A few years ago, in both American Programmer and a now defunct publication called Ed Yourdon's Guerilla Programmer, Ed conducted discussions regarding "good enough" software. Basically, the crux of the discussion was that software didn't have to be perfect to be usefull, and at some point enough of the flaws had been removed to make the software "good enough" for most people.

Does this concept apply to Y2K remediation? I believe that it does, but I also tend to believe that there are those among the other frequent posters and readers out there who do not, or would at least differ on where to draw the "good enough" line. How about it. Any one up to hack out a definition of "good enough" Y2K remidiation with me over the next few days?

Paul

-- Paul Neuhardt (neuhardt@compuserve.com), June 01, 1998

Answers

I just want to be the wiseguy and point at Windows as "good enough" crap. heehee.

I hope the fixes are at least good enough. I wish these 577 days would pass by faster. My life is too boring, ya know.

ken (-;

-- Ken Ott (k150@yahoo.com), June 01, 1998.

Paul,

Your point about the possibility of remediation being "good enough" is a good one. The problem with a definition of good enough is that it is application specific; good enough for an electronic funds transfer is different from an elevator controller, which is different from an IRS application, which is different from an air traffic controller program.

Sometimes I wonder if I'm overreacting to Y2K for exactly this reason. Granted that somethings won't be fixed, but will they be "good enough." The problem I see, though, comes from my own experience. Sometimes repairing a software application is like rebuilding a car engine. You pick the whole thing apart, make the changes, then put it back together again. If it's apart when you need the car to drive to the hospital it doesn't matter if you were just three hours away from putting the last piece in place--your car doesn't work and you don't get where you need to go. Some application will work relatively well in 2000, albeit with some remaining bugs. Others will still be picked apart and won't run at all.

Best regards,

David

-- David Palm (djpalm@compuserve.com), June 01, 1998.

Let's triage the problems into distinct categories and take a first stab at a definition of "good enough" for each category. We can always refine the categories and definitions later. (Yes, I know "triage" means to order into three groups, and I am defining four. Poetic license.)

1. Catastrophic: People die, or suffer life-long, severely negative effects. Does the fix for these problems have to be complete, or could it be aceptable in the short term to fix these problems to the point that they then fall into one of the other, less severe categories and complete the fix later, after 1/1/2000?

2. Severe: People suffer negative effects, albeit not life threatening. Still, livelyhoods, savings and possesions may be reduced, damaged or lost. Standards of living are reduced, perhaps considerably. Could a lessening of the effects of these problems also be considered "good enough" in the short term?

3. Problematic: Troublesome, but survivable. Some conviences of modern society may have to be done without for a while, but life rolls on. Cell phones don't work for a few weeks or months, food stocks in supermarkets are reduced by 25% and people have to get by for a while with simpler diets, fuel shortages a-la 1973 emerge, etc. While ot would be desirable to fix these problems ASAP, could we postpone these fixes if efforts could be re-focused on problems in the first two categories?

4. Trivial: Irritating or mildly inconvient at worst. No serious effect on quality of life. How many of these need be fixed at all?

-- Paul Neuhardt (neuhardt@ultranet.com), June 01, 1998.

By the way, that last email address isn't a typo. The old compuserve address will still work for awhile, but anyone sending me email should use the ultranet address from now on.

-- Paul Neuhardt (neuhardt@ultranet.com), June 01, 1998.

Paul,

He is my considered opinion of what is "good enough":

Fix all systems that need to be fixed to assure that we have a continued supply of electricity and food to all parts of the country from now through 2002 at no less than 90% of current levels.

Since much of the generation capability depends upon rail to bring fuel, that implies that the railroads are also fixed.

Since much of the food is transported by road, that implies that road travel, long and short haul trucks, is fixed. This also implies that fuel will be readily available, and that implies that the petro-chemical industry must be fixed.

Simce much of the petro-chemical industry depends upon shipping for its raw materials, that implies that shipping must be fixed.

Since no one or no company is going to do anything for anuyone for nothing, this implies that there must be some system of payments (or some plan for some system of payments) in place. This implies that the financial industry (at least the banks) must be fixed.

If all of these systems are fixed so that they will function at 90% of present levels through 2002, I'd say that is "good enough".

You may notice that I do not consider government to be necessary. If the other things are fixed and government is not, I forsee the church retaking its position from the government as the protector of those unable to care for themselves, a position that the government should never have been in in the first place.

You may also notice that I have not included "health care" as a necessary to be fixed industry. The reason for this is that doctors, left to themselves, much like lawyers, will find a way to ply their trade and garner wealth, and in the process doctors will perhaps help a few persons (lawyers, like bankers, only help themselves).

How's that for a set of biases?

George

-- George valentine (GeorgeValentine@usa.net), June 02, 1998.

George,

The problem I am addressing is the definition of the word "fixed." To restate a little bit, does fixed mean "dead solid perfect" or could it mean "it'll do for now." Obviously perfection should be an ultimate goal, but must it always be achieved?

If you agree to answer that question "no" the next question becomes "What must be done to make it do for now?" Certainnly the answer to that question differs by case, but for this discussion I bet we can come up with a few generic definitions that should be "good enough" for some debate.

BTW, that's quite some set of biases.

-- Paul Neuhardt (neuhardt@ultranet.com), June 02, 1998.

Just an addendum to my previous post: I think that "90%" may be a bit vague in terms of defining "good enough." For instance, if fixing 50% of all instances of Y2K defects in a system move it from the catastrophic category into the trivial category, or even the problematic group, might that not be "good enough?"

What I am hoping to arrive at is something like this, only more detailed and though out:"All catastrophic and severe failures must be remediated down to at least the problematic stage. After that, all problematic issues must be resolved in order of some priority to be assigned later, then trivial problem resolution may be considered."

-- Paul Neuhardt (neuhardt@ultranet.com), June 02, 1998.

Paul has added:

"What I am hoping to arrive at is something like this, only more detailed and though out:"All catastrophic and severe failures must be remediated down to at least the problematic stage. After that, all problematic issues must be resolved in order of some priority to be assigned later, then trivial problem resolution may be considered."

If any co. or corp. has a deadline contract or penalty clause contract, (because any delays will cause problems up stream), or if any vendor has trouble supplying their customers, in fast demand service, (hospitals etc.), then nothing less than the full remediation of all systems that are necessary to fulfill contracts, would be acceptable. Many will default because of the tremendous costs in penalty fees and loss of contracts if they cannot supply in the same fashion as the demand requires.

-- Dave Jones (dfj@fea.net), June 02, 1998.

Paul,

Let me try again. As usual, I added some text that was more noise than light.

He is my considered opinion of what is "good enough":

Fix all systems that need to be fixed to assure that we have a continued supply of electricity and food to all parts of the country from now through 2002 at no less than 90% of current levels.

My requirement really stops here. The other stuff I added was to show what I felt the scope of the problem was/is: HUGE.

I don't care whether every system is perfect or not as long as there is the *reliable* delivery of electricity and food to all parts of the country at levels that are no less than 90% of the current levels. This does not imply that every system has to be fixed so that 9 times out of 10 it works. This means that if 100 million pounds of food are delivered daily (weekly, monthly, whatever is reasonable) now that 90 million pounds must be delivered in periods of the same length through 2002. And it means that if 100 gigawatthours of electricity are delivered in a given period now that 90 gigawatthours must be delivered in periods of the same length through 2002.

It is my personal opinion that it is not possible to fix to my "good enough" level: there is not enough time left.

If we reach the fix to 90% level, I expect the financial equivilent of the 1970's (I don't remember the exact year) oil embargo or somewhat worse.

If we reach the fix to 80% level, I expect the financial equivalent of the 1928 crash and a nationwide dustbowl rolled into one.

I think that even at the 80% level there is a chance that we will loose society because of the world-wide domino effect (both inward and outward).

As levels fall below 80% it becomes more and more likely that we loose society.

These are only my opinions. Your milage may vary.

Gerorge

-- George Valentine (GeorgeValentine@usa.net), June 02, 1998.

Just a little note here. When talking about all these percentages, it reminded me that coal fired electrical generation plants used to use electrostatic precipitators to reduce particulate output.

Are these type of precipitators still in use?

I remember an engineer telling me that these would consume around 40%(?) of a plants output. If these are still in use, perhaps we could gag and bind the EPA and allow greater polution so people won't freeze to death?

-- Ken Seger (kenseger@tseinc.com), June 02, 1998.

Paul,

Good to see some thinking like this.

George and Ken: Good to see people thinking right away that when it comes to this whole mess, electricity is at the top on the list. (No other remediation effort will matter without it, whether it's good enough or perfect). In terms of "ultimate triage," definitely, electricity needs to be good enough. What that means or how that's achieved is way beyond me, so I'll keep myself out of the way on this one.

Except for this... In reading this thread so far, I was reminded of an excellent perspective on how to go about making the basic elements of the infrastructure "good enough." It's well worth reading, being aware of, and referring anyone to who's serious about finding a way to keep things up and running. It was written by Harlan Smith (a link to a background article: retired electrical engineer; tremendous amounts of "solutions-oriented" y2k thinking; many years experience working with "large scale embedded systems.") He calls it his Synergistic Mitigation and Contingency Preparation proposal, and it can be found at: http://2000.jbaworld.com/harlan/smcp.htm, and at http://www.scotsystems.com/harlany2k.html. (Same thing, different formats).

Something to know about, something to pass on and refer others working on the problem to... Good ideas about how to make it "good enough."

Good thread. Keep it going.

Bill

-- Bill (billdale@lakesnet.net), June 02, 1998.

What about the compatability of the fixes? North believes that data exchange is the biggest unsolvable problem of Y2K. He claims that most organizations have taken the easy route with the windowing technique instead of expansion which requires data file modification.

At the IBM site there are four remediation techniques:

2.4.1.1 Solution #1: Conversion to Full 4-Digit-Year Format 2.4.1.2 Solution #2: Compressed Date Data 2.4.1.3 Solution #3: Windowing Techniques 2.4.1.4 Solution #4: A 2-Digit Encoding Scheme

IBM then states:

"When selecting a proposed Year2000 solution, evaluate the following factors:

1.What is the external impact due to incompatible date format changes?

That is, what other programs or what output will be affected and to what extent will those programs require change if this solution is implemented for this particular program?"

I haven't found a free comprehensive analysis of the data import/exchange problem anywhere on the web. Anyone?

-- Reg Smith (rrr@ibm.net), June 03, 1998.

Dave said

"If any co. or corp. has a deadline contract or penalty clause contract, (because any delays will cause problems up stream), or if any vendor has trouble supplying their customers, in fast demand service, (hospitals etc.), then nothing less than the full remediation of all systems that are necessary to fulfill contracts, would be acceptable. "

I disagree on economic grounds. Suppose a business has the following two choices:

A. Complete remediation by 1/1/200 at a cost of $A. B. Partial remediation by 1/1/2000 allowing for operation, albeit slowed, and full remediation by 1/1/2001 at a cost of $B for remediation and $C for penalties, lost business, etc.

If $A > $B + $C, option B becomes an awfully easy choice to make. For that matter, it wouldn't be all that hard to make even if $A < $B + $C if there simply is no way to meet option A.

Paul

-- Paul Neuhardt (neuhardt@ultranet.com), June 03, 1998.

Paul has stated:

"If $A > $B + $C, option B becomes an awfully easy choice to make. For that matter, it wouldn't be all that hard to make even if $A < $B + $C if there simply is no way to meet option A".

Perhaps this would be true for the low impact service co.s.

Those who decide not to spend the $$ in critical service supply, will reap the cost on the other side, (perhaps in lives).

One needs to weigh the eventual costs. If remediation takes your savings account, your Corporate Assets, and your family fortune to remain in place, perhaps you should consider the cost. However, if you are lax in your remediation and the Lawyers catch up with you or your customers sue or there are real penalty clauses in your contracts, you may as well pay up front.

The other aspect in this question is, even though you come to full remediation and any other sector that impinges upon you does not, you have spent all that money and are no better off than if you did nothing regarding your remediation.

Then, if your employees ever get wind that you did not do all that you could have to protect their jobs, you might find that plan "B" was not good enough after all.

-- Dave Jones (dfj@fea.net), June 03, 1998.

I didn't say that partial remediation now was always the *right* choice, just that it might sometimes be the *easy* choice. And let's face it, easy often wins in this world over right.

-- Paul Neuhardt (neuhardt@ultranet.com), June 04, 1998.

"And let's face it, easy often wins in this world over right."

I would certainly agree with you in this Paul.

"Easy" will be a very short lived term. What most are going to be up against is the degree of "easy", in terms of remediation? What will happen as the timing spins down, is that most will loose the edge on awareness and the lethargic complacency of over confidence will set in. More bad decisions and shortcut measures will take place as the desire to walk away with at least some of your wealth. It probably won't take long for Business owners and Investors to see that the "remediation" monies would look a whole lot better in their pockets as they come to grips with the fact that even if they are compliant, up-stream and down-stream may not be.

-- Dave Jones (dfj@fea.net), June 04, 1998.