About This Web Site

This web site was started in the early noughties. It was a part-time venture, built in a garage that had been converted into an office, using some desktop towers that had been transformed into BSD and Linux servers. One of the most valuable resources that assisted in the construction of the site was the web itself, which was, and still is, rich in information about how to build a web site.

The Early Years -- What My Kids Refer To As The "Jurassic Period"

I had written my first computer programs in the late sixties. At the time, computers were massive installations that occupied many thousands of square metres of floor space, filled with bulky, heavy cabinets containing expensive circuitry, motors and cabling. They consumed a considerable amount of electricity and, as a result, required large and powerful cooling systems that consumed even more electricity and, for the largest complexes, quite a lot of water. The overall installation could have an energy budget that rivalled that of an entire village.

Later, in the mid-seventies, the observation made by Gordon Moore -- that the number of components that could be placed on an integrated circuit (IC) would continue to increase exponentially -- became popularised as the eponymous Moore's Law. Like so many of the so-called laws of economics, which, rather than being laws, were observations about data, often accompanied by a graphical representation involving only two variables, they were really observations about contemporary trends, behavioural norms and agreements rather than actual laws, such as the laws of thermodynamics ... Especially the 2nd Law ... [Now that's a law -- to paraphrase Crocodile Dundee]. Nevertheless, for the last third of the 20th century, Moore's Observation appeared to hold firm and was often quoted as a bona fide law.

As such, Moore's Law seemed to apply to research and development during the seventies. Logic circuits became smaller and, before long, we had an entire Central Processing Unit (CPU) on a single chip. This led to the microcomputer. And there were knock-on effects. Some programmers at AT&T Bell Labs developed an operating system inspired, in part, by a system named Multiplexed Information and Computing Service (MULTICS). They toyed with naming their new system Uniplexed Information and Computing Service (UNICS), but eventually settled on a rather nerdy play on words: UNIX.

In California, two computer enthusiasts started a company and designed and assembled a microcomputer kit that they named "Apple". In New Mexico, two young programmers founded a small company to develop software for the new microcomputer revolution. Their first major product was a version of BASIC for the Altair 8800. They called the company Micro-Soft [Software designed for microcomputers].

I had heard about some of these developments but was slow to adopt them. In the eighties, PCs began to proliferate in the workplace. There were increasing opportunities to work with mid-range and small computers. It soon became apparent that the combined amount of information stored on multiple users' desktop PCs was larger than that stored on the "real" computer in the data centre. And while the information on the desktop was growing, the size of the "real" computers was shrinking, whilst their capacity was growing and their power consumption was falling. I began to acquaint myself with Unix and C. And in the last decade of the 20th century, I transitioned to programming for PCs and mid-range systems rather than for mainframes.

Boom, Boom ... Or Bust?

The nineties were a brave new age of techno-optimism. When I wrote my first program in 1969, the approximate cost of a megabyte of memory had been measured in millions of dollars. Later, in the early nineties, the cost had fallen to tens of dollars per megabyte. That is an extremely steep decrease, but perfectly in step with Moore's Law. And rather than taking up half an acre of real estate and consuming as much power as a regional town, the computer sat on my desk and used less power than a toaster.

But the real kicker in the nineties was connectivity. A packet-switching network called ARPANET had adopted the TCP/IP protocol suite in the eighties, helping to create the modern Internet. Then Tim Berners-Lee proposed a system that came to be known as the World Wide Web. The name, although immodest, turned out to be accurate. It was a virtual web that really did wrap around the globe, connecting us all and super-charging the technological zeitgeist. And if Moore's Law really was a law, the price per MB of memory was going to continue decreasing every month, at an increasing rate!

It seemed that we were on the verge of a new age of communications. Data, news and knowledge could be shared easily. Publishing costs had dwindled almost to zero. Those of us who embraced the optimism opined that this would lead to an incredible ocean of information liberated from onerous servitude to a few oligarchs who controlled and distributed it via radio and television networks and massive printing presses.

To a certain extent, the optimists were correct. The World Wide Web did indeed inundate the Internet with data. However, not all of it was quality news, technical specifications, history, interesting and factual details of our real lives, or other useful information. Some of it was bullshit. This was only to be expected. After all, the content on the web was being created by humans. And humans, as has been well documented in so many historical, scientific, economic, sociological and religious contexts, do have flaws, not the least of which is a penchant for bullshit. So, surely we could tolerate a little pinch of bullshit in our brave new information techno-sandwich? It couldn't be much worse than the bullshit produced during the past six centuries of civilisation, could it?

But there was a more important and substantial ingredient in this global Internet pot-pourri ... And that was Money! There was obviously going to be a lot of it created and burned by this transformative technology. And in the nineties, any company with ".com" in its prospectus could benefit from a surge in its stock price. When the dot-com bubble began to crank up towards the end of the decade, some of the techno-optimists speculated that, since the Internet space was virtual, there were no limits to growth, especially since the US dollar was now untethered from the gold standard and the chairman of the US Federal Reserve was quite sanguine about opening the spigots and making credit cheap if the economy required topping up.

However, the no physical limits paradigm was just more bullshit. There are physical limits to growth for everything on Earth, even the new digital virtual universe. Nevertheless, stock-market values continued to surge and price-to-earnings ratios went into double digits. By the turn of the century, P/E ratios bordered on human mid-life-crisis levels.

As they say, the rest is history. At the turn of the century, gravity reasserted itself, dragging financial valuations down to a hard, realistic ground zero. At first, the most dramatic correction was in the stock-market indices, especially the tech-heavy Nasdaq index. Smaller or unprofitable corporations were wiped out. In the USA, the Federal Reserve responded, as ever, by cutting interest rates. Regulatory bodies became more strict about enforcing accounting standards, and the rules on mergers were relaxed to allow some of the casualties of the market pile-up to be consumed by still-viable entities rather than disintegrate into detritus littering the kerbs of the financial outflows from Wall Street.

The reduction of interest rates is a remedial action often undertaken by central banks to address drastic market corrections. It stimulates the economy, but can run the risk of fuelling inflation and over-stimulating the growth of asset values. At the time of the dot-com bubble and its aftermath, many central banks considered economic growth to be more important than potential risks. Some critics have argued that this relaxed attitude to monetary policy was an important contributor to the formation of the dot-com bubble in the first place.

Whatever the risks and/or benefits of this approach, the dot-com bubble had already burst by the time another severe external shock arrived. The September 11 terrorist attacks delivered a further blow to an already weakened economy. Commercial air traffic was paused. Wall Street closed for four days, the longest closure since the Great Depression. When it reopened, the market suffered a historically severe decline. The Nasdaq continued to fall for some time, and there was a further downturn in the technology and IT sectors.

In response to the broader economic weakness, the US Federal Reserve, and then, following its lead, central banks around the globe, cut interest rates and made credit substantially cheaper. The resulting flow of liquidity would eventually contribute to the enormous financial imbalances that developed during the following years, helping to create the conditions for the Global Financial Crisis that would arise seven years later.

My Little Corner of the Web

At the time, I was unemployed. Normally, this would not have been a matter of great concern. There had been sustained growth in the IT industry during the previous thirty years. Labour in the sector was mobile and could shift rapidly to the highest bidder. After the 9/11 attacks, the labour market tightened. Capital expenditure and business spending disappeared. In Australia, firms were not firing, but they were not hiring.

Fortunately, my spouse was able to assume the role of primary breadwinner, and I concentrated on domestic duties and caring for young children. In my spare time, I worked on creating a domain and a web site. Even though most of my earliest experience with computer technology had been with mainframe computers, I felt that I had learned enough about computing and information technology to be able to put together a web site, perhaps promote myself, and try to get back into the workforce.

A few weeks later, I was the proud owner of a bright, shiny new web site. I sat back and waited for all of those visitors who would pull over from the information highway and visit my humble site. ... But nobody stepped through my digital doorstep! ...

At the time, people had learned to search the web by typing a query into one of several search engines, all of which offered their services for free because the "clicks" were a valuable resource. After a decade of fierce competition between search engines, one of them had come to dominate the industry. Such was the success of this contender that its name became a verb. A verb that I was quite familiar with.

After all, I had Googled for most of the information I required to build the MTA host, DNS server and web host for the domain. I had also notified Google that my site was ready. And I could see in the log files that Googlebot had visited my humble little site.

The problem was that if anyone entered a query into the Google search engine, my site was not returned anywhere in the first several hundred pages of matches to the query.

A little research revealed that a web site's ranking in Google depended, among other things, on links from other sites. If a site seemed to be popular, it would be given a higher ranking. There seemed to be some other factors that influenced the ranking. The algorithm may have performed some analysis of the text and content on the site and the structure of the code, and assessed whether it complied with commonly accepted standards of coding.

It seemed that the best way to improve the ranking of my site was to create some content and contact other site owners and exchange links with them. I continued to devote my time to household duties and, in my spare time, worked at improving my ranking in Google searches by writing articles, doing research on matters related to the Internet and starting a blog.

Eventually, this toil bore fruit. By 2002, if anyone entered "DBA Melbourne", "Systems Administrator Australia" or a related search, my site would appear in the top five lines.

And then I started getting calls from potential employers. So I was back in the workforce. I gathered clients in Melbourne, a few at first, then started getting contacted from overseas and Sydney.

I tried to keep up with my web site, but gradually it started to degrade due to neglect. There were several scripts running on it, keeping track of browser agents, and some people would still visit to read about agent strings. However, without regular maintenance, the information became less relevant. Links on the site became stale as other web sites were upgraded or moved to different domains.

Plus ça change, plus c'est la même chose

Almost a quarter of a century later, I have retired. I am still doing a little part-time work; however, there have been some dramatic changes in the tech sector ... But, as the old French proverb goes, about the more things change ... The more intense the déjà vu ...

If the francophones will pardon the pun ... Now, in the second quarter of the 21st century, there is significant momentum towards AI. And lately, there have been growing calls for the construction of massive data centres that will occupy large areas of land and consume large amounts of electricity and water to feed the AI beast.

Furthermore, this build-out requires enormous amounts of capital expenditure. And the resulting rapid growth is leading to speculative investment that bears many similarities to the dot-com bubble at the turn of the century ... Hang on! ... No wonder I'm experiencing déjà vu ... I've seen this movie before!

This almost sounds like the early years. Except that back then, massive data centres came first. And then we got speculation, unrealistic estimates and an investment bubble. Good thing we don't have a secondary external shock? ... Unless, of course ... Oh! Wait! ... Don't mention the war! ... Or the Strait of Hormuz!

A Brave New World --- Or is it 1984?

I have now started to renovate this site. I am cleaning out cruft, broken and redundant links, old style sheets, deprecated code, etc. I've also re-read the scripts that I wrote originally. And I will endeavour to enlist AI to assist with the project.

So far, the code our friendly neighbourhood AI agents produce looks good, and the AI sector does appear to be ascendant. However, the business proposition appears suspect. Especially now that it seems that Moore's Law really was just Moore's Observation.

In the second decade of this century, the remarkable miniaturisation of digital circuits began to slow and became dramatically more expensive. With the added pressure of AI firms buying up chips and memory, prices for some types of memory have begun to rise sharply again.

And there may be speed bumps ahead for traditional software. The recent release of Microsoft's OS, Windows 11, can run more slowly than previous releases on some hardware. In the past, software companies could often rely on the ever-increasing performance of computer hardware to make their latest release look good, even if it was a little more bloated, as it often was.

Now that hardware improvements are becoming more difficult and expensive, and the industry is increasingly focused on integrating AI into operating systems, Microsoft may not be able to demonstrate the same effortless improvements in performance that were once possible. The old assumption that each new generation of software could simply demand more resources because the next generation of hardware would provide them may no longer be reliable.

In addition to this, as AI systems siphon off more of the data from web sites, the legacy link-based architecture of the old algorithm may be degraded as sites receive less traffic from webizens. This, combined with what appears to be an over-optimistic assessment of the economic and ecological costs of data centres, poses a threat to the AI business model.

I'll try to write about all of this while renovating the site and, like the rest of the world, pondering our dilemma: Are our AI overlords helping us build a utopia ... Or will it be a dystopia? ... And if so, will it be 1984, Brave New World or The Matrix meets Skynet?