I think it is time to try to codify some criteria for assigning a 'y2k relatedness confidence factor' (Y2KRCF) to reports posted here. This is my first cut at this and will thus be rough. Perhaps when we have finished the thread someone will prepare a succinct report to document the various ideas.

We need first to establish a scale to measure against. This scale should be linear and ofcourse can at best represent the 'best guess' of the reporter. The scale should not be too complicated lest people find it cumbersome. The rules should not be too detailed lest we wind up so many criteria that only an expert can make the assessment.

I propose a integer based scale of minus five thru five with five representing 100% certainty the reported glitch is related to y2k and minus five representing 100% certaity the reported glitch is NOT related to y2k. Let zero represent absolute uncertainty.

We now have a workable scale. We need to establish critera which can place points along the scale, for example:

Y2kRCF assignment criteria -- cut one

5 if the company (or someone with access to the source code) which owns the software exhibiting the anomaly (glitch) states that the issue is a y2k issue.

3 if users of the software state the issue is a y2k issue.

1 if directly affected clients of users of the software state second hand the issue is y2k related.

0 if there is no indication or certainty of either case

-1 if directly affected clients of users of the software state second hand the issue is non-y2k related.

-3 if directly affected users of the software state the issue is non-y2k related.

-5 if the vendor or owner of the software (or someone with access to the source code) states the issue is non-y2k related.

You will notice that I have not established criteria for certain of the values this scale allows. This is where you come in.

Serious/humorous responses encouraged. I will take the first obvious frivolous response. 5 if the government says the issue is non-y2k related. Now that we have that off our chest...

-- Michael Erskine (Osiris@urbanna.net), January 10, 2000

Answers

How about this?

1. Blow it out your ass!!

-- (Always@mazedATthe bullshit.com), January 10, 2000.

How does one score a self-reported "Fred with issues"?

-- PA Engineer (PA Engineer@longtimelurker.com), January 10, 2000.

Michael: Sounds like a good format and good logic behind it. But logic has'nt been high priority as of late on this site. If Joe Fartwunkles P.U. runs out of gas it must mean G.M. loaded bad chips in the ignition or Standard Oil is going dont for the count. Well, I'll go back to lurking and see how much I pissed of the "Joe Fartwunkles" out there....

-- Capt Dennis (capden@hotmail.com), January 10, 2000.

Always; Since, "blow it out your ass" is more like a SWAG than a client's opinion, I would have to opine that "blow it out your ass" is more likely a zero. Your thoughts?

-- Michael Erskine (Osiris@urbanna.net), January 10, 2000.

Michael,
If the text of a post indicates the source of the information reported therein, it's not clear what a numeric confidence factor would be adding. Perhaps the idea is to include the Y2KRCF in the subject of a post, but an artfully worded subject can also convey this information.

-- David L (bumpkin@dnet.net), January 10, 2000.

David; Thanks I was just thinking along that line as I responded to a mail I received from Ken... I think that such a tag should be established because it will be easier to parse databases and messages in a few months to isolate and identify information which can be used for historical analysis if for no other purpose. In otherwords if some variation of y2krcf or some variation thereof appeared in a message text that message could be singled out and studied prior to inclusion into a database. The presence of the tag would indicate at least one persons evaluation of available source material and may be helpfull.

-- Michael Erskine (Osiris@urbanna.net), January 10, 2000.

Michael,
I got to thinking more about this after my previous answer. The relevance of a report is actually dependent upon many dimensions. For example, a Y2K glitch reported by an anonymous CEO might not carry the same perceived relevance as one in which the company is identified. The government example in your post, though facetious, illustrates the role that a source's past conduct plays in our evaluations.

More broadly, I distrust labels, because they nearly always mislead. They may serve a valuable purpose, as illustrated by the categories into which the posts of this forum are grouped, but the faults of even that scheme are revealed by a post whose content does not concern just a single category. I would prefer that future researchers be left to devise whatever groupings meet their own needs. A grouping we devise to clarify things might end up obsuring them instead.

-- David L (bumpkin@dnet.net), January 10, 2000.

This seems like a good start, but how are we going to measure how serious the problem is? If Y2K is going to have a serious impact on the economy, it will be because it is affecting billing, orders, inventories, movement of freight, etc.

-- Dave (dannco@hotmail.com), January 10, 2000.

Dave; I view this idea as a simple tool that can be used to conduct trend analysis. The trend analysis is what would actually be the thing that can help us to understand how much the software aspects of y2k will affect us. We can also use the tool in a historical context to figure out how much it did affect us. The idea is just something that I thought might be usefull. If sufficient people use it it will prove to be usefull. If not, it will prove to be nothing more than another failed idea.

We really can't judge clearly the impact of these software problems without some kind of metric and even then we will need a method of dividing the available data into more managable data sets. There are probably thousands of ways this can be done. This is just one of them. Once the data is in for the month of Janurary we might be able to get an idea of the trend we are seeing.

Personally I expect we will see that we are stressing the system but it is holding. The $64,000.00 question is are the y2k related glitches increasing in frequency?

I have not watched the situation long enough to make a decision on that. I am pleased that I am able to say that at this point.

-- Michael Erskine (Osiris@urbanna.net), January 10, 2000.

Michael,
I suspect it would be hard to persuade people to include the Y2KRCF in their posts. I'd do it in any incident reports I happen to post, but there haven't been any and I hope this continues. 8^)

You might consider creating your own file or database of thread URLs and their corresponding factors. Many posts can be immediately ruled out as not pertaining to incidents, leaving far fewer than 100 to skim each day. If you can find just one other person with whom to share this load, it might not be too bad.

-- David L (bumpkin@dnet.net), January 10, 2000.

I like your idea. Could you make it a two digit number with a second number for severity.

-- JOHN LITTMANN (littmannj@aol.com), January 10, 2000.

From: Y2K, ` la Carte by Dancr (pic), near Monterey, California

Believe nothing until it is officially denied.

-- Dancr (addy.available@my.webpage), January 11, 2000.