Does Anyone Remember...

…this graphic?





It was shown by my friend Lawrence Perry >>> in his Community blog on 9-Jan-12 at 12:47am: How long can Thomas Power stay No 1 on Ecademy? >>> I have Lawrence's agreement to mention him herein. My comment >>> was, "What a load of nonsense. Who cares?" Well, an Important Person did care, said so, and gave me a good kick, for which he apologised, albeit less publicly, and I forgave him, but this blog is not about that. I mention it only because attention is hereby drawn to it once again. Believe me, it is "water under the bridge". As it happens, I think I was simply misunderstood, and perhaps that is a lesson for us all: to be as clear and unambiguous as possible, and bear in mind our likely audience and their possible reactions. Can we ever attain such a high principle? According to the philosopher Jürgen Habermas >>> in Moral Consciousness and Communicative Action >>>, it requires communicative competence. He argues that when speakers (and presumably also writers) are communicating successfully, they will have to defend their meaning by using these four claims.



1. That they have uttered something understandably — or their statements are intelligible; 2. That they have given other people something to understand — or are speaking something true; 3. That the speaker is therefore understandable — or their intentions are recognized and appreciated for what they are; and, 4. That they have come to an understanding with another person — or, they have used words that both actors can agree upon.


Challenging for bloggers, eh? But to get back to that graphic, and I think you will agree I have not digressed too widely, within possibly a few days, Lawrence kindly contacted me by phone and I took the opportunity to ask about it. I had been unable, any more than probably anyone else, to see what it contributed to the question of how long that nice Mr Power could retain his then current position - hence my stating it (the graphic, in context) was nonsense. Well, would you believe it, Lawrence agreed it was nonsense, and he should know; but also he said he did not know what the graphic really meant nor where it came from. Now I don't blame him for effectively falling short of Professor Habermas' ideals, for I am certain we all do. I seem to have done so with my ambiguous "nonsense" statement, probably not for the first time and maybe not the last - where would be the fun in that? But it spurred me on, to investigate what the graphic might really mean. I found out, and will show you. It has taken some time to get a Round Tuit >>> (remember them? Oh how I wish I had not mislaid mine! ;-) ), but better late than never I hope. As you will see, the original graphic is very far from nonsense in its own context, but nothing whatever to do with anyone's position in the Ecademy league. First, compare this graphic with the original one:





It is similar, apart from explicit labelling of the variables plotted, which helps clarity somewhat but still we need some indication of purpose and meaning. The graphics came from the website of a firm called Innovatrics, which works at the "cutting edge" of technology in the field of biometrics. Probably the best-known biometric technique is fingerprint recognition, and matching systems are the topic of interest. Now in automatic recognition of fingerprints or any other biometric system, and for that matter in statistical methods in research of any kind, there are four possible results to consider, as shown in the 2 x 2 matrix.



http://<a href="http://www.flickr.com/photos/97818214@N00/7342856684/" title="Truth Table by nbadagio, on Flickr"><img src="http://farm9.staticflickr.com/8168/7342856684_8ac2bd96b9.jpg" width="330" height="165" alt="Truth Table"></a>


As shown, so long as the results show a match (positive) when there really is one, and reject match (negative) when there is not, they are true. But it is possible to get a positive result when it should be negative (False Acceptance) or a negative result when it should be positive (False Rejection) - both of which need to be kept at rates as low as possible. In a trial of the system in 2004, comprising tests of all its algorithms (codified processes), the results were tabulated for Innovatrics' product in competition with some 40 others from a variety of firms. This was FVC2004 - Third International Fingerprint Verification Competition >>>; whether or not it has recurred since is not relevant. The results were plotted in a graph with horizontal axis FAR (False Acceptance Rate) and vertical axis FRR (False Rejection Rate), which means it compares how bad all the competing systems were. (The terms FMR - False Match Rate - and FNMR - False Non-Match Rate - respectively, are sometimes used.) Anything further left or down would be good. The EER (Equal Error Rate) Line shows the case when FAR = FRR. Innovatrics are justifiably proud of beating all competitors. Thus, the graphic is very far from being nonsense in its own context. But as it has nothing whatsoever to do with people's relative positions within Ecademy it was indeed nonsense in that context - all the more because there was no attempt at explaining it. The moral of the tale for bloggers is that if you include material in your blogs, either it needs to be self-explanatory, or you need to comment in clear and unambiguous terms about it; never, ever, put unrelated stuff in; and if you need to say something about it is nonsense, make it plain that you mean the specific material not the whole topic. Oh, and if you get others to write your blogs and they put obscure stuff in without telling you where it came from and why, sack them. I can understand why you might... this has taken a lot of time and no little quantity of research. Best Regards

  Maurice


MPC Wessex business clarity. Using mapping to understand your business better


Maurice Poole

mauricepoole-96500

Does Anyone Remember...

Well, Steve, me ol' pal me ol' beauty, the whole point is how to be understood, and if you don't care whether anyone understands or not, the communicative competence doesn't matter to you. I would ask what the point might be of saying/writing anything if understanding it is of no concern. To anyone fortunate enough to remember The News Chronicle (it closed before many people here were born I should think) the daily cartoon series Colonel Pewter >>> by Arthur Horner (which might have continued after the Daily Mail takeover) had an Ancient Briton character, "Glub", revived from suspended animation in an ice block found by the Colonel, say at every opportunity, "Ar blukkalumpf" or something like that, and lack of understanding by anyone didn't matter either. Would that be something approximating to your level? ;-)   Maurice

0 comments

RealSteveHolmes Fading away soon

cvsage-38854

Does Anyone Remember...

Alas I do not agree with the claim that it is necessary to make yourself understood. All original thought is unintelligible until the hearer clear their mind and opens to its meaning. I am incapable of non-original thought. Hence I do not care whether any given person understands me or not. Those that are able to, will. People can only understand according to their maturity and honesty. Most haven't got a clue and to speak to them you can only say tripe.

0 comments

Richard Derwent Cooke

richardcooke-98882

Clarity

I fear that Maurice's key point seems a little eclipsed in comments about individuals. Keep it simple (wherever possible) Nd keep it CLEAR!

0 comments

Maurice Poole

mauricepoole-96500

My apologies...

...to those that still could not understand the graphics. Having thought about it a bit more, it seems to me that in sets of data from the trials, for every value of FRR there is a corresponding value of FAR and together they are plotted as a point on the graph. I don't know how the trials were conducted, but it seems highly believable that when a system errs often to False Acceptance it is unlikely to err often to False Rejection, and vice versa. Those are the opposite ends of each curve. In the middle is the case where the two failure modes occurred at exactly the same rate, shown by the Equal Error Rate Line. The labels FAR100 and FAR1000 correspond to False Acceptance Rates of 1 in 100 (10^-2*) and 1 in 1000 (10^-3*) respectively - duplicating labels really, and the same rates for FRR could have been labelled on the corresponding horizontal lines, but weren't. * shows my ignorance of how to type "10 to the power of" in html allowable on Ecademy! ;-) And that exhausts my knowledge of what the graphics show. Best regards   Maurice

0 comments

Maurice Poole

mauricepoole-96500

Does Anyone Remember...

Not sure what you mean, Stella. None of the stuff I tried to explain, so far as I am able, applied to any of us. Did you suppose it would? Best regards   Maurice

0 comments

Stella Holman "The Connector"

stellaholman_theconnector

Does Anyone Remember...

Interesting Maurice, do you suppose the system is making us all very competitive?

0 comments

jet rotmans

jetrotmans-173923

Does Anyone Remember...

dear maurice it is hard to understand to be honest because we are lazy to understand it and we like the drawings of the graphs we congratulate you with the writing of the blog good wishes warmest regards jet

0 comments

Angus Whitton

anguswhitton-88567

Does Anyone Remember...

Maurice, What a tremendous amount of work - you must have had a busy weekend! I can't say I understand all the graphs I'm afraid, however I think I'm with you in principle, as I use benchmarking extensively in one of my businesses.... The benchmarks are of course externally set and adhere to a common recognised industry standard. They are not set by my company - otherwise they would be of little value. Our clients use these standards to evaluate (benchmark) our standing against our competitors, and is applicable all the way through, from tendering to SLA. It comes back to: at what point are management more use working in the business than on the business? My role is mainly to keep clients 'happy', which means profit for the company, and that is the benchmark by which our shareholders judge us - the fiddly internal bits are just garnish on a Friday night kebab. If I get some time I'll try and dig up my old scribblings from my army days about why the last person - well almost the last person ;-) I wanted to share a trench with would be a General, and said scribblings are actually quite complimentary about the General! Angus

0 comments