Tuesday, December 25, 2007

Mastering the Mess

Mastering the Mess
An article found in ACM Communications in April, 2007.

What this articles talks about sounds quite obvious and straightforward, but not many people understand it.

When people face problems that are hard to solve, then after a while, people give up trying to solve it and they just adjust themselves to the problem and change their life around it. When someone finally finds that answer to the problem, people often resist accepting the solution because it will change the way they've been working whether it was right or wrong.

The author seems to say, the "innovator" should accept this fact of people not trying to change old ways, and try to bring gradual change without people noticing it, rather than bringing in a sudden revolution. The process could be slow but with that way, the innovative solution could eventually work out without scaring the majority.


This is another dilemma. Innovators are often hot-tempered, and fast-paced people who just can't stand the problems they are faced with. And they want to see their solutions to kill the problem right away, so they push it hard. Often it's too hard, and it scares majority of people who are just too afraid of the unexpected things that it could bring. So his innovation gets bullied and killed, even though it would've solved the problem.


These kinds of things happen everywhere. In life, in politics, at work...
Those who followed the author's approach often eventually won the crowd at the end. The ones who try to turn things upside down in a short time often get squashed by the scared majority. I think Korean president, Roh Moo Hyun, had good intentions and answers for solving South Korea's many decades-old problems, but he pushed his intentions a bit too hard - and he tried to do them all while he was in presidency, only 4~5 years. Guess what happened? His policies were too much for people to take and thus his sides was crushed by the top opposition in the last week's presidential election big time.

The ultimate convergence of IT and machinery tech

The marvel of machinery technology must be automobiles. I mean it requires lots and lots of new and old technologies related to machines - engines, suspension, tires, materials, etc.

Now there is this Nissan GT-R, which I must call it the ultimate convergence of IT and machinery. It's got so many features accomplished by IT. I hear from the news that one could change the car's settings like damper ratio, gear changing speed, and etc, with touchscreen LCD, just as easily as one could do it in a video game.

It even has an location based speed limiter. It has a GPS receiver and the limiter allows the car to be driven up to 180km/h if it's outside a race track, and allows unlimited speed if it's in the track. BTW, this is an overkill and I am sure there will be hackers who will do some tricks to disable this features.. I mean who wants to drive that nice car to drive up to only 180km/h...?


Anyways, I personally don't really like cars that have too many computer assisted features. I am probably more like European drivers, who like firm and stiff rides more and don't care much about anything else but driving when driving. (This sounds funny, by the way, but there are lots of people who buy cars not based on how it drives) One thing Japanese car still are behind Europeans is in making more rigid body structures, which help make the car feel more solid and handle more torsional force. What they lack in that department, they make it up with all those IT assisted features.

By the way, I really think European, especially German makers, could throw in those IT assisted features, but they probably just choose not to. More features require more parts, and more parts mean more costs and more possibillities for malfunctions. Also, their cars are great enough so that they just probably don't find a need for them as much.

Friday, August 17, 2007

Java VS C++ : parameter passing

Parameter Passing : Pass-By-Value & Pass-By-Reference.

Parameter passing is something people who moved from C/C++ to Java (or the other way around) get confused.

Let's just explain it with few examples.


void main () {
int myAge = 10;
string myName = getName (myAge);
exit (0);

string getName (int age) {
return "My Name";


What is the value of myAge after getName(myAge) gets executed?
in C --> 10
in C++ --> 10
in Java --> 10

In all 3 languages above, the parameter that is of a primitive type is always being passed by value. What it means is that if getName() function is defined like the one above, the parameter value outside getName() function will never be changed even if the user manipulates the parameter value (e.g. age++) in getName() function.

Then, how about this case?


void main() {
Car myCar = new myCar ("Ferrari");
int price = getPrice (myCar);
exit (0);

int getPrice (Car aCar) {
return 10000;


What will be returned if myCar.getName() gets called, after getPrice(myCar) gets executed?
in C --> ERROR! there is no class in C. :-)
in C++ --> still a Ferrari.
in Java --> Hyundai!!!!! (Where the hell is my Ferrari?!?!)

This part is the critical difference between Java and C++ when passing parameters to a function.

In Java, if non-primitive data types are passed as a parameter, reference to the data will be passed, ALWAYS! In other words, if manipulation on the parameter is done inside the function, it will be reflected outside the function as well.

To be more precise, Java never holds a object itself like in C++. Instead, it always holds a reference to an object. Also, Java always passes parameters by value, and the parameter being passed to a function is a reference value. oooh, this is not quite so easy to understand, by the way. :-(

Probably many C/C++ programmers will wonder why the name has been changed to Hyundai, or many Java programmers will wonder why the name has not been changed to Hyundai, if the same code was run in C++.

Java VS C

Differences between Java and C as described in "Java In A Nutshell"

These are something that may well be asked on a university CS exam. (e.g. "List at least 5 differences between Java and C")

1. No preprocessor (e.g. #define, #include, #ifdef)
2. No global varibles -> no possibility of namespace collision
3.Well-defined primitive types :
- Size of short, int, long types are platform (e.g. "What kind of CPU?") dependent in C.
- Not so in Java
4. No pointers
- Java classes and arrays are reference types
- Java references are "opaque" -> no way of converting it to a premitive types or to manipulate it (e.g. ptr++; )
5. Garbage Collection
6. No goto statement
7. Variable declarations anywhere : also true in C++!
8. Forward references : funtions can be used before being defined/declared.
9. Method overloading : multiple methods with same name can be defined as long as parameter lists are different.
10. No struct or union types
11. No bitfields
12. No typedef
13. No method/function pointers

Monday, August 13, 2007

rss test

Wednesday, July 25, 2007

How to make a cross-over ethernet cable

Here is how to make a crossover cable

1 : white orange
2 : orange
3 : white green
4 : blue
5 : white blue
6 : green
7 : white brown
8 : brown

12345678 on one end

36145278 on other end

Thursday, July 05, 2007

Military Service in Korea and its problems

Here in Korea, every man is supposed to be serving in a military. Exceptions apply to those with mental or physical deficiencies, whether it's a big or small, or those who are deemed to be "smart" - those ones are assigned to some companies to do some more "productive" works.

The problem with this system is that people who served in a military, whether they liked it or not, are losing 2-3 years of their young career, and the society as a whole is not kind enough to compensate them for their "loss of opportunity". The root of this problem is that the Korean society in general is not so appreciative of the people who served in military. Okay, this is just my opinion, but you should agree with me on this one if you compare how the military personnel or veterans in U.S., Canada, U.K., or some other countries are being praised and appreciated by its people and society on a regular basis.

More importantly, it doesn't matter what the person did in the military. He may have assasined Bin Laden, he may have flied jetfighters, he may have built roads for military use, he may have destroyed a whole country all by himself, or all he may have done was cooking meals and washing dishes. It simply doesn't matter what he did. If he (or she) served, then he'll get appropriate and deep-hearted praises and appreciations from general public.

In Korea, this doesn't happen. And when people who served in military comes up with a question like "We are disadvantaged, so we do want some sort of compensations in some forms!" Then the others (e.g. women or people who worked at some companies) often go nuts and say compensating them is a "unfair advantage to the selected few" and "discrimination" against them. Bullshit.

I think what the people who served want is not money or anything material. What they truly need is proper and warm appreciation and recognition for their time and sacrifices.

The problem is most people do not see this fact.

Saturday, May 12, 2007

A Video that puts smile on my face.

Watching the Brazillian players in this video makes me think
When you love what you do and enjoy doing your works, that's when you are becoming goot at something.

Motorola KRZR and my thoughts on...

About a month ago, I lost my cell phone - A black LG "Chocolate" phone. I got it when it first came out and I've had it for more than a year. Where did it go? It somehow slipped out of my pocket in a taxi on my way home from seeing my buddies.

So after about a month with a big, bulky, yellow! Casio phone that I digged out of my cousin's closet, I finally got a new phone - a black Motorola KRZR.

Having used that Casio phone, it's a huge upgrade in terms of how small it is and how it looks. Anyways, after many years of having used LG and other Korean made phones, this is my first taste of an American designed/engineered phone. Its material and finish is superb and I am quite happy with it.


It doesn't seem to have as many functions or features as Korean made phones. For example, it doesn't have as much internal memory as other phones here in Korea as a lot of phones come with from 512MB to 1GB of internal memory. Even my lost LG Chocolate Phone had 512MB.

By the way, I think Asian companies have a tendency to throw just about anything they have in their hands to a product and hope people will like it even though many of them probably will never use them at all. As for the internal memory, I have never used any more than 40-50MB when I had the LG phone.

I think American companies like to concentrate their efforts on few features - features that they think people will use the most - and just take out anything else. It is easier on engineers, too, as they have few things to work on.

Why is there this subtle difference? It's probably due to the consumers. In my opinion, American consumers like things that are simple and easy to use (think iPod!), yet Asian cosumers like things that are just fully loaded (even though they'll probably use it once or twice!)

Which one do I like better? I like products that have an excellent fit and finish, look good, have more features for the same money all at the same time! Am I too sophisticated? Probably not... I am sure we all want such products!!

Monday, May 07, 2007

Going back to school..?

Lately, I've been thinking of going back to school for further study in a field other than computer / engineering related. Probably I am running into a so-called "3rd year salaryman crisis" as one of my friends at my last work described.

Last time when I visited my University, looking at the people with backpacks and books running around campus for classes made me think "geez... I don't want to go thru this again. Thank god I am out of school!"

I don't know what has happened to me since then.. maybe it's my desire to know more about economics / finance / business. I always thought those things were somewhat useless or could be learned better in a real world, but something has changed. Maybe it's books that I've read lately. But the bigger thing is that I finally realized that to be a better "engineering architect businessman" I have to know more about other sides - not just how to program multi-thread applications in object oriented ways using C++ in Linux. Also people who know both engineering and other stuff, both of them quite well, are in high demands, as there are just handful of them.

Oh well... I don't know if I can really go back to school, as there are things I will surely lose for something that is not 100% predictable. I have to begin my own analysis on opportunity costs of going back to school. And plus, I don't know if my marks in undergrad is good enough for getting me into some of the well-known, good schools. Plus..... this is sad facts, going to school is costing a lot!

Will I view it as a pure expense, or an investment? That, I have to do more investigations and analysis.

By the way, the schools that I wish to go to..
- London School of Economics (LSE) : I don't know.. I just have some fantasy about that school.
- Columbia University : NYC!
- Univ of Chicago : where "Freakonomics" authour teaches at!
- Univ of Toronto : hometown school. easier to get together with family and friends. plus tuition is the lowest among all as I am a Canadian.

Tuesday, April 24, 2007

Steve Nash the footballer?

For those of you who don't know who Steve Nash is.
He is a 2-time NBA MVP - a basketball player.

Having car accidents is a lose-lose game

Driving thru busy streets of here in Korea today, I saw some people jaywalking a very wide street where cars are driving at 80km/h.

At the same time, the thought and scene of that person being hit by a car went thru my mind. When it happens, we all lose. It's a lose-lose situation, an opposite of a win-win situation.

First, people get hurt physically and mentally.
Secondly, money and time has to be spent to repair cars and people. The resource is something that didn't have to be spent. It's basically a waste as it doesn't add any value to the society or any individuals.

The only plus side of a person jaywalking? He could have probably saved a little time, maybe 5 minutes at most. But is his 5 minute saving is worth more than his well-being and his and others' resources? I don't think so, but there are still a lot of people thinking otherwise. They are very selfish bastards, in my opinion.

Certainly, people need more serious education in how to calculate things and economics/sociology.

Wednesday, April 18, 2007

Linux Kernel expert wanted!

Someone said here in Korea, it's super hard to find anybody who knows Linux Kernel inside out and be able to tune it if necessary. It's so true. The company I work for right now has like 1300 employees, and I have not encountered a single person who are Linux Kernel guru. Maybe I haven't met many people, but when I look up the names of the teams and departments, most of them are about developing the front end GUI and user applications.

It seems like the people who program and tune at "lower" abstract level are becoming more and more scarce. Maybe there isn't much demand for them - after all, how many companies are willing to write their own OS or tune their OS? Most of them are just using whatever it's currently available off the shelf, and my company is one of them.

Anyways, if you know of anybody who's good at Linux Kernel here in Korea, let me know, please.

Tuesday, March 27, 2007

OpenSSI on SATA HDD controller

OpenSSI Kernel (version 1.9.2) comes with a number of SATA driver modules.
The most common SATA driver used is ata_piix.ko

Here are the issues that I've had this driver.
Apparently, the situation is as follows.

1. Dell Dimension 8150C - requires ata_piix.ko ver 1.05
2. Fedora Core 3 comes with ata_piix.ko version 1.03 --> You can't install FC3 on Dell 8150C with a "normal" way.
3. CentOS 4 comes with ata_piix.ko version 1.05 --> CentOS 4 can be installed nicely.

Note : Both Fedora Core 3 and CentOS 4 is on Kernel

4. OpenSSI comes with ata_piix.ko ver 1.03 --> Even if you could've installed CentOS 4 on Dell 8150, if you installed OpenSSI there and rebooted the machine, you are screwed because the new kernel image (OpenSSI) tries to load ata_piix.ko version 1.03.

Solution :
- Find and add ata_piix version 1.05 source code and its related files to the OpenSSI's source kernel tree.
- Rebuild the kernel.

But, the question is...
1. What files should I add/modify/update/delete?
2. What parts of the codes should I modify?
3. Is there any subtle difference in other related dependent files (e.g. scsi_mod)?

Since OpenSSI is not very widely used, there is no support from the h/w vendor or anybody, at least from anybody that I know of. I will probably dig everything by myself.

Arghhhhhh.... what a nightmare...
Linux sucks in a situation like this.

Sunday, March 25, 2007

Some interesting facts

1. The year of 1971 :
I never knew the year of 1971 was such a significant year. It was 1971 when the U.S. government (led by Richard Nixon, the president by then) abandoned gold standard. From there on the dollar became just a piece of paper. It could no longer be guaranteed to be exchanged to gold. Apparently Americans have been buying too much of goods from Japan and Europe, and Japanese and Europeans were exchanging dollars for gold from the central bank of the U.S. (Federal Reserve) The U.S. Government by then didn't like their gold leaving the country, so they decided to abandon the gold standard. Was it a smart decision? Probably not, because it gave them a power to print money as much as they want.

2. From Washington (1789) to Clinton (2000) administration, a total of $1.01 trillion dollars were borrowed from foreign governments and finacial institutions.(how many 0's are there in a trillion, by the way?)
Between 2000 and 2005, the Bush administration borrowed the total of $1.05 tirllion dollars.

Friday, March 09, 2007

Top 10 Goals

Yes, Top 10 OWN goals!

Thursday, March 08, 2007

Tuesday, March 06, 2007

Use of CDSL (Context Dependent SymLink) in OpenSSI

CDSL (Context Dependent SymLink) feature in OpenSSI can be quite useful.

First of all, what is it?
Normally, when you create a file or dir on a OpenSSI cluster, there will be only one copy/image/piece (or whatever you want to call it). That means node 1, node 2, or node N all access the same file when a process running top of that particular node tries to make a file I/O.

There are times when you want to make this file I/O independent of the cluster. For example, if you want "onnode 1 vi /testdir/hello.txt" and "onnode 2 vi /testdir/hello.txt" to load different /testdir/hello.txt file (that is, depending on which node the vi process is running on), how would you do that?

The answer is to use CDSL. It's quite simple. Simply type
"mklocalfile /testdir/hello.txt" will turn the file into a CDSL.
"mkglobalfile /testdir/hello.txt" will turn it back into a normal file.

Monday, March 05, 2007

Joel On Software : Translation into Korean

Why aren't there no up-to-date Korean translation of Joel On Software?
I can find Chinese, Japanese, and other languages that are quite up-to-date. Not in Korean, uh huh!

So, let me do it!

Here is my first work.

Joel On Software 한국어 페이지

Calculation Office Spaces - 사무실 사이즈 계산하기

Sunday, March 04, 2007

Linux Cluster explained!

This site explains quite well about overview / definitions of what a cluster is and its various use.

Things I've done at LGE from 2005 - 2006.

May 2005 - Oct 2006 at Mobile Communications Division in LG Electronics

- Wrote scripts to do automatic nightly builds.
- Customized end user menus for various languages.

KG 920
- Coded in C on top of ADI ARM 7 baseband processors.
- Used ClearCase for collarborative works.
- Participated in design and implementation of Layer 1 API for file system I/O and - multimedia feature control.
- Developed end user UI for File Manager, Video Recorder, Video Player, and Photo Viewer.
- Incorporated Multimedia features with basic features such as voice call and text messaging for multi-tasking.

GS 3
- Experiments of new development platform for new ADI ARM9 baseband processors.
- Successfully ported KG920's multimedia features to the new platform.

Wednesday, February 28, 2007


For this 9-10 minutes, I couldn't help myself standing still and just watching the TV. A performance like this makes me think about all those times that I've tried NOT to take music lessons. If I had a sense of appericiation and mind or saw performances like this when I was a kid, I might as well be a pretty good amateur musician now.

Tuesday, February 13, 2007

How to write a document that sells.

1. A short introduction.

2. Explain current situation -> problems and challenges that users/customers are faced with but do not realize they are really up against them. In other words, you have to find/create needs for them!

3. Offer various solutions that are available.

4. In the end, give your solution that beats all others.

A good example : http://www.lustre.org/docs/selecting-a-cfs.pdf

Sunday, February 11, 2007

How fast can Hyundai manufacture a car?

- Jan 30 : I met a dealer nearyby my place and said I would like to get a 07 Hyundai Click (This car is not available in U.S. and Canada)
- Jan 31 : He put an order
- Feb 3-4 : A Weekend
- Feb 6 : My car is done built.
- Feb 7 : Delievered to me.

So it took like 3-4 days for Hyundai to manufacture my car.

Pretty quick, isn't it?

Monday, February 05, 2007

How quick will Google search engine be?

I just typed "tkang blogspot openssi" in google.com's search box to see if my previous post can be searched. Of course, Google search gave me no result since I uploaded my last post few minutes ago.

It will be interesting to see how long it will take for Google to search for my previous post.

Finding the best node in OpenSSI cluster

When programming with OpenSSI's API (cluster.h & /lib/libcluster.so), there are times when you want to find out the best available node. There are two ways to find the best node.

1. Easy way - Use CLUSTERNODE_BEST macro

- Defined in cluster.h file.
- Returns physical memory address that points to the node number with the lowest load level, hence printf ("%d", CLUSTERNODE_BEST) will print some strange number.
- The speed at which this value gets updated is slow.
- Can't exclude nodes --> a node will be picked from all available nodes.

2. Alternate method - write your own function!

How do we know load levels of nodes?
- *_/proc/cluster/nodeN/load{_}* file contains node N's load level, hence we will compare all values and find the lowest node number.
- This is also useful if you want to find the best node from a specified group of nodes.
(e.g. Pick the best node from all nodes except for master node (Node 1))

I prefer to use #2 because it gives me more control. It is probably more costly to use, though.

system() vs fork()-exec()

system() vs fork()-exec()

system() :
- Uses a shell and the shell does the actual execution
- Dependent on the currently used shell.
- There will be two processes created as a result.
- Not as efficient.
- Easier to use?

fork()-exec() :
- No shell used - only one process will be created
- A little bit more complex to use.

If you want to run some commands with pipe operations (e.g. |, >>, <) then using system() is probably much easier.

Sunday, February 04, 2007

Samsung's MP3 Player?

Samsung is promoting their new mp3 players a lot. They appear in a lot of music videos these days. Pretty good marketing, indeed!

About iPod's Reliabilities

Is iPod a reliable product?
From my experience, it's not so reliable.

My last iPod with 20G hdd is dead after about 1+ years of use. Too bad it broke down after a year. What's even worse was to fix it, it would have costed a lot of money to fix it since it's more than a year old. So, like many others who have a broken iPod's with its warranty expired, I ended up buying a new one.

This new one (4G iPod Nano) that I have is starting to show some signs of.... rustiness. Today, it just got hung up. I couldn't turn it off, I couldn't do anything with it until I plugged the USB sync cable.

Would I buy another Apple product? Probably. Apple products are sort of like Italian or British cars. They are very elegant in design and there's something about them that keep attracting people.
I know Apple products are probably not very sturdy and will break down after a year or so, but I will probably give it another try until someone else comes up with their level of "coolness"

By the way, I think Apple really gives no shit about building sturdy products.

Thursday, February 01, 2007

Time to unload e-bay stocks?

Isn't it time to unload e-Bay stocks?

First of all, I do not own a single stock of any companies right now. Nor am I in any position to influence not even a single stock.

My reason on why you should unload e-Bay stocks now is very simple.

e-Bay is not like what it used to be. It sucks. Big time.

There are just too many phony cheaters out there in e-Bay, and more and more people are turning their backs on e-Bay and go to some other sites like Amazon.com when they want to buy something online.

BBQ recipe for $0.19!!! No S&H charge!!! (or course, you morons, e-mailing doesn't cost a penny!!)

Normally, when people wants to buy something from a seller on e-Bay, the first thing they do is to check the seller's rating. The problem right now is that there are just freaking too many meaningless positive ratings. If you investigate closely, a lot of those positive ratings are for selling/buying some useless/valueless stuff - like "my recipe for making a cookie." "a poem my 3 year old kid wrote." "serial number for whatever". Of course, they are very cheap - something in between $0.01 and $0.99. It's basically buying positive ratings for pennies. There are just so many of those cheaters out there with 1000+ positive ratings, a lot of people including myself would rather go to other places rather than taking a chance with them even if it may be slightly more expensive to shop somewhere else. Not to mention those assholes who sell fake stuffs.

So unless e-Bay does have some actions in plan to fix the rating systems or penalize those cheaters, it's probably a good idea to sell your e-Bay stocks if you own any because more and more people are turning away from e-Bay.

Tuesday, January 30, 2007

Interesting Story : South Korea's 'dependency' on MS

A story from CNet News.com

Yes, the stories about the police raid and how companies react to them seem true. I've heard about such stories many many times, too.

Flash HDD

Most bottlenecks when processing large sized data exists in file I/O. Two most notable bottleneck locations are HDD and networks.

Google must've known this problem better than anyone and it uses its specially designed data processing algorithms (Map and Reduce) on it s powerful distributed computing environment which also runs distributed file system (Google File System). Google's algorithm tries its best to do all the data processing in one machine so that there will be minimal amount of data transactions over the networks.

But more serious botttleneck problems exist with the current HDD. The fastest HDD out there cannot even match data trasaction rate of the gigabit ethernet which is common nowadays, not to mention fibre channel which moves data at the speed of light (in theory!) Speed isn't the only problem, either. HDD has got many mechanical parts that fail quite frequently as well.

Semiconductor manufacturers know this problem, too, and they know there are demands for faster HDD. So they come up with Flash HDD, like the one pictured above!

Someone at Samsung once said something like conventional HDD is dead and the age of Flash is upon us. I totally agree. If they can get the cost down, which they will surely do, it will be everywhere.

Maybe it's time to buy Flash manufacturers' stocks...

By the way, I wonder if Google replaced all its HDD with Flash HDD how much faster their search and other services will be.

Monday, January 29, 2007

Why companies want grid/clustering technologies?

Why do companies want grid/clustring technologies?

Once again, there is that difference between companies desire to build or deploy a clustering/grid software.

Companies with just a bunch of PC's that do not require massive computing power won't need it. Sure, it will be better if they have it, but the reward for them to use it isn't that much.

On the other hand, companies that do quite a lot of massive calculations (e.g. banks, insurance companies, bio-tech companies, etc.) must've bought those super expensive computers (e.g. a 8 cpu server, mainframes, etc.) because there are times that they need the power of those machines. Not 24 hours a day, but there are times when they seriously need them. What they really want to do is maximize the use of their investments. The opportunity cost lost for their machines to sit idle is just too much. So, they research/develop/deploy/buy grid/clustering software, and even though grid/clustering software is often expensive, its cost will be covered by the opportunity cost gained by having their super machines up and runnig at close to 100% of their capacities.

What if those expensive machines become cheaper? Then, the needs and demands for grid/clustering technologies will have to go down, too.

Higher in the ladder...

I've been working at a various companies or government jobs for almost 3-4 years.
One thing I have noticed is the difference between people's commitments to their works. This is very much like an economics question. Everyone has got different value, or different motivation, and that difference is why people behave differently at work.

Over the years I have found that people that are more "intense" at work are often higher in the ladder. If they score big, they will get rewarded big time. But if they screw up, they will get nailed big time, too. So they work hard and they want others to work hard because it makes the score big.

People who are super intense at work - for example, sweating to death or saying f-words if things go wrong - are most likely high managers or directors. Why? because if things go wrong, they will probably be fired or brought down from their positions. For the working drones at the bottom (e.g. me!), the rewards or punishments are not as big as those of the higher in the ladder and it's probably the reason why people at the bottom are more friendly and less intense at their work.

So don't go mad even if your boss or your companies directors scream at you. They are just scared - more scared than you are because of the fears of what they'll lose.

If you were in that situation, you would've done the same thing (even though you swear that you will not be like one of them, now.) and I think it's just a very natural thing.

Saturday, January 27, 2007

Top Gear - Bling Bling Escalade

This is so funny.

"A bling vehicle will get you from A to B, but you get to B very late."

Wednesday, January 24, 2007

Can't fuck with the Bank of Korea

In the days of "IMF" - those are the days in 97-99 when many Asian economies were rampaged by merciless hedge funds and financial institutions - the Korean government (precisely, the Bank of Korea) did not have any foreign exchange (US dollar) left in their pockets.

Of course, the smart and opportunitistic money makers from the world with handful of dollars flooded into Korean market and played the situation to their advantages. They knew Korean economy needed US dollars but the Bank of Korea did not have any dollar to support the economy. So what did they do? They held on to US dollars, creating big big shortages of US dollars and causing big big havoc for Korean econmoy. Eventually, the Korean government had to borrow dollars from IMF (International Monetary Fund) and people in Korea call those days "IMF period."

Those were good days for those bastards.

Now, the Bank of Korea has so much US dollar in their hands, they can match anybody or any groups who try to play around with the foreign exchange between Korean Won and US dollar. One of my close friends work for a pretty big hedge fund company in New York and he said he and his people sometimes played around with Korean foreign exchange but as soon as they senseed or heard a news that the Bank of Korea was about to make a move, they would back off. Sure, they run a lot of money, but even they are no match for the Bank of Korea.

This is the way it should have been in 1997. They should have crushed those bastards, but unfortunately the Korean gov't at that time was a stupid dumbass government who could not even fight a single group of hedge funds.

Cluster/Grid Technologies. Who uses them?

PBS Pro, LVM, Mosix, SSI, IBM's Load Leveller, etc...
I never knew exactly who actually are using these stuff.

But after meeting some vendors who try to re-sell some clustering software, I realized most big companies are using or planning to use some sort of clustering software.

One thing that I am 99.9% certain is that any big businesses who make their living by working with numbers ($$$) - financial insitutions, banks, insurance companies, etc. - all use some sorts of grid/clustering technologies.

By the way, I wonder how much they pay for those pieces of software... but I am sure it's probably way more than I can imagine... maybe in the range of a few thousand dollars per piece. Sure building clustering s/w is not so easy, but it's not exactly a rocket-science, in my opinion.

The best that I have encountered so far, in terms of features? OpenSSI!!
Its problems? Scalabilities and High Availability features.
And those are pretty big factors when choosing a clustering software and maybe they are why OpenSSI is not used as widely as we thought.

Tuesday, January 23, 2007

Larry Bird

"A winner is who recognizes his God-given talents, works his tails off to devleop them into skills, and use these skills to accomplish his goals."

Larry Bird

The truest hustler on and off the court.

Why is Hyundai more expensive in Korea?

왜 현대차는 한국서 더 비쌀까?
Why is Hyundai more expensive in Korea?

There have been news articles the same cars manufactured by Hyundai are so much cheaper in other countries (U.S., Canada, Japan, Europe, etc.) than in Korea, the birthplace of those cars.

This is simply unacceptable and very much a mystery to many people. Selling Hyundai cars in U.S. requires shipping, paying some tariffs, and a lot of extra paperworks. It's more time and money and energy consuming.

Then is Hyundai more expensive in Korea? (or Why is Hyundai so cheap in other countries?)

Let's approach this question from a different perspective.
We can ask "Why are imported cars SO MUCH MORE EXPENSIVE in Korea?"

The answer, as you may all have guessed, is the Korea's not-so-open market.
Simply put, an open market is free of restrictions and there is no friction in transaction.

However, there are so many frictions and restrictions in Korean automobile market.

1. Obvious, "everybody-can-see-it", restrictions such as import tax (tariff)

2. Not-so-obvious "not-easily-seen" restrcitions : You are treated as a cancer to Korean society/economy because you own and drive an imported vehicle, and Korean tax man goes after you the year you register an imported vehicle under your name. Doesn't matter if it's a shitbox from Ford. If it's from abroad, there goest the Korean taxman. (I don't know if this is still true.)

3. Others : Import Dealers often say the reason their prices are so high is because Korean market is so small and there is no chance they can achieve economies of scale. It's somewhat true, but I think it's just a BS. You just ship the car from its origin, mark up, and sell it.

Because of these reasons (and there are probably many other factors, too.), imported cars are very expensive in Korea and guess what it does to Hyundai. They become relatively more competitive in terms of price. They don't have to do anything. Those advantages are just given to them as if they were gifts from the God.

Hyundai is a corporation and a corporation exists for making most profits. And guess what. They do, too. So what do they do? They bump their prices up and up and up until they can sustain their pricing advantages. This is bloody obvious.

But, what would happen if they bumped their prices up in U.S. just like they would do in Korea? Then, they'll probably go out of business. Why? because imported cars (okay, Ford and GM are not "imported" cars in U.S.... anyways.) are so much cheaper there. In other words, Hyundai does not have that gift of automatic price advantages. And what that means is that, in the U.S. auto market, Hyundai is much less competitive.

Then what would have to be done if we want to see the Hyundai's price in Korea is as low as that in other countries?

Open the damn market!

Then, imported cars in Korea will have to be cheaper than they are now. Then what will Hyundai have to do?

1. Work on details, improve their products, close the quality gap between their products and imported vehicles like Toyota and BMW.
2. Lower their price!!!!!!! What else can they do?

What if the market simply won't be opened up?

Let's give an example.

- In Korea, a BMW is 50% more expensive than a Hyundai.
- In U.S., a BMW is 20% more expensive than a Hyundai.

Then, In Korea, Hyundai will increase their prices until people in Korea say "BMW is 20% more expensive than a Hyundai in Korea." (By the way, this is what is happening right now in Korea.)

In the end, the percentage difference between a Hyundai and a BMW will be the same in U.S. and Korea. But, people in Korea pay more money for the very same product. Why? All because of the closed market.

This is why closed market is no good. In the end, sellers/manufacturers like Hyundai and BMW can sell the same stuff for more and make more profits, but buyers/consumers take all the hits.

So, Don't blame Hyundai for selling their stuff very cheaply in U.S. Don't blame Hyundai for giving 10-year warranty to the buyers in other countries. don't blame Hyundai building theirs cars to be sold in Korea like junks. (Okay, this is just a rumour, but many people believe so and it may be true...)

What they do as they do is very natural. They want to make profits and squeeze the market until there is no juices left. It is their mission. "Make as much profits as possible!"

If you want to blame anybody, blame that government and beauracrates who created this market full of frictions.

Tuesday, January 02, 2007

New Year's Resolution

06 is already behind us. It's now time to put up some New Year's Resolution!

1. Read 12 non-technical books.
2. Run 4 miles x 4 times a week.
3. Workout + gain 10 more pounds (I mean, 9 pounds of muscle + 1 pound of fat!)
4. Contribute and participate in at least one open source project.
5. Start working on our project. Yes, we'll make it!
6. Find a girlfriend, please.

Seems like a stretched goal, but I will make it.