One thing to remember with this whole Y2K retrospective is that this was only a problem because of the endlessly shoddy code behind Windows and DOS. Yes even windows 98, published just two years before Y2K, was not fully Y2K compliant.

Alternately even System 7 for Mac (first published in 91) was fully Y2K compliant. It is quite likely that earlier systems of the Mac OS were compliant too there was just no testing done on system 6 and earlier (since by that point virtually noone was using it given how vastly superior 7 was). Yes Apple solved Y2K decades before the monolithic windows machine did and with absolutely no expanded IT, no recoding, no billion dollar expenses and no headaches.

This is, of course, the difference between Apple and Microsoft summed up well. Apple builds first, builds correctly and works while Microsoft makes bad decisions (IE pinning Windows to DOS, 2 digit years, Internet Explorer vulnerabilities) and then keeps going down that mistaken path for years if not decades. Its virtually impossible MS did not know about Y2K by 1998 and yet, they put out Windows 98 without compliance anyway shifting the cost to Windows users.

In 1999 it wasn't even clear whether there would be a patch from Microsoft in time for windows 95, again, they can't think 5 years into the future to see that going from 99 to 00 might be a problem?

This argument is best summed up in the classic 1999 Apple superbowl ad. I guess IT workers are lucky Windows remains so prevalent. If everyone switched to macs Indian outsourcing would be the least of their problems. How do you find jobs for millions of IT problem solvers when the problem set is reduced by a factor of 20?

link to the superbowl ad here <link>

And Microsoft was stupid enough to wear its Y2K noncompliance on its sleeve: Windows "95," Windows "98." How did they not see that one coming?

Patching Windows was relatively easy. The biggest alleged problem at Y2K was the older systems from the 60s and 70s, which were so well-designed that they were still in use decades later.

Tom_Tildrum:

Patching Windows was relatively easy. The biggest alleged problem at Y2K was the older systems from the 60s and 70s, which were so well-designed that they were still in use decades later.

Yes, the big problem was programs written in COBOL for mainframes. For a few years, COBOL programmers who had been considered dinosaurs were suddenly in big demand as you can see in this Dilbert cartoon

It was APPLICATIONS that were the problem, not the OS itself.

As for the comment of the "endlessly shoddy code behind Windows"... please. The memory manager on Mac System (1-9) NEVER WORKED RIGHT. You could never fully rely on VM to properly deal with a multi-tasking environment. Windows from 2.0 onwards never had this problem.

As for MS-DOS, please. 4 generations of technology had come and gone in the interim.

Nor was Windows "pinned to DOS". Instead it was user demand that existing DOS Applications not be obsolesced by any changes that made this happen.

This argument is best summed up in the classic 1999 Apple superbowl ad. I guess IT workers are lucky Windows remains so prevalent. If everyone switched to macs Indian outsourcing would be the least of their problems

About as ignorant a comment as I've seen in a tech discussion in a long long time

The "endlessly shoddy" remark, while throw-away is important. Software, unlike just about everything else that gets made, does not have the legal threat of "Product Liability" for defective design.

For the record: Product Liability and in particular, the strict liability variant, is a legal concept to express an idea that the societal cost of harm caused by "endless shoddiness" should be bourne by the manufacturer and all the downstream sellers, and that the way to inspire this is to make the bar so low for plaintiffs that all they have to prove is harm, not negligence.

Software, as I said, does not have this idea. Nearly all software is purchased on the basis of a revocable user license where the user has no remedy of refund for defect and no remedy for harms resulting from use. Caveat Emptor.

Despite some comments that Manooj didn't research this article, its not a myth that lots of companies spent lots of consulting dollars inspecting each technology possessed in search of this one defect. Companies paid attention both to their cost estimates as to what the defect would cost to: detect, fix, and not fix.

Sadly, Y2K did not result in a major disaster causing a massive cost to society. If it had, then software makers would have been required to write software that actually works and is not buggy. Practices like weekly patches distrubuted through the internet would have been replaced by crippling lawsuits and bankruptcies forcing software makes to not release software unless it was bug free to begin with. That possibly would have stopped offshoring, it certainly would have prevented Windows Vista.

More importantly, it would have created a revolution computer science. Programing, unlike legitimate forms of engineering like civil, chemical, electrical, mechanical, etc. suffers from a very low level of discipline such that ludicrious ideas are given legitimacy like Open Source "if we let anybody write whatever code they want, good code will eventually emerge, because what members of the public really want to do is scrutinize details to get it just right" and basic things like scheduling, revision management, communicating in teams, and quality inspections remain very very very elusive despite decades of industry experience.

Cranky, I'm going to have to disagree with you here. Part of the reason for "shoddyness" in programming isn't due to lack of effort, it is due to the inherent complexity of large software projects. Don't let the concept of "its just software" fool you. Large software projects can be infinitely more complex than many other engineering tasks, simply because they are far more inputs and more variables involved. And programming is just as legitimate a form of engineering as any other that you mention. I know several civil engineers and many programming jobs are much more difficult and intellectually demanding than many civil engineer jobs (not to mention the fact that most electrical engineers such as myself end up working as programmers anyways).

There is also the issue that software has no "wear" problem. I wrote some code for MSDOS back in 1984, That code is everybit as functional today (assuming you ran it in the right execution context) as it was back then. OTOH, the car I had in 1984, got a new engine installed in 1986 and then was sold to the scrap heap in 1994 after it "crashed".

Because code 'doesn't wear out' there are companies that are STILL running code that was written in the 1950s (some of the Bill of Materials Mgmt software for NASA). No other engineering discipline has that sort of design legacy trailing behind it. NONE.

Regarding software, you've also got the complexity that the programmer cannot possibly test the software under all real-world conditions. Not only do you have a huge number of hardware possibilities, but computers can run a combination of software programs reaching infinity. You simply cannot test your software with the combination of other programs with which it will be running in the real world. Therefore, you cannot guarantee that a particular program will run properly on a particular computer.

I ran into this problem in the mid-90's when I was writing mass-market educational software. It turned out that there was a certain screen-saver which was incompatible with one of my programs. If the program was running when the screen-saver came up, the program crashed. This was not something I could have foreseen, because I couldn't test my program with every screensaver, and even if I could, other screensavers would be released after my program and I couldn't keep up with them all.

I imagine there is some funny story to why you say your car "crashed" in quotes. :)

It was not just applications, it was also stupid design decisions made by the Windows team. An example is that the system defaults to a two digit year. This was a 1995 OS, so did noone on the design team think "hey this will be really inconvenient in 5 years"?

There are countless stupid and careless mistakes like this in Microsoft software. Another great example is code for the Zune which does not compute Leap Years correctly causing the Zune to break every 4 years. Connecting the endlessly hackable internet explorer right into all the core features of windows is another decision that has caused endless billions of dollars in headaches from all the worms, viruses and more that have come down that particular lane.

Microsoft codes poorly and doesn't think through important decisions correctly, it has been true for decades and remains true today. Yes, there are a handful of Apple mistakes especially in middle the Steve Jobsless years but overall there are very very few incidences of million or billion dollar break downs from shoddy design. Put it out broken, we'll fix it later is certainly the mantra at one of the two big OS companies, and its certainly not Apple.

As to being pinned to DOS it wasn't just a user issue, it was a marketing issue. Originally Windows was heavily pinned to DOS because Microsoft had no from the ground up GUI options available to compete with Apple's Mac (which they shamelessly copied after being given a preproduction model to make Office programs on). They couldn't start from scratch in time so instead they tacked a GUI onto DOS, a decision which persisted for more then a dozen years with version after unreliable version of windows being released until finally they picked up the core from the NT series.

As to MS DOS being gone, many MANY windows 95 applications used DOS heavily and Windows 95 was still a very prevalent OS in 1999. DOS was integrated tightly into Windows through and through until the switch to NT, thus yes maybe DOS as the primary OS was gone, but it was still part of almost everything Windows did for a decade after it stopped being a stand alone solution.

As to the memory issues with OS 7-X (the OSes available in this 90-99 window) A) that is a separate issue not related to stupid design decisions like a 2 digit year or making a giant security hole with IE and B) it worked fine especially if you bought systems with sufficient RAM and its not like RAM usage in Windows has ever been a strong suit for the platform. I've used macs for more then 2 decades now and I never found myself unable to use my computer due to memory issues.

Finally I stand behind the IT comment. I work at a small business that is half mac, half PC. literally 95% of the tech support issues are with the PCs, its not even close.

I've had to deal with viruses, Vista Home Premiums complete inability to do any kind of VPN (good thing its the standard OS on almost all new computers), malware, networking problems (why Samba is so hard for Windows machines to handle correctly is beyond me) and much much more. If macs were 90% of computers instead of the other way around I have absolutely no doubt that the demand for IT support would halve at the very least, I would guess even more of a drop then that.

I put a new engine into it, gave it to my sister. She eventually rolled it. So it was a "operator error" rather than a system crash.