensibility I needed.

What I really wanted was a shopping component – some ASP .NET controls that could be dropped onto any ASP.NET page and would somehow work together to implement a shopping cart. There would be a control that would “add to cart” a single product – and a page could have any number of these control to display multiple products. There would be a shopping cart page that would allow modification of quantities or deletion of items from the cart. And there would be a checkout control.

In my ideal solution checkout would be handled by an external processor such as Paypal, and once the order was confirmed my code could issue license keys, download links, send out customer Emails and so on.

After searching through and installing trial versions of numerous packages, I came to realize that none of them came close. The truth is, I didn’t want an online store – I wanted an online store component.

So I wrote one. The excuse I used to justify the time was that it would make a great application note for the latest version of our licensing system. But in truth, it sounded like a lot of fun – and it was. I ended up using authorize.net for the credit card integration (they have the best online documentation for developers of any of the card processors I found other than perhaps Paypal).

There’s still some work ahead to turn it into an application note for distribution, but the store is now live at http://desaware.com/purchase/store.aspx.

Above all, the experience confirmed to me the rightness in the way we decided to extend the licensing system for the newly released version 2. Instead of piling in features, we focused on enhancing the extensibility model. Our licensing system is not so much a licensing application as it is a licensing component – and right now it’s hard for me to imagine a licensing scenario that it can’t implement.

Loans

There are no sub-prime loans in the Microsoft economy. Nothing sub-prime would ever be shipped. True, many customers would end up with version 1.0 loans that would occasionally run into problems – payments might mysteriously fail to get recorded or interest rates change without logical reason. But don’t worry – everyone could get their loan upgraded for a nominal cost when a new version comes out – refinancing would go smoothly for 99.5% of customers, only a small number would crash into spontaneous foreclosure.

Microsoft loans will come with legal agreements as long and incomprehensible as any we see today. And just like today, nobody would read them. Unlike today, it would never matter.

Derivatives and Credit Default Swaps

Microsoft has extensive experience creating incredibly complex products that almost nobody understands (COM, WCF, etc.), which makes it the ideal organization to take over the development and marketing of financial derivatives. However, unlike today’s financial products, Microsoft derivatives would undergo extensive testing both internally and through a massive pre-release beta and preview program to customers ranging from individuals to the largest banks.

With thousands of amateurs and experts able to download and examine every derivative and credit default swap before release, most potential problems will be discovered long before they are actually deployed. Sure, they will ship with some bugs, but the odds of bugs significant enough to crash the economy will be very small. Any such serious bugs discovered after release will be quickly patched using the Microsoft automatic update service for financial instruments, that will be included with every copy.

Executive Bonuses

Microsoft licenses products generally by machine or user. Microsoft doesn’t care if you are using Word to write a 3rd grade paper or a billion dollar contract, so it won’t care if a Microsoft fund has a thousand dollars or a trillion. From loans to hedge funds, once you’ve bought the product, there are no further charges (other than support or upgrade fees). Fees for financial products will be based on number of customers, but not the size of the product. Fees and costs will drop – and large bonuses and salaries will vanish. Of course Microsoft product managers will be paid a decent salary, and will have some company stock options that might become worth something someday. But the days of huge executive bonuses at institutions that are losing money will be over.

The overall economy

Our existing economy is a complex and bloated system with a long history of solid growth and productivity punctuated by booms and crashes. Microsoft Windows is also a complex and bloated system with a long history of solid growth and productivity punctuated by booms and crashes.

However, unlike the real economy, when the Windows run economy stutters, most of the time it can be fixed in short order with a quick reboot. Those rare crashes that occur are typically also solved by a reboot, and only occasionally need a complete reformat and restore.

Imagine an economy that can be rebooted in hours instead of years. One with extensive documentation and transparency as provided by MSFN (Microsoft Financial Network).

The mind boggles at the possibilities. The Bernie Madoff’s of the world would have no chance in an economy where dozens of vendors sell SAV (Securities anti-virus) software. Hundreds of thousands of developers will build financial products for the new economic system (run your own bank, publish your own credit card). And there would be no secrets – anything you ever wanted to know about the economy or any financial product or institution would be instantly available to view – on Google of course…

Now, if it’s not a very old computer – say, a year or three, it’s easy: just restore the computer from its restore partition or backup DVDs, run a program like Eraser to overwrite the unused space with random data to make sure no personal information is left, and give it to a friend or donate it to a local organization.

But what if it’s a really old computer – say 10 years or so? In that case you have a much bigger problem. Disposing of the hardware itself won’t be hard – your local electronics recycler will be glad to take it off your hands. But that’s the easy part.

I know what you’re thinking. What am I doing with a 10 year old computer? Well, as you know, I run a software company. And we still occasionally get support calls on very old legacy software, which means we need to maintain the ability to debug or build some pretty ancient code. So when a new machine comes in, the previous machine remains frozen in time.

Now you may think you don’t have this problem, but let me ask you – how long have you had a computer? It may well be 10 years or longer. If so, you may have some old documents from back then that you’ve been carefully backing up or copying to your new machines each time you upgraded. But can you read them?

Yes, I know – the file can be read – but can you read or view the contents? When companies upgrade software they often upgrade the file formats. The updated applications can typically convert from previous versions of the file formats, but usually only a few versions back. Even Microsoft does this – For example: Office 2007 can’t open Word documents from 10 years back. And if you used an application that no longer exists, you could be out of luck.

Virtualization to the Rescue

So whether you are software developer, or just want to assure yourself of the ability to read old documents, it’s not enough to backup files from the old system. Fortunately, virtualization technology has reached a level where it is possible to create a virtual image of the original machine. Possible, but not trivial.

In my case, the box was a dual boot Windows 98 and NT4 box. Creating a working image took the following steps:

• Create a Ghost image of the disk using Ghost 2003
• Create a new virtual hard drive, then using an existing virtual machine, copy the image to the vhd.
• Create a new virtual PC and drive, then attach the vhd containing the ghost images.
• Restore the partitions to the new vhd, and create a new partition (you’ll see why soon)
• Dig through MSDN to find a Windows 98
• Boot into Windows 98 on the virtual pc. Go through numerous cycles of “found new hardware” until it finally boots.
• Boot into Windows NT – watch it fail miserably in numerous ways (blue screen of death, reboots, freezes, etc. – NT4 does not handle hardware changes well).
• Dig through MSDN to find an NT4 CD and an NT4 service pack CD
• Boot back into Windows 98, and in a DOS window create a set of NT4 startup disks
• Boot the NT4 startup disks and install a new NT4 on the empty partition.
• Install the latest service pack.
• Write down the video driver and network driver used by the clean install.
• From the new NT4 OS, navigate to the system32 directory of the one you’re trying to bring up.
• Backup the Drivers subdirectory and copy over the one from the current OS.
• Backup the system32 directory files, then copy any files from the current system32 directory that are newer than those on the partition you’re trying to bring up (windiff can help identify the files).
• Boot into the restored NT4. With luck it will now boot.
• Install the display and network drivers based on the information you wrote down earlier.

It’s a lot of work – almost as much as bringing up a new machine. And if you’re fortunate enough to be converting a newer OS, things should be much easier. But though annoying, the investment offers valuable insurance for the future.

What, more backups?

That backs up the hard drives and operating environment, but remember, this is a very old machine. That means old hardware – including a tape drive and a Jazz drive (a removable hard drive device that uses 1GB cartridges). The Jazz drive still worked so I was able to back up all of the cartridges. The tape drive did its best, but whether it was the drive or the 10 year old tapes, there was little luck there. Fortunately, the tapes were just backups so did not contain anything important (I hope).

Back to the Hardware

The hard work done, now all that’s left is the final preparation for disposal. Removing the hard drives for erasing is a first step (though you can avoid this by using a recycler you trust to shred the drives). In my case I also had to pull and store the graphic card. Why? because this machine used an old Matrox Rainbow Runner card. This card enabled video editing, but used its own proprietary hardware video encoding (remember, machines back then weren’t fast enough to handle CPU video encoding and decoding). A few years back I converted a bunch of old videos on that machine to MPG, so I don’t think I’ll ever need that card again, but I’ve archived it just in case, and Matrox does have legacy drivers on their web site just in case.

All that’s left is erasing the old drives. I generally use the Eraser program (from sourceforge.net) to overwrite the drive with random data. If I can’t access the drive, I open it and physically damage the platters (along with running a strong permanent magnet over them). The old Jazz cartridges got the physical treatment, as did the now unusable tape cartridges.

So after (too many) hours, with data destroyed, the system is finally ready for the recycler. It’s seen loyal service, but I can let it go peacefully, as it’s spirit (in the form of a virtual image) will always be there if I need it.

This brought forth a great deal of comment, one of the most interesting questioning why anyone would encourage their kids to go into the software business in the first place. Why would you go into an industry where it’s increasingly difficult to keep up, where jobs are increasingly being outsourced, and where there is rarely long term career viability? Doctors can continue treating patients into their 60′s and 70′s, Lawyers continue lawyering, bankers banking and politicians politicking until they drop dead (and sometimes longer in the case of politicians). But past 40 or so, software developers who aren’t fortunate enough to have solid management gig are likely as not to find themselves unemployed or in some other career.

I’ve been particularly fortunate in this regard. Through a combination of being self-employed, honestly passionate about technology, and a bit of talent, I’ve had the opportunity to continue coding. What’s more, I’ve had the freedom to diversify into all sorts of new technologies beyond my core expertise. But keeping up is an effort. I often compare the life of a software developer to that of Alice in “Through the Looking Glass” – where Alice is told by the Red Queen that she has to run as fast as she can just to stay in place, and to make progress she has to run even faster

Why would anyone wish this on their kids?

It seems that our interests as a nation are radically different from our interests as individuals. There is no doubt that we need a lot of skilled engineers and programmers going forward. Not only do we need to maintain our existing technological base, there’s going to be increasing demand for power engineers and some incredibly sophisticated software to handle the future smart grid (or EnergyNet” as Juval Lowy calls it). Where are those developers going to come from?

I suppose we can continue to import them. But will the United States continue to be a desirable place for migrant technologists? To some degree it is likely, but as their local economies develop, they may find better opportunities at home. And many engineers consider foreign workers to be part of the problem – competition that makes it even less desirable to enter the business. But let’s assume for the moment that for the foreseeable future foreign nationals will continue to provide needed support our technological society, and that it is in our national interest for them to do so – thus making home-grown talent less essential.

What do you encourage your kids to study if not technology?

Medicine is good – the Baby Boom generation is getting older and there will be demand for more health care. Lawyers? It’s hard to argue that the U.S. needs more lawyers. Investment bankers? It seems unlikely that will soon return to being the path to instant riches it once was. Therapists? That’s not bad – there are going to be plenty of stressed-out unemployed people who will need treatment.

I suppose there’s always the option of flipping burgers or greeting folk at Walmart (not that there’s anything wrong with that).

It’s not enough to say that technology is a hard field – it is. But hard isn’t necessarily bad. And training in engineering sharpens one’s problem solving skills in a way that few other degrees can manage.

So here is my answer to those parents who wants the best for their kids: absolutely encourage them to get a technical degree. But don’t let them think they can rely on that alone to create a good future for themselves in a tumultuous world. Make sure they develop great communication skills, written and verbal. Teach them about money and accounting so they can use the relatively high income from the early part of their careers to establish a foundation for what comes next. Teach them to how to evaluate risks and opportunities so they can jump when the time is right and perhaps create their own businesses. And teach them to follow their passion, so that no matter what happens with their career, at least they’ll have some fun along the way.

At the time, I ended up upgrading an old Thinkpad 240 – a machine that today would be called a Netbook, but that anticipated that class of machines by almost a decade. And it served tolerably well. It was slow enough to be annoying, and the battery life was worthless, but it was good enough.

This summer I wanted something better, and with all the talk about NetBooks I decided it was time to take a look and see if any of the new machines would serve my needs. After reviewing specifications and reviews, I ended up snagging an Asus Eee 1000HE PC.

Honestly, I did not have high hopes – laptops are almost by definition studies in compromise. But I ended up being blown away. This is a nice machine. You can read the full review in my gadget column. Suffice to say that with a 2GB memory upgrade, it has plenty of power for casual development work, is easily small enough to fit into a book bag or backpack, and at a hair over 3 pounds with over 7 hours of battery life is truly portable.

Oh yes, and I will never buy another laptop without a multi-touch touchpad.

Read my full review

No problem, right?

Being an experienced .NET developer, I figured this would be a good time to try the Visual Studio Office system – to create the add-in using VB .NET. I mean, how hard could it be?

Not hard at all as it turned out. The project template for the add-in worked well, and it took no time at all to find some simple examples online to build on. I had enough familiarity with the Excel object model to make short work of the task at hand, and the integration that allowed debugging of add-ins in Visual Studio worked brilliantly.

Then it came time to test deployment – as I needed to provide this add-in to a couple of clients.

What a fiasco.

The setup project that was created didn’t work. The instructions on MSDN and other sites for diagnosing problems were complex and unreliable. I found additional articles with suggestions for creating custom installation actions to handle the security issues that might have been the problem. There were clearly differences between versions of the Visual Studio office runtimes depending on framework version and versions of Excel to support, with no clear explanation of which to use and how to create a deployment that will work with both.

I spent as much time trying to figure out the deployment as I did creating the add-in when I decided I was wasting my time. The Visual Studio office support may be nice, but the deployment solution and documentation is abysmal. Microsoft should be ashamed to have released it. It is perhaps the most disappointing experience I’ve had with any Microsoft developer technology.

I ended up porting the code to VBA in an Excel workbook and creating a .xla Excel add-in. Though I missed the intellisence of Visual Studio, the building and debugging experience was fine. And deployment was trivial – save the workbook as an add-in. Deployment consisted of browsing to the add-in file and enabling it – and the add-in worked perfectly on Excel 2003 and 2007 the first time I tried it.

I’d come to think of .NET as current technology rather than the wave of the future. It’s clear to me that at least when it comes to Office, .NET is still future technology – costly, complex and unreliable. If I need to do any more Office add-ins, you can bet I’ll stick with VBA.

First, it is not a copyright violation.

If you read a book out loud, is that different from reading it silently? If someone reads a book to you, does that mean you both have to buy a copy of the book? What if you hire someone to read to you? Of course not. So why would an automated reading device be any different? It is not.

Now a true audio book is different from a printed book. Why? Because it is a derivative work – a performance of a book. It is a new work that is derived from the original.

Some might argue that speech-to-text is also a performance of a work and subject to a new copyright – and it would be, if you tried to sell and market such a work. A similar situation exists with translations. If you wish to translate a book and sell the translation, you have to get permission from the copyright holder. But if a friend comes over to read a book in a foreign language and translates it for you as they read, that is perfectly fine. Text-to-speech is that high-tech friend.

But the copyright argument is not the biggest reason that the Author’s Guild is wrong about the Kindle. The real problem is that they are acting against the best interests of authors.

Here’s why.

Let’s consider audio books on CD in two categories. In general fiction Amazon.com shows 13867 results. In SF and fantasy, 1891 results.

Why would someone buy an audio book? Possibilities include:

• Unable to read (visually impaired)
• Too lazy to read
• Wants to utilize commute time (while driving, on public transit).
• Enjoys the performance.

Let’s assume that the first three of these represent 75% of the market, and that it can be replaced by text-to-speech. Let’s also assume that few people would buy both the print and audio book. Since audio books cost more than print books, text-to-speech technology should result in some drop of income to these authors as people choose to buy the print book instead of the audio book. If audio books represent 10% of a book’s total sales, and if we assume the audio book pays an author twice what a print book does, the author will lose 50% of 75% of 10% of their income – a drop in 3.75% of their income.

Of course, this would have a much greater impact on audio book publishers – but then why isn’t the audio book publisher’s guild complaining? Surely the Author’s Guild wouldn’t make such a fuss over a 3.75% drop of author’s income.

Especially when you consider the following:

Amazon lists 403,000 results just in general fiction, almost 90,000 books in SF and Fantasy. Or put another way, maybe 3% of printed books have audio books available. If there’s one thing we know about the market – when prices drop, people tend to buy more. Text-to-speech effectively reduces the cost of audio books which means people will buy more – and now they’ll be able to choose from any title, not just those with audio books available. Ultimately this will benefit far more authors as book sales increase overall.

While the numbers I use are largely hypothetical, the principle is clear – text-to-speech is good for authors. It makes their existing books more accessible and opens them to markets (commuters, visually impaired) that were otherwise closed to them. Authors win. The consumer wins. A few authors might lose a small amount. And audio book publishers potentially lose – they will have to market their good purely based on the quality of their performers, not just on the fact that it is an audio book.

The Author’s Guild should live up to its name and acknowledge the fact that the Kindle’s new text-to-speech feature is neither a copyright violation, nor is it counter to the interests of authors.