codesurgeon blog

Slides for my TUM Grid Computing talk: "Cloud Computing Is Not Cotton Candy ... Or Is It?"

2010-10-14T21:59:00.000+02:00

Cloud Computing Is Not Cotton Candy ... Or Is It?

Slides for Talk on Selected Features of C# 4.0 and the .NET 4 Framework

2010-07-05T22:39:00.001+02:00

2² C# 4.0 and .NET 4 Selected Features

These are the slides for the short version of my talk 2² C# 4.0 and .NET 4 Selected Features, presentation time for those is approx. 30 minutes. A slide set which goes into more detail for the DLR and .NET dynamic features is geared for 90 minutes incl. live code demos. It's available upon request.

Most slides in the original Office 2010 PowerPoint file are animated in order to not overwhelm the audience with slides bursting with text and information. The embedded Scribd version is based on Office 2010s PDF export. It is missing the animations and intended for print & screen reading purposes.

As always, your feedback is appreciated in the comments or via email to mustafa ÄT codesurgeon DÖT com

The iPad ... it tastes just like chicken!

2010-04-07T03:07:00.003+02:00

... yet it seems to make men, predominantly aged between their quarter- and midlife crises, to roll on the floor, snarling and biting.

Being an all-around geek, former co-host of Germany's most popular Mac-centric podcast, founder & co-producer of the biggest indie video gaming show north of the Alps and software engineer, I have been closely following the birth of Apple's latest aluminum, electronics and glass ~~sandwich~~ pancake. After all, the prospect of a "usable" tablet computing platform, bearing interesting challenges and new opportunities for a range of applications, tingles more than a single sense in the geek mind.

I have been excited about the concept ever since Bill Gates couldn't stop talking about it around the turn of the millenium. Unfortunately - as with phones afterwards - Microsoft's implementation of the idea left much to be desired. The Redmond behemoth, employing an army of bright and talented engineers, stuck too close to its Windows OS UI paradigm.

It wasn't until 2007 and Apple's initial iPhone release that a touch user interface successfully premiered in my and tens of millions of other personal computing universes and secured itself a place to stay. As we've all experienced by now, Apple's execution of the idea was close to flawless and their success well deserved. On the other hand, right now it seems they've been drinking a little too-much of their own kool-aid and seem a little lost in hubris - making themselves believe they've invented mechanisms (multitouch UI, OO graphics subsystem ... WTF?) which they have not.

Let's all hope that Microsoft ups their game, Android gains more traction, handset makers rise up to the challenge and the closed nature of Apple's platform leads straight into a wall sooner rather than later ... for the sake of us all, especially Apple's; for we all know that once Steve Jobs leaves the mothership, it will force the company to crash-land and need all the karma-cushion it can build up until then.

Yes, I am aware of the thousands of people working for Apple - among them bright devs - usually kept out of the limelight ... well, actually any light at all aside from their release into the safe confines of WWDC once a year. By the way, being a developer, to me this is probably the single most disturbing aspect of Apple's corporate culture, absolutely contrary to e.g. Google and Microsoft and any vibrant tech company in its right mind.

Over more than three decades, Jobs has proven to pick the right people and projects and be absolutely uncompromising (interestingly enough, even self-destructive) when he's set his mind on something.

Imagine an executive meeting at Apple today. Steve is presented with an idea, feature or prototype that he doesn't like. That pretty much settles the project's fate. Forever. Who is going to override his decision? Who is going to argue with the man who can rub you in the face that he's been ~~there from~~ the beginning of personal computing as we know it. Who is going to go rogue with a project inside Apple as he himself did with the Mac in the eighties? The man who financed George Lucas' former wandering hippie circus of CG-genius PhDs long enough for them to churn out the world's first fully computer-animated feature film? The list goes on and no matter what picture the various Jobs' biographies and personal stories we've all read and heard draw of him, he has obviously gotten some things right.

Now imagine the same meeting post-Steve ...

Apple *is* one auteur's will manifest in corporate form. If creative dictatorship is the company's only successful modus operandi (Sculley, anyone?), will Apple still work when it turns into a soviet? It's much less a traditional company than for instance Microsoft is. Is Apple up for a slow death-by-compromise/committee? We'll see but until then let's not steer off further from course and explore the shallow waters of gadget love.

The iPad has been released in the US four days ago, on April 3rd. The internet has been abuzz with videos, columns, spec sheet recitations and reviews. As far as those types of articles go, I have nothing more to add to what a Google search currently reveals.

I'd much rather share my initial hands-on experience with the device: it's much less magical than it is useful in a utilitarian sense. As with the iPod and iPhone, Apple has taken a rough concept others only pursued half-heartedly and made it accessible and joyful to use. Contrary to countless posts (authored by hyperventilating authors?) on the web, there is *no* black magic involved.

The hardware is as slick as all the photos on the internet convey. The solid aluminum casing, sealed on the front by a single glass-covered screen, is a joy to touch. As for holding the iPad for longer periods of time, that is another story. It's a little on the heavy side. Then again, so are textbooks and laptops. Just don't expect the convenience that comes with the light weight of an Amazon Kindle during handheld reading.

IMHO, none of the apps available right now are groundbreaking. The Marvel app would have been close if Sony Computer Entertainment Europe wouldn't have already come up with their PSP comic reader half a year ago. Nevertheless, it's nice to have a compact device with a screen of this size.

I have no doubts that the best is yet to come. The iPad's form factor is in a region where size *does* make a non-trivial difference. Compared to the iPhone/iPod Touch, the additional screen real estate makes it possible for applications to utilize screen input with both hands without occluding most of the display. That is definitely the distinguishing feature of the device and it matters more than what the "but it's just a bigger iPod Touch" blows might suggest. For the purposes that I see the device serving, the lack of a keyboard is liberating - no more cumbersome lid opening and lap balancing.

The display size makes the iPad perfectly suited to replace our living room computer - a white MacBook. It's what we use when we feel like browsing the news on the couch, checking up on friends and photos on Facebook or digging into Wikipedia when mutual stubbornness deadlocks a discussion. Now if only Apple would have included a front-facing camera, the iPad would have become the ultimate home-bound communications device and helped propel video chat into living rooms. I guess, for that to happen, we won't have to wait any longer than for the next product iteration.

As it is right now, the iPad feels right at home next to a stack of magazines, easily replacing your current terminal of choice for internet lookup.

It's different, it's cool ... and it tastes just like chicken.

--
Special thanks to geek chick extraordinaire Winnie Teichmann for dropping by, joint iPad-ing and taking funny photos.

The Windows Cache Manager - A Quick Overview

2009-11-27T05:28:00.000+01:00

Well, if I have to spend my days (and currently nights!) with Windows cache and memory management, it should at least be good for a blog post ;)

The Windows Cache Manager is a central operating system facility managing the system cache, which transparently supports all Windows filesystems. It is capable of handling file data - including multiple data streams for filesystems that support such - as well as filesystem metadata. Since the operating system provides for a centralized external storage caching mechanism (including support for network and disk based access), filesystems are alleviated of implementing proprietary caching functionality.

The Cache Manager operates on a relatively high level for a kernel mode component, interpreting files as byte streams vs. individual disk blocks on physical storage media. Abstracting away from physical file layout allows for simplified analysis of data access patterns and resulting optimizations for instance in terms of better read-ahead performance.

A segment of the kernel virtual address space is assigned to the system cache by the Windows Memory Manager. The Cache Manager further slices this segment into 256KB chunks to hold cached views of filesystem data.

Mapping file contents from disk into the system cache is 256KB-aligned - that is, when reading 100 bytes from a 1MB file at byte offset 400,000 the Cache Manager will fill up the corresponding 256KB chunk in the system cache starting at byte offset 262144 through 524288-1. Assignment of filesystem mappings into the system cache 256KB slots happens in a round robin fashion.

The filesystem tracks the physical bits of a logical file and the type of I/O requested (buffered vs. unbuffered) within corresponding file objects. Multiple file objects can represent the same file on disk. The former share a "shared cache map" which serves as metadata structure for the file on disk being represented by the file objects. It contains a mapping from file offset for a particular view to system cache virtual address space. The Cache Manager utilizes the shared cache map to ensure cache coherency among the different file objects for the same file - for instance by only caching the same view into the file exactly once. Data relevant to individual file objects alone - that is, access patterns to handle read-ahead - is kept in object-specific private cache maps.

The Win32 API exposes flags to influence cache read-ahead behavior via FILE_FLAG_SEQUENTIAL_SCAN and FILE_FLAG_RANDOM_ACCESS When not specifying any of the file flags at all, the Cache Manager tries to analyze and determine an access pattern (strident access, sequential forward or backward scan) in order to optimize its performance.

A number of lazy-writing threads that wake up periodically, sequentially go through the system cache and flush out a predetermined number of modified 256KB chunks to persistent storage in each wake-period. A list of linked shared cache maps indicates system cache chunks that have been written to and thus need to be considered by the lazy-writing threads.

It is interesting to point out, that with Windows' particular kind of memory mapped file caching, the Cache Manager remains oblivious to which parts of the cache actually reside in physical memory (RAM). Instead, the Windows Memory Manager is free to optimize physical memory utilization to balance system cache vs. demands of user processes.

The overall physical size of the system cache depends on the operating system - for 32Bit systems it maxes out at a little over 500MB. As for the size of the kernel virtual address space assigned to the system cache, it depends on the amount of RAM in the system and is calculated via the formula: 128 + (RAM in MB - 16) / 4 * 64.

Resources:

Fourth Edition of Russinovich's Windows® Internals, Fourth Edition - an updated version has been released in June 2009, covering the more recent versions of the OS: Windows Internals, Fifth Edition.
2005 MSDN Channel 9 interview with Molly Brown, the engineer responsible for the Vista version of the Windows Cache Manager: part 1 and part 2

Interview in Germany's #1 College Magazine: The Google Experience and CS Careers

2009-07-20T10:25:00.000+02:00

In November last year, Dominik Hahn, editor at presstige and fan of a couple of the podcasts that I'm collaborating on, contacted me for an interview in their publication. Little did I know at that time, that their magazine boasts a print run of 10.000 issues. IMHO, that is pretty impressive for an entirely student run magazine.

He was particularly interested in career advice I could give for fellow CS students and my personal experiences with working for Google.

I have embedded the magazine, neatly turned to the right page for your viewing/reading pleasure, right after the jump.

If you don't speak German, let me distill the article for you: Yes, go work for Google.

It's a special place, where you are given any and every chance to evolve as an individual and engineer. This is not to say that there aren't any other workplaces that qualify, but Google definitely gets a lot of the things I was looking for right. And then some.

Open publication - Free publishing - More presstige

WoW: Wrath of the Lich King Review for games™

2009-02-13T18:54:00.009+01:00

Oh Noes! A fan of our podcasts just dropped me an email, inquiring about WoW matters and made me realize that I totally forgot to plug my very own review of World of Warcraft: Wrath of the Lich King in games™ issue 01/09.

When Soenke Siemens, associate editor in chief of German games™, asked me for a WotLK review for their print publication in October 2008, how was I supposed to say "No!". Especially given that such an inquiry translated to an excuse for burning time hunting down Arthas in Blizzard's latest extension to Warcraft lore?

My personal history with WoW started as early as fall 2004 with the US Beta and has been unfolding ever since. Admittedly, the journey hasn't been without frequent pauses of several months. After all, real life obligations require their share of attention and levelling up too :)

So, I guess it shouldn't be surprising that after my initial three weeks of intensive adventuring in Northrend, thanks to a beta key provided by good friend and former Google colleague Sumer O. (Blizzard employee, now) and the big ol' collector's edition box by the magazine, I had to attend to my other - i.e. RL - character, Mustafa.

I might get around to pick up a little WoW-ing over the course of the next four weeks though. If you feel like teaming up with our friends & family guild on German server Khaz'goroth, I suggest you drop me a note. I have a dormant US account too, but chances for me to pick up that one too, while living in Germany and barely finding time to dwell in the European server pool, are currently very low.

Distilling the essence of the massive expansion to a two page review wasn't easy, especially when aiming for an original article that does not mimic a fact sheet or worse resembles little more than a mutated press release.

Judging from the feedback we received for the article, coming up with a compelling review seems to have worked though. Thanks to everybody for the kind words and feedback.

If you're interested and lucky, you might still be able to pick up a copy of the issue at a newsstand of your choice, before the February wave of magazines washes it all away.

On the other hand, thanks to games™'s PDF offer, you can always opt for the cellulose-free way of digital distribution and get access to any back issue you'd like ;)

Review: I Love Katamari (iPhone/iPod Touch)

2008-12-20T13:50:00.020+01:00

Roughly a week ago, in a surprise move, Namco released an iPhone game based on one of their more recent and creative IPs, that is Katamari Damacy. The official title of the version for Apple's handheld combo is dubbed I Love Katamari.

The basic game mechanic remains unchanged to what fans of the series have come to expect from the various incarnations of the title for the PSP, PlayStation 2 and Xbox 360.

The goal is to roll around a sticky ball, that is a Katamari in the game's lingo, and snowball it to a certain size by rolling over items strewn around in 3D stages.

As you roll around the first level of five for the iPhone title, you quickly realize the nature of the challenge ahead: you can't just roll up objects in any arbitrary order.

Items of a certain size will only be picked up by a Katamari that has reached a respective minimum volume. Thus you are required to iteratively identify objects you can pick up as the monstrosity you are pushing through the level reaches bigger and bigger sizes.

I Love Katamari features four modes which provide for slight variations while keeping the core game mechanic intact. In Story Mode you get a taste of the crazy background of the series, where you are prince to the King of All Cosmos, who requests you to roll up a particular object, varying from stage to stage and involving items such as a dog or a truck.

You start out with a blank Katamari that is just big enough to pick up the smallest items, depending on the stage setting these range from sushi to footballs.

A tight time limit forces you to hurry up and grow your Katamari as fast as possible in order to be able to pick up the requested object. Stages for all playmodes are made available subsequently through successfully completing challenges in story mode.

Once you have cleared a stage it is available in any of the other three modes. Time Attack gives you the opportunity to roll up a Katamari as big as possible within a set time limit, Exact Size Challenge requires you to do exactly that and Eternal Mode is your chance to enjoy stages without having to watch any limitations while building a behemoth to your liking.

The standout feature of the iPhone/iPod Touch version of the game are the controls. Navigation through the leves is accomplished by tilting and turning the device in a very intuitive fashion. To me this feature made me grok the Katamari Damacy fascination for the first time, despite having played the PS2 version for a short while.

From my personal experience I consider the game absolutely suitable for short bursts of playtime. Over the last five days, I have fired up the title every single day in order to try and please the King of all Cosmos while riding the subway or waiting in line.

Unfortunately not all that shines is gold and thanks to frequent slowdowns, occasional sluggishness and control issues every now and then, I Love Katamari makes sure we don't forget.

More than the obvious technical bugs, I am disappointed by sloppy design decisions. For instance, at certain points your Katamari goes through a distinct growth phase - which by the way always causes a slowdown and makes the framerate drop to something around 2-3 fps - where I'd expect to see some kind of consequence to the sudden growth.

Something along the lines of all of a sudden being able to pick up significantly larger objects than was possible before. That usually doesn't happen though until you've rolled over a couple of more items, that would have already stuck to the Katamari before the growth sequence, which doesn't help to make sense of the whole growth animation/framerate meltdown.

Despite the issues that the game suffers from and which Namco will hopefully address in a future update, my verdict is very positive. This is quite in contrast to the conclusion Luke Plunkett draws in his micro-review on Kotaku and Chris Kohler on his Game|Life blog on Wired.

If you are in for a game that does not mimic the bad half of 80s game concepts - as they seem to be flooding the AppStore - I suggest you grab a copy of I Love Katamari and start nagging Namco to release a patch ASAP - not telling them of course that you already enjoy the game - as I do ;)

Update - Dec. 22nd, 2008: Namco has released an updated version of the game which resolves most of the technical issues I Love Katamari was suffering from. The title in version 1.0.1 plays much smoother now.
The bug fixes also result in a significantly decreased difficulty level. For instance I was trying to get past the Park stage for the last two days, but failed miserably each and every time I attempted to. After the patch, it took me exactly one try to roll up the requested light truck and pass the level.

The UI stayed pretty much unchanged, aside from an element showing the extent of tilt the iPhone/iPod Touch sensor reports to the game. I suggest you either update your version of the game, if you have purchased it already that is, or use the new release to get it in the first place ;)

Space Siege Preview for games™

2008-07-17T16:23:00.004+02:00

My Space Siege Preview for gamesTM, originally uploaded by Mustafa K. Isik.

As listeners on Germany's most popular gaming podcast games und so already know, Alex and I met up with Chris Taylor and talked about his studio's latest game Space Siege.

gamesTM asked me for a preview of the title which has been published in their latest, i.e. August 2008, issue. The magazine is available at newsstands and bookstores as well as a slick DRM-free PDF online.

It is jam-packed with previews (e.g. Civ IV: Colonization, CoD: World at War, Treyarch's Bond, Dawn of War 2), reviews (Battlefield Bad Company, Alone in the Dark) and a LucasArts Indy retro flashback - well worth shelling out the bucks ;)

Google Tech Talks: Wiring Hacker Synapses & More

2008-07-09T15:57:00.004+02:00

Some of the talks given at EclipseDay at the Googleplex are available as videos via the Google engEDU Tech Talks Channel on YouTube now.

For your (and my) convenience, I have embedded our very own Eclipse Communication Framework Wiring Hacker Synapses talk right after the jump. But if you are really in for a ride to remember, I suggest you indulge yourself in the full list of EclipseDay Grand Teton room recordings.

Eclipse @ eBay, Michael Galpin
How Mylyn Changes the Way I Develop, Bjorn Freeman-Benson
GWT in Eclipse, Bruce Johnson
Plug-in Development Tips and Tricks, Chris Aniszcyk
and Wiring Hacker Synapses, Mustafa K. Isik & Scott Lewis

Wait - there is even more ;) Photos.

Visiting the Friendfeed Global International Headquarters ...

2008-06-28T16:59:00.015+02:00

... and helping out with the move.

Since I was in Mountain View, CA for the first half of this passing week, I used the chance to make up for all the missed friendfeed open house and TGIFF invitations. On Monday afternoon I took a break from preparing my slides for EclipseDay at the Googleplex and dropped by the Friendfeed Global International Headquarters - new and old.
It was great to finally meet up with the team behind the service of which not only I think that it is excellent in scope, execution and promise.
After a warm welcome by Ana and a chat with Paul, I joined the team in moving from the old office to the new location, contributing my share to friendfeed's continued success ;)

Among the non-trivial tasks of relocating a global international headquarter was mounting a high-tech mailbox to the precise specifications laid out by Ana - that is "So that our short and chubby mailman can reach it". Paul and I had to make more than one attempt to get it right.

Over the course of my stay I was successively introduced to the rest of the team and I have to say you guys rock. I had a blast talking to you all and I am sure that my Monday afternoon friendfeed diversion helped with the success of the talk at Google the following day :D

P.S.: Ana thank you for the shirts. I took mine out for a walk in Munich today. My girlfriend says thank you for thinking of one for her too - since you and I could not find the girl cuts in the box, it looks like as if I'll get around to wear that one too though :)

Corresponding items on my friendfeed:

Hot Off the Press: Slides for Cola Tech Talk "Wiring Hacker Synapses"

2008-06-25T02:44:00.007+02:00

These are the slides for my Cola: Real-Time Shared Editing talk at Google for EclipseDay.

You can download the PDF version of the slides by clicking through to the scribd page for this presentation.

The embedded slides/PDF version have been stripped off any animation, that is are fully built and thus should be well suited for non-presentation style consumption.

Wiring Hacker Synapses

Cola: Real-Time Shared Editing - Screencast

2008-06-19T07:54:00.001+02:00

The Ganymede release is looming on the horizon and the Eclipse community is busy wrapping things up. Our team - the Eclipse Communication Framework - is no exception there.

In the course of getting ready for launch, I have put together a short screencast, showcasing the first incarnation of "Cola: Real-Time Shared Editing" for Eclipse.

Follow my friend Thomas K. and me, lead you through an exemplary shared editing session.

Cola: Real-Time Shared Editing from Mustafa K. Isik on Vimeo.

You can enjoy high resolution versions of the screencast at the vimeo page (higher than the embedded version) or by clicking on the HD icon when doing a mouseover (full 720p glory).

EclipseDay hosted by Google: Attendee list filling up fast

2008-05-30T12:04:00.005+02:00

On Tuesday June 24th, 2008 we will hold an Eclipse event at the Googleplex. The first of its kind, hosted by the great folks of the Google Open Source Programs Office (OSPO).

As Ian noted on his blog, I suggested an event of this type while I was interning with the Build Tools team at Google last year. Soliciting the eclipse committers mailing list for feedback concerning a potential DemoCamp at Google revealed great interest in the community and knowing of Google Engineers' love for all things technologically slick - and Eclipse definitely qualifies ;) - ultimately led to an event of larger scope.

Ian Skerrett, Marketing Director for the Eclipse Foundation, Leslie Hawthorn, Program Manager for Google's OSPO and Rob Peterson, Manager of the Google Build Tools team, signed up for the idea right away.

We started brainstorming ideas for the conference and bouncing them off of each other via emails and phone conferences late last year and have since been supported by more members of the Google Build Tools team and OSPO. I would like to single out my host Rob Konigsberg, colleague Rob Clevenger and Tiffany Griffith for their efforts on making this event happen.

Even with so many people involved and the event having developed a dynamic of its own, I like to think of EclipseDay as my personal Waynestock and if you attend, it shall be yours too.
I hope for a similarly fun get-together :)

The Google-hosted very first EclipseDay will manifest as a gathering of one hundred eclipse-enthusiast developers. The talks will provide for a general hands-on approach and should prove to be insightful for a broad audience. I suggest you have a look at the agenda.

Space is limited so please sign up if you would like to join - only nine spots are left on the attendee list.

It would be great if you could make it to the talk Scott Lewis and I will give on ECF work: "Wiring Hacker Synapses: Collaborative Coding and Team Tooling in Eclipse"

See you at the Googleplex.

View Larger Map

On Being an Easter Egg

2008-03-04T17:56:00.007+01:00

This morning I woke up to the realization, that I had been easter egged.

I can't tell for sure whether it had something to do with the fuzzy feeling in my head or an email I got from my Google Santa Monica buddy and Wine hacker Lei Zhang, asking me to have a closer look at a screenshot in his Google Lat Long Blog post.

Thanks Lei, this made me feel a little like Sam & Max in various LucasArts adventures. I'd like to share the spot in the limelight though - Lei is the second from the left :)

Macbook Pro Modding: Resurrection of the Rainbow Apple

2008-02-22T19:13:00.017+01:00

The first computer I consciously laid hands on was an Apple II.

It must have been sometime between 1983 and 1985 while visiting family in Dortmund, Germany. The exciting beige box, crowned by an amber monochrome monitor, belonged to Uncle Celal, a cousin of Dad and computer science student at the time.

I vividly remember playing Choplifter on the huge, at least to my childish eyes, machine for years to come - and I remember the Apple logo on the box. The prominently featured, colorful symbol of joy.

Even though given this early and intensive experience, I did not turn into an uncritical fanboy. For most of the early nineties I was much more interested in the Amiga - for the games - and IBM-compatibles for their flexibility, my teenage love for Turbo Pascal and even more games. I enjoyed to hack then popular fire demos with inline assembler, devise systems for my friends and then assemble them.

But the Apple logo stuck with me ... unfortunately it did not with Apple Inc.

Apple came up with Mac OS X, I started to use their machines as of 2001, but gone was my chance to ever own a piece of equipment bearing the Rainbow Apple.

Until recently that is. Browsing the web for a classic Apple logo sticker that I could use on my Macbook Pro, I came across iColours.ca, a Canada-based retailer selling plastic inlays to be used with the backlit Apple logo behind notebook screens. I was instantly sold.

Below I have documented the process of opening a Macbook Pro and replacing the standard white fill with a custom inlay.

In order to remove the back panel of the notebook display, you have to unscrew two tiny screws to the far right and left on the bottom of the screen frame.
Once you are done, use a sturdy but thin credit card to carefully pry open the screen casing. I suggest you start inserting the card/your-other-tool-of-choice towards the very bottom of the screen. Work on the inner side of the thin plastic rim that circumvents front and back part of the notebook screen casing.

Work your way towards the top on both sides of the screen. Three latches on either side of the screen hold the front and back parts together. Be careful not to break any of them.
After loosening the sides, the back panel will slide off towards the top of the screen.
Remove the white fill cover above the logo window.
Use some heat-resistant adhesive to secure the custom inlay in place.

The rest is straightforward reversal of the opening steps to close and finish your custom logo MBP mod.

Augmented Reality Tracking with Wii components

2008-02-22T08:53:00.007+01:00

The AR Roadmap spin-off prototype we built for BMW Research in 2004, relied on tracking a large number of items equipped with (passive) infrared reflective markers. We used high-end tracking cameras built by advance realtime tracking GmbH, capable of tracking a high number of markers with good resolution/precision.

If you would like to play around or even build your own homegrown AR system, Johnny Chung Lee from CMU's Human-Computer Interaction Institute, shows how to use components of Nintendo's Wii game console for active marker-based tracking. The WiiMote is used as tracking device, transmitting location data of active LED markers, as for instance the Wii sensor bar, to a PC via Bluetooth. His C# SDK and more info on his approach are availabe via his project website.

The results are impressive, well unless you need surgical precision, but then again that can easily cost you 10K+ USD/EUR.

Enjoy Johnny's video on the head tracking system built by utilizing the WiiMote.

Feature Detection with Phosphorous Make-Up @ GDC

2008-02-21T15:05:00.003+01:00

Sometime in 2002 and to no little extent thanks to Prof. Gudrun Klinker, I developed a profound interest in Augmented Reality. Until 2004 I attended all classes and lab courses offered on the topic and worked on a couple of related research projects, e.g. FixIt: An Approach towards assisting Workers in Diagnosing Machine Malfunctions.

FixIt: Diagnosing Machine Malfunctions with AR - Upload a doc

At some point in 2003, Gudrun approached me with an offer to collaborate on a project at BMW Research, yielding the BMW Augmented Reality Roadmap, which doubled as my Bachelor's Thesis. I helped with the realization of a spin-off from the BMW AR Roadmap, project Ida: An AR System for Visualizing Deviations between the Real and Planned Shape of a Car Body.

Ida: Visualizing Measurement Data on a Car Body - Upload a doc

I figure given this intro on how I was involved with AR, it is understandable that a video on facial feature detection, that Joystiq recorded at this year's GDC, got me all fired up. Tracking of real-world elements is key to all Augmented Reality systems, integration of reality and virtualizations usually being a primary concern.

There is a wealth of sensor- and marker-based techniques to aid with the detection of real-world objects, specifically their location and orientation. Though the holy grail of tracking is feature based tracking, that is tracking of objects without the need for modification of objects of interest.

Steve Perlman of motion capture studio Mova describes more-or-less markerless (aside from phosphorous make-up) tracking of facial features with impressive results.

GDC - Mova Contour Reality Capture Technology from Joy Stiq on Vimeo.

If all this AR talk made you curious, I suggest you head over to Gudrun Klinker's sites for her Introduction to Augmented Reality and Advanced Topics in Augmented Reality courses. Her slides should provide for a good introduction to the subject.

Google Talk and GMail Unicode Support

2007-09-15T05:36:00.000+02:00

Due to my Turkish heritage, my name is riddled with non-ASCII-conforming characters.

The small letter "s" at the end of my middle name "Kurtulus" is actually a "ş", that is a small s with a cedilla.
It is pronounced like the Latin small letter "esh", the sound of it being described as voiceless postalveolar fricative, represented via ʃ in the IPA.

Both, the capital and small letters "i" in my last name, written "ı" for the small one, represent the Turkish alphabet's respective dotless letters, also known as close back unrounded vowel to the linguistic community, I suppose :)
The IPA uses the Latin letter "m" turned upside down ɯ to represent it.

Even though my German passport reads

Mustafa Kurtulus Isik

following the correct and intended spelling of my name should result in

Mustafa Kurtuluş Işık

Well did you know, that if you have to deal with such letters in your name, Google Talk and GMail won't leave you out in the rain? They support all the Unicode goodness you could ask for.

If for instance you feel like changing your GMail sender name, just go to Settings->Accounts->edit info and re-enter your name with the special characters.

You can use a UTF-8 character table to look up the four-digit hex code point encoding, which you should enter holding Ctrl+Shift on the keyboard. Upon entering the last of the digits, your desired special character will appear auto-magically.

If you lack the name, but not the desire to play around with this, why don't you go for a trademark sign ™ or an end of proof ∎ in your Google Talk status messages?

By the way this blogpost follows a marathon debugging session of open source code that I am in the process of reviving ... makes you do weird things ... ☻

Web Crash 2007

2007-07-23T12:07:00.000+02:00

Were you wondering why your office internet connection was so slow this morning?

ONN has the Breaking News on it.

Breaking News: All Online Data Lost After Internet Crash

After wiping off the tears of laughter from my face, I came to think of redundancy in the internet's topology.

Even though most of us consider internet routing paths to be laid out greatly redundant, the truth requires a little more thought, as for instance becomes apparent by some comments on Bruce Schneier's blog.

The whole concept of the internet is not compatible with the notion of being breakable. It is parts of the network of networks that carry the risk of being weakly linked to the rest of the world.

The internet protocol (IP) does not require packets to follow predetermined paths, thus allowing for data to be routed flexibly and with a high probability of arrival at its intended destination.

Just too bad when an ISP does not provide enough redundant connections out of its own network. All the routing flexibility of IP does not help, if there is only one path to the outside world, constituting a single point of failure.

Redundancy costs money, without necessarily providing any imminent and easily perceivable benefit to customers.

The lack of redundancy marketability and network outages being potentially cheaper to ISPs than continuous investments into redundancy measures, makes such efforts even less attractive.

For ISPs having to cater to the free market, costs and prices are much more powerful figures.

With the internet becoming more and more important for almost everybody's personal and business lives, wouldn't this be an area where government regulation, such as requiring a minimum number of connections to different Tier 1 ISPs and/or IXPs, would make sense?

iPhone: Stroking vs. Poking in UI Design

2007-07-21T12:41:00.000+02:00

Some facts for starters:

I am a geek.
Robert Scoble went a little overboard with his enthusiasm for the iPhone ...
...but he still hits that occasional interesting interview.
Thanks to Apple's devilish gadget, Stephen Colbert of the Colbert Report sees us marching to our enslavement.

Of course me being a geek has translated to me being all curious for everything iPhone-related over the last weeks. With seemingly everybody else's anticipation for the little device, especially Mr. Scoble's, exceeding the human sanity spectrum, I was determined not to post anything related to Steve Job's one ring to rule them all.

But just like in good old Channel 9 days, the Scobleizer managed to run into somebody that had something interesting to say.

Nicole Lazzaro shares insights on how user input on the iPhone via stroking gestures vs. conventional button poking makes a difference on how the device is perceived and the emotions users associate with it.

At the same time Stephen Colbert warns of Apple's hidden agenda to enslave us all :)
Enjoy it while it's hot, since comedy central embraces the not-so-bright practice of outfitting some of their video clips with expiry dates (for some even-less-bright licensing reasons).

Toys can change the world - Will Wright on TED Talks

2007-07-18T08:14:00.000+02:00

I really like the TED conference. I have been treating myself to quite a few of their recorded talks ever since Google helped them publish an initial batch of presentations.

More recently the TED conference launched a website to host many more of their recordings, all freely available, the latest of which I enjoyed a lot.

In March 2007 Will Wright - one of my (two) handful of personal game design heros ;) - held a presentation titled "Toys that make worlds" at TED.

Over the last two years, especially at GDC, Will Wright has given very good presentations featuring his upcoming title Spore and the reasoning behind its concept. This talk is different from past demos, in that it focuses more on the educational aspect of the game.

Will gives insights on how he thinks Spore might foster long-term thinking considering environmental issues. The demo of the game he gives throughout the talk shows how well suited the title is to playfully convey an idea of evolutionary scale.

If the term edutainment would not have been overhyped and associated with so many mediocre software titles, that were neither educating nor fun, this title could become a prime example for what that class of software should have been like.

You might want to pay particular attention to

the mating habits of slug-type creatures (and the accompanying soulful music) at 00:05:11
the editor which seemingly makes modeling entities to your liking a breeze while not being overwhelming at 00:05:43
procedurally generated texture maps at 00:07:11
procedural animation of dynamically created beasts that looks natural at 00:07:31
terraforming and interplanetary travel :) as of 00:10:45
and Will's thoughts on helping to change the world in terms of encouraging long-term thinking at 00:16:07

If you enjoy the content of Will's TED talk and his presentation style, GDCTV has a recording of his 2005 GDC Game Design Keynote speech up for sale. If you're into game design or electronic gaming in general it is worth the fifteen bucks.

Google Video hosts a whole bunch of Spore videos from past GDCs and E3s.

Parameter Passing in Java

2007-07-12T10:46:00.000+02:00

A couple of days ago, I felt like refreshing on refactoring knowledge by digging up my copy of Martin Fowler’s Refactoring: Improving the Design of Existing Code.

Once I reached Remove Assignments to Parameters (131) I came across past notes of mine, in color :) - which usually is a good indicator for raised attention levels.

The refactoring item elaborates on the bug-inviting practice of reassigning values to parameters.

Upon re-reading the refactoring and my notes, I remembered that prior to initially working through the book years ago, I thought that in Java primitive parameters were passed by value whereas objects were passed by reference.

Even though the item points out that this isn’t so - all parameters are passed by value - I would have liked to see a little more light shed on the issue.

At first thought, the Java language’s strict adherence to call-by-value parameter passing conflicts with the notion of being able to modify passed objects.

In contrast to primitives, objects do not live on the call stack. They reside on the heap instead and outlive method execution. Object states can be modified from within methods. Giving in to intuition this sounds much more like call-by-reference handling of non-primitive parameters.

It is assignments to parameters within methods, that proves intuition wrong.

A tiny example is to serve to illustrate the issue. Class Foo is trivial, just holds a private int instance variable, initialized via its constructor, and provides getter and setter methods.

class CallByValueGoodness {

private static void incrementBar (Foo aFoo) {
  aFoo.setBar( aFoo.getBar() + 1 );
}

private static void incrementReassignedBar (Foo anotherFoo) {
  anotherFoo = new Foo( anotherFoo.getBar() + 1 );
}

public static void main (String[] args) {
  Foo firstObj = new Foo(5);
  Foo secondObj = new Foo(5);
  incrementBar(firstObj);
  incrementReassignedBar(secondObj);
  System.out.println(firstObj.getBar());
  System.out.println(secondObj.getBar());
}
}

Running the main-method will output:

6
5

Even though both Foo-objects have seemingly been updated, secondObj does not reflect any changes.

This is due to Java’s parameter passing by value.

When calling a method with a non-primitive parameter a new reference, i.e. a copied reference, is passed instead of the original reference.

Any reassignments to the parameter reference, that is anotherFoo in the example, do not influence the caller’s reference to the object.

Effective Java

2007-07-05T15:08:00.001+02:00

In 2005, in my everlasting pursuit for slick coding challenges, I took some time off college to work at Avid Technology Inc. as software developer. I applied for a six month internship position in the cross-platform team (developing in Java and C++) and ended up staying with them for two years :).

The work was highly interesting as was the team, which was made up of former NXN Software developers, mothers and fathers to the brainchild of company founder Gregor vom Scheidt: the Alienbrain SCM and Media Asset Management System.

Thanks to the people I met at the Munich development office, where I have to single out my friends Thomas Krammer and Austin Moore in particular, I could gather extensive experience with agile methodologies and a whole slew of better coding practices. During that time I extended my personal library of required and essential reading by a whole bunch of books, all very interesting and insightful. One of the books finding its way into my mahogany paneled, wing chair equipped reading room and claiming an extra-special spot, is Joshua Bloch's Effective Java Programming Language Guide.

The book is an excellent compilation of subtleties of the java language and class library, best coding practices and insights from the author's extensive wealth of experience with all things java (among other things being the head behind the Java Collections API). The book draws upon the style of Scott Meyers Effective C++ series, presenting its material in a strictly itemized manner, making it suitable for quick reading bursts and usage as reference literature.

Effective Java having been published in 2001 does not feature any discussion on the new language features introduced with Java 5, which in my view does not diminish any of its value to java geeks though.

I suggest you don't wait any longer to get your hands on this wonderful read. In case you'd like to dwell in the land of the unknowing for a little longer, you could also pre-order and wait for the next edition of the book, Effective Java(TM) Programming Language Guide (2nd Edition) (The Java Series), which is supposed to respect the latest updates to the language.

Being a Google Tech Talks addict, I treat you to an insightful and interesting talk given by Josh Bloch on Good API Design.

P.S.: The part with the mahogany paneled and wing chair equipped personal library is not entirely true .... but can you blame me for wishful thinking :)

See it all coming together ...

2007-07-04T12:26:00.001+02:00

Now that the big guys in form of MTV and EA step in to provide some licensing and distribution muscle to Guitar Hero makers Harmonix, there is land in sight on the horizon for those travelling to rockstar island to form their own Rock Band.
Watch the latest video from Gamespot to see what Harmonix' interpretation of an integrated version of multi-player Guitar Hero, Singstar-esque Karaoke and drumstick-whirling DrumMania-type gameplay will look like. I am _so_ waiting for this ... omg more partygaming on the way ;)

Distributed Software Systems Code

2007-07-04T10:44:00.000+02:00

This semester's class on distributed software systems, more specifically the lab class going along, proved to be fairly insightful. Even though I have done my share of multi-threaded programming and using Java Sockets, I never got around to play with Java's Remote Method Invocation API or anything remotely utilizing RPC concepts for that matter.

Some of the assignments were pretty dull others for all intents and purposes interesting. I liked a modified version of Dijkstra's Dining Philosophers problem in particular, where the number of philosophers exceeds the seats at the table. This results in not only having philosophers sitting at the table competing for chopsticks/forks, but unseated philosophers challenging each other for available seats as well.

Given the slew of activities eating up my schedule this semester, I decided to not join the other students in the lab, coding under professorial supervision [for a whole semester] :) Much rather I opted to hack all of this semester's tasks on a couple of free evenings last week and the one before.

So it wasn't until turning in and presenting my solutions to the class-instructing professor, that I witnessed quite a few fellow students being totally stressed out and struggling with the assignments. Since I am not a believer in closed code, especially in an academic environment, I am sharing mine via an open source repository.

Head over to Google's excellent Project Hosting and feel free to play around with, even better improve on, my beads'n'threads code. As always, there is room for improvement. For starters you could take a look at the issue list.

Please let me know either via email or a comment to this post, if you'd like to have committer access to the repo.

I don't want to encourage my fellow classmates to revel in cheap plagiarism. But in the end it is and should remain everyone's own decision, whether to use the code as

inspiration
learning material
means of communication
basis for collaboration

or simply obfuscate its source and try to resubmit it for class (which is really not a bright thing to do).

Code shouldn't be locked up, especially in an environment where learning is supposed to be of prime importance.