Jesse Ezell Blog : Microsoft

My First Microsoft Interview

Jesse Ezell — Sun, 13 Jul 2008 20:47:00 GMT

I'm happy in my present situation, but working at Microsoft has always been in the back of my mind. So, when a recruiter for the Business Solutions division tracked me down, I was certainly going to find out more. The project sounded interesting enough, so when he asked if he could take it to the next level and ask some technical questions I said, "why not?" The questions he asked basically ranged from basic stuff everyone should know to basic stuff anyone who got a CS degree should know. I've been writing code just about every day for the past 15 or so years, starting with ASM. When I started out, I had an 8086 Zenith laptop with a CGA monitor that couldn't really do much of anything else. I actually wrote programs for quite some time inside the command line debugger until I managed to find some cheap assembler program at a local computer swapmeet. As a kid, I spent so much time programming that my mom would ground me from the computer, not from going out. My first first fight was with my programming buddy that "stole" my code and changed my copyright with a hex editor (Yeah, I learned that software piracy was a bitch before I could drive). Later, I got a bigger machine and moved on to C and then C++ and did a ton of work with the STL since those were still the days when Windows was still a DOS program :). These days, I write code literally from the time I wake up at 5:30 AM till I sleep (which usually results in far less than 8 hours of sleep). It's not that I have to. I just really enjoy it. I'll hit the clubs or go out on the town every once and a while, but most nights I'd rather be coding. But... I didn't get a CS degree. To the dismay of my teachers who all had their own ideas of what I should do, I was actually going to get a Theology Degree... I never could kick the programming habit though and ended up dropping out after two years to start a consulting company.

Although I've written and optimized pretty damn complex code like H.263 video compression, path finding AI, etc., worked on teams with people that had shipped products for Macromedia and Microsoft, and shipped a few products personally, I haven't had to write my own heap class and I certainly wouldn't be using some of these slow ass sorts in my own apps (but hey, the .NET framework itself uses a quicksort which is pretty easy to implement). I faired well on the majority of the questions (like "what's the difference between a mutex and a critical section"), but it really made me feel silly that I couldn't instantly pull the answers off the top of my head to a couple of them. Why? Because it was basic stuff. Not the kind of stuff you may ever have to do in practice, but the kind of stuff you would definitely have learned if you went to school. If you didn't go to school, you have some programming chops, and some basic math skills, it's the type of stuff that you could teach yourself in a few hours. It reminds me of when I was in Junior High. We took a test to measure everyone's math skills for placement. I always enjoyed math, so I figured I would ace the thing. And I aced half of it. The complex math section. Somehow, I managed to get one of the highest scores on the complex math section, but totally screw up on the basic math section. Fortunately, the complex math score is the one that mattered, but I was always perplexed by how the hell I managed to score like that. In the end, I told myself I needed to slow down and double check the simple stuff. Treating it like it's braindead easy just makes you make braindead stupid mistakes.

This weekend I decided to go and close this little knowledge gap. I was going to start reading POSA 5 this weekend, but that can wait till next. The process has been suprisingly fun. I expected it to be pretty lame, but I was up till 4 am last night optimizing my radix sorts and heap sorts. It wasn't till my girlfriend's snores on the couch woke me up from my trance that I realized how fast the time was flying. If you didn't go the academic route, I highly recommend the excercise. Yeah, you might feel a little silly doing it, but it should be pretty easy to work through--and if you are a nerd like me, you might even enjoy it.

We'll see what happens next. Maybe I won't like the team or the offer. Maybe they won't like me. Maybe I already shot myself in the foot. In any case, I am sure as hell going to prevent that same foot from getting shot in the future :)

In junior high, I wrote a GUI using GCC and Allegro for a report about Bill Gates. We used to call him the "super nerd". The one that was so smart that it didn't matter he was a nerd. It would be funny to end up at his old company one day... though I don't think it would suprise anyone I grew up with.

Flash Lite on Windows Mobile

Jesse Ezell — Mon, 17 Mar 2008 18:43:00 GMT

Lots of people are pointing to the release of Flash "Light" on Windows Mobile by Microsoft and claiming that this is to "hold people over" until Silverlight gets here. That is just utter BS. Who the hell has ever used a Flash Light site on their phone? Hell, who really uses Windows Mobile to do any web surfing? Come to think of it, I don't think I've ever used a Flash "Light" site. Ever. From ANY device. This is certainly not something that is happening because users are demanding it and it sure isn't going to sell any more phones. More likely, this is a reflection of two things. First, Microsoft under Ozzie is making it a higher priority to support other people's products. Second, this could help limit the chances of another lawsuit over in EU land should Silverlight get big. By providing support for Flash "Light" well ahead of the Silverlight for mobile release, Microsoft is showing that it wants to win this fight fair and square.

[1] http://www.techcrunch.com/2008/03/17/microsoft-adopts-flash-lite-for-windows-mobile-as-a-stopgap-measure/

Silverlight vs. Flash: The Developer Story

Jesse Ezell — Thu, 03 May 2007 22:03:00 GMT

A few people didn’t like my proclaimation that Flash is dead. This is understandable. It is a bit premature to make such claims, but the Silverlight model is pretty amazing. As someone who works with Flash on an ongoing basis, I thought I'd chime in with a more in depth look at the issues.

First off, let me explain my background for those of you who may not know. Way back in the day, when Flash 4 was the latest and greatest, Macromedia decided to “open up” the Flash file format. They released documentation (which was poor at best) and an SDK (which was horrible at best). I saw the potential here. Finally, the format third party developers could unleash their creativity and usher in all kinds of amazing tools. Unfortunately, the documentation was full of errors and the SDK was so riddled with bugs that you spent more time debugging it than using it.

Nevertheless, debug I did. I came up with quite a list of fixes to the SDK, fixes which would render it near complete and bug free. I signed up for a free hosting account and promptly placed a list of the updates you had to make to the SDK in order for it to be bug free online. Macromedia’s response: cease and decist. Rather than integrating the changes themselves or acknowledging that they solved a serious problem, they told me that it was a violation of the license agreement to be posting that kind of information…. Some definition of “open” they have there.

Not to be dismayed, I determined that the source code license was just far to restrictive then and I would create my own SDK. As familiar with the spec as I had become, it didn’t take long for me to put something together. In a few short weeks, I had an SDK that was far more complete, far less buggy, and far easier to work with than the Macromedia SDK. It wasn’t long before hundreds, then thousands of developers were downloading the SDK and posts to Macromedia’s own open-swf forum turned from questions about the Macromedia SDK to questions about this new alternative SDK. It wasn’t too long before Macromedia completely discontinued their SDK (rumor has it that product teams internal to Macromedia even considered using the SwfSource code for their own projects).

Ever since then, I've been working with the Flash File Format. I've helped put together some award winning tools that are arguably some of the most successful SWF generation tools outside of Macromedia. This gives be a unique perspective on the differences between the two formats and how these formats enable developers to create tools that work with each of them.

Admittedly, my view point isn’t the same as a lot of people, who are perfectly content just buying the Flash IDE and can do everything they will ever need to do from there. I create tools that work with the File Format itself, tools that export their content into the Flash Format. So, if you are a software developer like myself (which is probably a good chance if you are reading this blog), then you would almost assuredly come to the same exact conclusion as myself if you knew the details of the two formats.

Animation
The Flash format itself has no notion of animation other than transformation matrices. You can apply a matrix to an element on a per frame basis to move it around. Want to move something across the screen in 3 seconds? Calculate how many frames 3 seconds will take, then calculate the matrixes required for each frame along the way. Oh, and don’t forget that the player won’t actually maintain any frame rate unless you embed blank audio tracks, so that 3 seconds might turn out to be 2 or 6 or 5, it just depends what kind of mood the machine is in.
Silverlight supports the WPF animation model, which is not only time based instead of frame based, but lets you define the start and end conditions and it will figure out how to get there for you. No need to deal with matrixes. No need to calculate positions on various frames. It just works.
Shapes
Flash stores its shapes using binary shape records. In order to write shape definitions, you will need to either license a 3^rd party Flash file format SDK, or build your own. It isn’t too difficult, but it does require a bit of a learning curve and the ability to manipulate things at the bit level, since shape records don’t align on byte boundaries. Needless to say, it isn’t the kind of thing most people can write and have all debugged in one afternoon.
Silverlight uses XAML. XAML is text based and can be output using a simple XML object. No need to buy special libraries to write files. No need to write your own libraries. Just stream some text to a file and you’re done--easily the type of thing that can be debugged and finished in an afternoon.
Text
Flash stores its fonts glyphs using the same exact shape definitions that are used for any other shape. The player itself does not understand TTF files, so you’ll end up digging deep into the Win32 APIs and the fairly vague definitions in the Flash file format documentation to come up with something that sort of does the trick. You’ll probably spend ages trying to deal with all the intricacies of fonts, because it turns out that typography is actually fairly complex… and you will have to deal with all those complexities yourself.
WPF/E lets you embed true type font information directly into your projects, and download that information with the downloader object. No need to do anything special. No need to handle anything yourself. It just works.
Video / Audio
Flash supports multiple video formats. The latest codec is really high quality and the bandwidth usage is nice. There is one problem though if you are creating a tool that outputs Flash content… the formats it supports aren’t really used by anyone else. The original video codec, Sorenson’s proprietary H.263 implementation is a mutant version of H.263. The compression follows the spec fairly closely, but there are a bunch of features dropped out and you can’t exactly just go find a complete spec on how to build your own encoder. The later codec from On2 puts you in an even worse position. Licensing Sorenson’s codec isn’t that expensive, but On2 will rape you with fees. They are relying on revenue from licensing the codec used by Flash to revive their $2 a share stock price. It is also a completely proprietary format (where at least the Sorenson one was loosely based on a standard). The audio formats Flash supports are all proprietary, except for ADPCM, which no one uses because of its horrible compression, and MP3, which is decent but dated, and still requires licensing fees and 3^rd party conversion libraries.
Compare that to the Silverlight story. Silverlight implements industry standard VC-1 codec for video, as well as offering support for WMV and WMA. Just about everyone already has Windows Movie Maker, but if they don’t it’s not a big deal. Why? Because Microsoft makes available a free Encoder SDK for producing WMA and WMV. So, not only are you using formats that people are more likely to be able to encode themselves, but Microsoft also provides your product with SDKs if you want to do the encoding yourself. The best part about it is that Microsoft doesn’t rely on WMA/WMV licensing revenue to keep themselves alive, so not only is it easier to integrate, but it’s also cheaper.
Scripting
You can reuse C# classes from your tool inside your exported content. There is no development environment out there for creating real desktop applications which is based on ActionScript. If you go the Flash route, this means that all your classes and objects have to be written twice. You need .NET classes to handle the author time experience and Flash classes to handle the run-time. If you have server components, once again you need to switch back to .NET and throw out all the classes that the run time is using. For example, let’s say you are creating a tool that outputs rich media quizzes. With Silverlight / .NET, the same entity classes you use to deal with results in the player could be reused on the server side. With Flash, you’d have to write all that logic 2x and keep it in sync as your tool changes.
Tools
You can create Silverlight content with the same tools you use on a daily basis. Visual Studio.NET is by far the most powerful and most popular IDE. You can potentially have all the code for the server components, the authoring tool components, and the runtime/player components inside the same project. No extra skills required. No needing to hire some special Flash guru to do the graphics junk. Every developer can contribute to every part of your application.

The bottom line: about the only thing Flash has going for it from my perspective is adoption. Adoption isn't hard to achieve, especially for the people that ship the operating system 90% of the world uses. As such, it's just a matter of time till that is no longer part of the equation. Adobe has a lot of work to do in the mean time, and the clock is ticking. Open sourcing Flex is a really good start in the right direction... unfortunately, Flex was built on top of the wrong platform from the start (something I told the Flex team while Flex was still in Alpha), so this last effort, while a good one, still might not be big enough to turn the tide that is coming. Now, this isn't to say that Flash isn't a great format and doesn't enable a lot of scenarios (like I said, my job is working with Flash and I'd be doing something completely different if it wasn't for Flash). So, Flash is great. Silverlight just solves a lot of the major problems that I've run into with Flash.

Open Source: It Costs Too Much

Jesse Ezell — Thu, 05 Oct 2006 07:27:00 GMT

Our startup honestly wanted to use OSS products. We do not want to spend time for any OSS bug fixing so our main requirement was -official support for all OSS products-. We thought were prepared to pay the price for OSS products, but then we got a price sticker shock. Now behold: QT is $3300 per seat. We have dropped the development and rewrote everything to C# (MSVS 2005 is ~$700). Embedded Linux from a reputable RT vendor is $25,000 per 5 seats per year. We needed only 3 seats. We had to buy 5 nevertheless. The support was bad. We will go for VxWorks or WinCE in our next product. Red Hat Linux WS is $299. An OEM version of Windows XP Pro is ~$140. A Cygwin commercial license will cost tens of thousands of dollars and is only available for large shops. We need 5 seats. Windows Unix services are free. After all, we have decided that the survival of our business is more important for us then 'do-good' ideas. Except for that embedded Linux (slated for WinCE or VxWorks substitution), we are not OSS shop anymore. [1]

Microsoft has been arguing the TCO part of the equation for a long time now. Are people finally starting to listen?

[1] http://ask.slashdot.org/article.pl?sid=06/10/04/0452244

Microsoft = Evil

Jesse Ezell — Tue, 05 Sep 2006 20:15:00 GMT

Do you see the devil in this picture on MSDN?

[1] http://msdn.microsoft.com/vstudio/images/products/vssafe/hero_SKU_vss.jpg

[2] http://reddit.com/info/ge0o/comments

Live Services are in Trouble

Jesse Ezell — Tue, 08 Aug 2006 20:18:00 GMT

Nial Kennedy is leaving Microsoft. His reason? He joined MS to work on the Live initiatives and after MS stock fell when they announced their Live goals, they scaled back everything. Sounds like Live won't be nearly as cool as it could have been.

[1] http://www.niallkennedy.com/blog/archives/2006/08/leaving-microsoft.html

Rory > Joel

Jesse Ezell — Tue, 20 Jun 2006 01:14:00 GMT

Seriously, not only does Rory provide 10x more insight than Joel Spolsky's "I'm 10x Smarter than Anyone at Microsoft but My Software Still Sucks" blog, but he is a million times more entertaining. Take, for instance, his recent article about the Ballmer criticisms:

"...To begin with the Ballmer bashing, think about the man and his background. He’s a sales guy. That’s what he did for the company before running it, and, from what I understand, he did a magnificent job. During my nearly two years at Microsoft, I’ve met some “old-timers” who have endless stories about Ballmer’s methods (entirely fair and legal – no joke there) that would blow you away. The guy could sell condoms to the Pope..." [1]

[1] http://neopoleon.com/blog/posts/21095.aspx

Ballmer Ain't Goin' Nowhere!

Jesse Ezell — Sun, 18 Jun 2006 10:59:00 GMT

Contrary to a lot of talk lately, David Kirkpatrick--the Senior Editor of Fortune--thinks Ballmer is here to stay and that it is a good thing.

[1] http://money.cnn.com/2006/06/16/technology/fastforward_fortune/index.htm?cnn=yes

More Calls for Ballmer to Step Down

Jesse Ezell — Fri, 16 Jun 2006 17:42:00 GMT

Looks like Business 2.0 is in agreement [1] with Joel [2] that it's time for Ballmer to step down from Microsoft.

[1] http://money.cnn.com/2006/06/15/technology/business2_workingtech_0615/index.htm?cnn=yes

[2] http://www.joelonsoftware.com/items/2006/06/15.html

Joel On Ballmer

Jesse Ezell — Fri, 16 Jun 2006 04:30:00 GMT

"As of now, Microsoft stock is surprisingly quiet given the announcement that Bill Gates will step down. It should probably be going down. Ozzie is smart but not in the same class as Bill Gates. And it's really Ballmer that needs to go." [1]

Ouch...

[1] http://www.joelonsoftware.com/items/2006/06/15.html