Well before this version of subnet142.com, I've written a lot in form of e-mails, newsgroup posting, articles everywhere etc. Actually, part of the reason of setting up this subnet142.com is to cut down my (perhaps unwelcome) mails to my buddies before they start to block me as spammer... Anyway, some of the writings are still an interesting read after all these years. Have fun!
(From The Archive, originally posted 2005 May, somewhere else on Earth.)
I was on a leisure trip to Taiwan last week. It was fun, and most importantly, I seemed to feel life again. Well, it wasn't due to Taiwan, but perhaps after my 2 years ordeal of midlife crisis, my hormone levels were finally back to normal (or finally my brain had accepted my degraded body). Hope that I can write regularly from this point onwards. It took me a while to figure out that I have either to write programs, or anything except corporate memos, in order to keep my brain functioning.
For the young programmers out there, my advice: if you are really good at your hack, then keep on programming. Unlike the false perception imposed by your managers, who were probably lousy programmers from the beginning, there is no age limit. The perception that you couldn't deal with the influx of new technologies all thru your life, and at some point, you have to back off to the management seat and start to do "strategic" planning, is deceptive. The truth is: there are not that many new technologies out there. For the past 20 years, relational DB is relational DB and OO is OO. New languages are only old wine in new bottles. If you are really good at it, you should know it and absorb the "new" technologies by just re-mapping from your foundation knowledge(1). If you don't feel the same, then you are not good at it - be a manager in this case and enjoy the self deception of "strategic" planning(2).
Anyway, perhaps I should focus back to the title and talk more about the Dilbert stories later on.
So, there was a strange debate when I was in Taiwan. Some think that Taiwan now lacks behind in the global competition, partly in terms of international communication (a.k.a. English), and partly in service level. Service level here refers to the general consumer business such as restaurants or retail shops. There is also a feeling that the economy is not as vibrant as in the past.
While I have no doubt that those who said this didn't exaggerate and really learnt the hard facts from their daily lives in Taiwan, it doesn't reconcile with my perception since 3 years ago that Taiwan conquered almost every sector in the PC industry, e.g. kind of a 70% market share of PC motherboards, globally. Even this alone could make Taiwan self sufficient for being a global leader, not to say the other industries which drive the rest of the economy. So, am I outdated such that it's China now who enjoys the 70% market share, or simply the technology industry isn't as huge as I thought in the total economic picture of Taiwan?
Fortunately, I didn't need to delve into some lengthy research for an answer. So happened that it was also a topic on Business Week (news link), with the e-mail newsletter arriving right after my return of the trip, free of charge. (Ok, ok, the corporate influence of Internet is not all that bad - some goodies do happen from time to time...)
So, Taiwan is still leading (70% of chip foundry service, 72% of notebook PC, 68% in LCD monitors, 33% in servers, 34% in digital still cameras, 79% in PDAs, in the global market shares.). This is spectacular. So, you thought that the prominent iPod and Mac Mini are from Apple? They're manufactured by Asustek. Dell and HP make their own laptops? They're by Quanta (hey, Quanta's founder was born and raised in Hong Kong, but moved to Taiwan to attend National Taiwan University). Who supplies chips for the world? It's TSM, the biggest chip foundry on the planet. And Sony's Play Station, LCD panels, etc. from all the others, all by Taiwan. All together, these high tech companies tally to USD122 billion in revenues. With cross reference to the CIA World Factbook, this is about 21% of the country's GDP.
So, where has the chain effect gone? Supposedly, even though this is not 100% of the country's GDP, it should generate a chain effect which benefits the downstream industries. At least, those techies have to go out and eat somewhere in Taiwan, and benefit the restaurants, after making all those money from the world. But where are they?
Well, perhaps they are all made "by" Taiwan, but "in" Taiwanese-owned factories in China. "As many as 1 million Taiwanese live and work on the mainland.". This is about 5% of Taiwan's population. It perhaps explains why China is so tolerant to the naughty behavior of the current Taiwanese government. Other than political pressure from U.S. et al, a large chunk of China's high tech export is blessed by the Taiwanese.
On the other hand, while Taiwan does enjoy some back-feeding to the downstream as China has now overtaken the US to become the largest export market of Taiwan, the general public cannot feel the full benefit yet. Note that what the general public perceives is not from economic figures but what they see on the streets daily. A more prominent "economic" impact would be say, a prosperous local retail market, which might only account for 2% of the GDP, but 10% of residents whose lives depend on it, and then a 2nd tier of 10% who benefit from the chain such as the entertainment industry, which now have support from the 1st tier who's got the money. It doesn't count in terms of economy, but it counts in terms of votes. We (HK) had that experience - mainland easing travel restriction to HK - our local market boomed immediately and everyone all of a sudden started to see the sun again, even though, yes, perhaps 2% gain on GDP.
So, that's why the biggest gift from the PRC government to Taiwan will be to relax the travel and export restriction between the strait, because it will put the chain effect back into its full circulation. It will also be the most tempting bait that PRC will put in front of Taiwan President Chen Shui-bian to change gear - risk to lose his original supporters on one hand, but possibly gaining even more votes from the other end.
So, where will HK be? Never mind, the re-export due to cross-strait restriction accounts for USD17 billion, i.e. 7% of our total re-export size of USD240 billion last year. USD17 billion seems to be a non-trivial figure but note that re-export doesn't represent our income, it's only perhaps 10% from the USD17 billion of logistic handling which will be our income on the transactions. So, USD1.7 billion, 0.7% of HK's GDP, I guess we can suck it up considering the best interest of the region for a more stabilized cross-strait relationship, with the ever more interwoven economic ties between the 2 parties.
Some years ago, I was surprised to learn that most SQL programmers didn't even know set theory, and yet they were supposed to handle a hybrid of relational algebra and calculus in the form of SQL proficiently. Set theory was one of the subjects at high school in the '70s. So, nowadays, learning a programming language means learning its syntax as interpreted by most. This is like saying that you've learnt English literature because you've learnt the grammar. And not to our surprise, most did interpret literature as a language set of grand wordings and difficult adjectives, rather than the content and impact of a writing...
(From The Archive, e-mail in 2005 Feb. Re-formatted for re-post.)
This is perhaps the longest software project that you could possibly have heard of - 15 years of development since inception. Microkernel was hype in those days, but now on every Windows OS already (Microkernel since NT 3.51). Anyway, glad that they have finally completed the first console application on HURD, just 10 years off schedule... BTW, Linux (except the L4Linux as mentioned in the Slashdot news) is still the older monolithic kernel design. No matter how much you hate Microsoft, they have their technical edges that shouldn't be misunderstood to be just marketing tactics. Don't degrade yourself to think like a marketing executive - Microsoft succeeds because they know how to treasure their technical asset, and not just marketing strategies.
If talking about visionary (Microsoft is not though - they are pragmatic), that's what before NT of an OS with many advanced features - the NeXTSTEP. A Steve Job's demo archive, back in 1992... (Perhaps I shouldn't have forgotten Amiga, the first commercial true multi-tasking, and a fast one, PC O/S, when Windows was still 1.0 and behaved like an idiot. Mac looked better at that time, but still not pre-emptive multi-tasking.)
It took 10 years for Steve to really bring NeXTSTEP to the market, as Mac OS X. The price of being visionary - the market just wasn't ready for it until 10 years later. (Although the technology was ready, the machine was pricey and beyond consumers' acceptance.) Well, at least it's better than most mathematicians - 200 years for Fourier transform to be used in audio processing, 30 years for fractal geometry to be used in video compression... For those who think that social investment on technology means eBay or PCCW, think again - it's about stuff that you probably cannot see any return before you die, and that's what we are certainly lack of in the Eastern culture.
BTW, for nostalgic reasons, you might be interested to learn how those folks compete in the 80's (Win 1.0 vs. Mac <-21MB download!!):
We must be too conservative and thus cannot escape the corporate jailing...
(From The Archive, e-mail in 2005 Aug. Re-formatted for re-post.)
NY Times News Link (free membership logon required)
Perhaps my own mourning only. Moog was the father of (analog) synthesizer. I started with digital sampling synthesizers in 1994 so it's well before my time. A bit before that, in 1985, I started with FM (Frequency Modulation) synthesizer which was a concept evolved from Moog's pure sine/triangle waves addition/subtraction method. FM was also a brilliant idea because it could produce waves that were so complex that you would literally need a few hundred oscillators to produce with the Moog's method. However, this complex wave forms were also its drawbacks - it was just too complicated to master and therefore creativity remained in the hands of a few top sound designers only. So, most musicians would only use the factory presets bundled with the Yamaha DX syns, and yes, DX was once so hot that every band would have one.
The Moog synthesizer was something in the 60/70's, but still highly regarded for its original sound which no sampling device can match. For one thing, this is true creativity because we make up waves from the fundamentals, whereas for the modern sampling synthesizers, we copy sounds from real instruments or environments, or well, sometimes from the Moogs, but not really a creation. Remember Ryuichi Sakamoto? His debut surprise to the music industry was not his musical talents (well, which is pretty average - I would say), but his unworldly sounds with absolute perfection created from his array of Moogs.
So, analog still survives. One of the recent Moog imitators that I know of is Nord Lead, which got famous because of alternative rocker Trent Reznor. BTW, Reznor's debut album "Nine Inch Nail" was such a shake-up to the boring rock scene in the 90's. Well, Reznor is old by now as well...
Anyway, it's part of my memory.
(From The Archive, e-mail in 2005 Feb. Re-arranged for witness protection.)
With the way that most corporate programmers write sqls, this server will become a necessity very soon for running one sql, minimal configuration 100GB memory (the sql wont' run unless the whole database is cached).
It has to be accompanied also by 10Gbps custom built fiber links without repeaters in between servers across ocean (probably need to join hands with Xerox and AT&T since such technology hasn't yet existed). This is to solve the performance problems associated with a typical EJB on RDBMS design by using COBOL style sequential programming in favor of set operations, on a set machine (this interesting mismatch puzzled me for a long time but now I finally understand why - Sun is a hardware vendor - like giving out free shavers for razors business in the long term).
Well, the funny thing is - such configuration might even be cheaper than to train a typical CIO/CTO to understand what system building is all about...