The Fat lady has been singing for years...

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

In his article "Y2K, I Know what I Know", Yourdon makes this statement: "And what I know about large projects in large companies is that a substantial percentage of them are finished late, and/or over budget, and/or riddled with bugs." He then goes on to make the usual "doomer" argument that this is one reason that Y2K bugs won't be fixed on time. In a recent article here, Dale Way essentially says what I've heard from EE guys for years. Notice how as he goes down the list of possible problems, physical systems are mostly OK (that's the EE realm), but the more things rely on software (support systems, the realm of Software Engineers, if you can call them Engineers) the more problematic they become (buggier essentially). I'm a software guy, and I've heard this talk from EE guys for many years now. And the truth is both Yourdon and Way are right. Most software projects, and software systems are garbage. The thing is, they have been for years. This is one of the major reasons Y2K won't be that big a problem from this point on.

Why?

Because companies, governments and people just like you have learned to live with it.

Take away Y2K and these systems will still fail on a regular basis. They crash, they report incorrect data, they crash, they transfer wrong amounts of money from one account to another, they cancel accounts incorrectly, they crash, etc. etc. etc. This has been happening for years. Its just not advertised like Y2K has been. The primary concern for Y2K is that failures would be so widescale all at one time that it would make it difficult to fix in the usual amount of time. But fortunately the problems are not at that scale, and from this point on it is _highly_ unlikely that they will be. Y2K bugs are now just one more set of bugs to add to all the other failures that are going to happen, and have been happening all along. In fact this has _already_ been the case for over a year now. There have been far more Y2K failures in the last year than have been reported in the media or wherever. Just this last week I ran into 2 YK2 failures on my own home system that must have hit others, but were never reported. But at the same time, I added a new hard disk to my system (win98 with all known y2k upgrades) and the system started freezing solid several times a day. This wasn't Y2K related and was much more severe than the Y2K problems I did see. I also played several computer games which crashed the system several times last week (driver and bios problems). I had several other programs cause GP faults. Win98/95 (and NT) might be the buggiest piece of garbage on the earth, but its also one of the most most used pieces of garbage on earth, by many major corporations, governments and people. Its been that bad for years, but it hasn't ruined the economy and it hasn't brought any governments down (yet).

So while hundreds or possibly thousands of Y2K failures have occured in the last year, and more will happen this year, many thousands of other failures have occured worldwide at the same time, and will continue. Non have yet caused the end of the world.

A server product I'm famiar with has a mean time between critical failures of about 3 days. It's used over the internet by millions of people everyday. Most never know it goes down. Since it's known that it will fail, more than one server is run at a time, and its setup to restart quickly and automatically, and to send out alarms in various ways when it does fail. Bugs are fixed in it almost daily. And believe me, this is not the only system like this.

So the real problem with doomers is not that they are too pessimistic, but that they aren't pessimistic enough, at least about software systems. But software systems have been this bad for so long that failure tolerances are big enough now to expect and handle most such problems. This is not a world where all the systems run near perfection and any jolt or failure will send it into chaos. Rather, its a world built upon systems that fail regularly. So just like there are plenty of car repair shops out there to fix your near perfect car when it breaks down unexpectedly, there are plenty of Software Engineers and IT guys around to fix systems problems as they crop up, because we all _know_ its going the fail. Its called job security.



-- Fred Bryce (fredbryce@hotmail.com), January 03, 2000

Answers

Hey Fred, I like your name..

You are completely correct. I'm an SE as well. 'Stuff breaks all the time, (nothing I write of course..). That's why they pay us to be here.

I think software is akin to a chain mail. As I write new systems, people buy them. After a while, I move onto another project, group or company, then someone has to maintain what I wrote.

I think I've created about 12 IT jobs over the years..

We are good at fixing things, and yes we found out the hard way that there is not 'general' Y2K bug that kills everything, or causes cascade failure.

That means Y2K bugs will be fixed as they come up, just like any other bug.

Bryce

-- Bryce (bryce@seanet.com), January 03, 2000.


Fred, Well said. I said that all large projects come in when they are due because the due date has been reached a couple of months ago. When the projects slip it is because the higher level management is willing to continue throwing good money after bad. With Microtrash, they realize the customer will work around their junk so after only 40000 bugs are left of the original 100000, they release.

I don't think there were that many millenium bugs that were serious in the first place. I would like to see what kind of problems were remediated by the y2k teams.

As for the systems where a totally new implementation was the answer as in ERP systems where SAP was purchased, this was great business. I am sure there were lot's of kickbacks for America's CIOs as they engaged in an orgy of wasteful spending on new systems and software.

I believe that the companies that did this will be cash starved for the next few years..

-- William R. Sullivan (wrs@wham.com), January 03, 2000.


Fred,

Supposedly 6,000 gov't systems remediated out of 70,000, 10-20% of the overall code. What function do the non-critical systems play in the government? This question is never answered anywhere. And if it has who has answered it? Basically we're had triage from the start of this just to make it. Maybe. What were some of those systems? Non- critical mission to the government triage people, but maybe critical to the people that derive benefit from them. Extrapolate this worldwide. Tell me what the effects of this will be? I'm guessing that it is impossible to know.

-- PJC (paulchri@msn.com), January 03, 2000.


Well I knew the software side was pretty bad and that is the side that I have been mainly looking at. How high the failure tolerance can go I don't know, it's violatile enough as it is!. It depends if the fault tolerances will be small bumps in the road, or the road gets so bumpy that the nuts get loosened and the wheels starts slipping off the cart and the programmers will be busy trying to hold onto the wheels to stop them coming off, apply band-aids or anything, patch it with tape, or something that will hold it in place and at the same time they are doing the repairs, they are being carried up and then over and then back down as the wheels turn over in the economy carrying us on top of it, programmers are getting a rough ride themselves, right underneath everything else. In other words you need the programmers to keep everything to stay together

-- Brent Nichols (b-nichol@ihug.co.nz), January 03, 2000.

PJC, you are right, the question is not answerable. However I do believe that I can say with complete confidence that even though 80% of governments systems may still be untested for Y2K bugs, 100% of government systems have non-Y2K bugs in them. Should this scare us? Perhaps. But I'm not going to live my life in constant fear. I'll never fault anyone for living life "prepared". But nothing you could ever do will %100 guarantee you will still be alive tomorrow. Survivalist or not. And no software engineer can ever guarantee that there are absolutely no bugs in their program. Well, maybe a "hello world" program, but even that usually goes through library startup and exit and uses a print function which may have bugs in it.

-- Fred Bryce (fredbryce@hotmail.com), January 03, 2000.


Seems like some sort of "atta boy" "we're great, the glue that holds everything together" thing going on here. I'm outta here, hope you folks are right.

-- Michael (michaelteever@buffalo.com), January 03, 2000.

Michael, actually its quite the opposite. Software Engineers don't hold anything together. Thats the whole point. Its the "doomers" who believe the whole worlds going to fall apart because of a software glitch, implying that everything _is_ dependent (and thus held together, glued together) by software/computers. My point is that software bugs have been around for a long time, and society continues inspite of them. Software and thus Software Engineers and the IT groups are _NOT_ the glue that holds society or the economy together. That is one main reason I don't believe Y2K was or is a TEOTWAWKI threat. It is the "doomers" who inflate the importance of software systems, and indiretly the ones who support those systems.

Not that a massive failure all at once on Jan 01 would not have been a major pain. But I don't believe we were ever in danger of TEOTWAWKI, at least not any more that we are from non-Y2K bugs the rest of this year or the next or the next...

-- Fred Bryce (fredbryce@hotmail.com), January 04, 2000.


Its nice to read an intelligent post by someone else who has actually worked in the software industry.

As actual programmer who actually has had to fix Y2K problems. I too had to make some last second (unforseen) Y2K fixes, but they have taken the grand total of about 24 hours work.

All the other bugs ... call me in 2999! :)

-- N. Zax (smalltalkin@yahoo.com), January 04, 2000.


Moderation questions? read the FAQ