CLPEX Chat Board

Haven't had a chance to read this yet. (It is 248 pages.)

This is the full article that Christophe posted about last week.

Latent Print Examination and Human Factors: Improving the Practice through a Systems Approach

http://www.nist.gov/oles/upload/Latent- ... actors.pdf

Yes, it is lengthy. It covers a breadth of topics and views. But I think some of the chapters are very strong and helpful. I like the process map on p. 3 too. Helpful for showing the decision points of ACE-V and showing that there is some structure and logic to the framework of ACE-V.

There is a summary chapter (ch. 10) I think which puts all the recommendations in a combined manner to stress the major recommendations of the group.

Like any document from a large group of disparate experts, it represents a lot of view points to greater or lesser degrees in some individuals, but overall, I think there is a lot of good info and resources in the document. And lots of good references.

I (and I am sure many of the other authors) would love to hear the public voice on this and hear feedback from the community. It would be nice to know if we accomplished anything helpful for the community, or wasted 2.5 years of our lives, and killed lots and lots and lots of trees during the reviews.

Cheers,

g.

Dear g. --

Can you, please, in 50 words or less, sum up the salient points of any serious problems you found? Not necessarily the arguments du jour of the critics and defense attorneys, but any real issues that we need to be aware of, or that need to be addressed through research?

Thanks,
t.

In 50 words or less??? We barely got the TITLE to be 50 words or less!!!! (it was 27 words!!)

In all seriousness, Taz, I don't think I can reduce to just a few succinct points. The report is lengthy b/c all players involved I think appreciate the complexity of what we do. I don't think I should be the spokeperson for this group. Kasey Wertheim and Melissa Taylor wrote an article related to the detection and reporting of errors and were chairs on this committee. Perhaps one of them might have a short snippet to contribute.

My sincere advice is read Chapter 10. Pages 197-210. It's 13 pages long and hits all the high points of the report. If you want a snapshot of what the report contains read that. Some people who read it (on this forum) will be incensed, some will think we didn't go far enough, and some will see where we came up with the best compromise for a group of many people of disparate backgrounds, knowledge, and view points.

Sorry I can't be more helpful. This one is in the court of public opinion now.

g.

Well, okay then. Thanks for at least keeping it under 50 pages. I'll start with that chapter and expand out from there.

I think you should've named this 'What Glenn thinks' and stapled all your published works together and called it a day.

In all seriousness though I've only read chapter 3, and why I'll give credit to the fact that it's trying to address issues, it doesn't say much.

My personal interest these days is the valuation of information so that's why I read chapter 3 first because it deals with interpretation generally and feature weighting specifically.

Thoughts: it gives way too much real estate and weight to 'bias' and primarily talks about feature weighting in terms of probabilistic frequency. Ever wonder why Dror et al don't ever provide a causal link between bias and say...feature extraction. Let's say I'm in the most caricatured hypothetically biasing situation in the world: I'm a cop/examiner, and someone killed my wife. They find a print on the knife sticking out of her chest and they caught a guy fleeing with blood all over his hands. They give me the latent, the 10P card and everyone stays around and watches. These factors in and of themselves do not A) alter the physics of light to distort en route to my eyes B) do not affect the visual regions that perform brute force feature extraction (read gestalt organization in the visual cortex). This leaves only higher order cognitive functions left to hone in on and at least gets us to a plausable place where valuation takes place such as the prefrontal cortex, parietal and limbic lobes and those areas associated with the processing visual information and of abstraction to name a few. I may be dense, but please point that section of Dror's paper to me. Anyone familiar with Semir Zeki's Laws of the Visual Brain and V.S.Ramachandran's Eight Laws of Artistic Experience with regards to neuroaesthetics can then at least begin to speculate about the true nature of bias, because bias is at its core 'undue weighting'. Bias has just become synonymous with 'boogey man' (I just tell my kids that Itiel Dror is under their bed)

If this publication doesn't even begin to recognize what bias is, or how valuation is performed outside the discussion of probabilism, well I have a serious problem with that. I said this in another post, but I'll say it again, examiners are not determining the rarity of features in a population. Call the evaluation a function of subjective probability if you like because it captures the cognitive acceptance of evidence, fine. Say that some features like deltas, cores and ridge flow for that matter are binarily deterministic (either they're there or they aren't) due to their unchanging frequencies in the population, fine as well. But if you intend to tell me that examiners can keep track of frequencies of combinatory features, I call B.S. Experiments show you can’t even remember a 7 digit phone number while holding a piece of cake. However, this book demonstrates the serious deficit of understanding that resides at the level of those who are supposedly ‘leading the way’.

Thanks Boyd, I've come to expect no less...

That said, I am not so sure we disagree that much. There was tremendous discussion and compromise about the extent to which bias should be discussed in this report. As said, it represents the viewpoints of many authors. I would look carefully at that list of authors and you can imagine all the thoughts and personalities that go into those long and lively discussions. I also agree that humans are not generally capable of explicitly computing frequencies of rarity for combinations of features, but my recollection of Chapter 3 is that it says that analysts are implicitly, subjectively, and subconsciously doing this. As you yourself say, there is a weighting process and bias is just "overweighting". These views are not at odds.

I don't get your point about these texts though. If I read your post correctly, you seem to be indicting the authors of the report (and any other cognitive researcher such a Busey, Lin, and others emerging) for not using these texts as "authoratative treatises" on the definition of bias. i.e. if you don't cite these texts, then CLEARLY you don't know what bias is. I have little knowledge of cognitive psychology at all so I guess this point is lost on me, other than that you feel this report and all the other researchers know little of what bias truly is. Am I misunderstanding your view point?

g.

List of Authors

David H. Kaye, JD, MS, (Editor in Chief) Distinguished Professor and Weiss Family Scholar, Dickinson School of Law and Graduate Faculty Member, Forensic Science Program, Eberly College of Science, Pennsylvania State University
Thomas Busey, PhD, (Editorial Committee) Professor, Department of Psychological and Brain Sciences, Indiana University
Melissa R. Gische, MFS, (Editorial Committee) Physical Scientist/Forensic Examiner, Latent Print Operations Unit, Federal Bureau of Investigation Laboratory
Gerry LaPorte, (Editorial Committee) Forensic Policy Program Manager, National Institute of Justice
Colin Aitken, PhD, Professor of Forensic Statistics, School of Mathematics, University of Edinburgh
Susan Ballou, MS, Law Enforcement Standards Office, National Institute of Standards and Technology
Leonard Butt, Chair, Scientific Working Group on Friction Ridge Analysis, Study and Technology
Christophe Champod, PhD, Professor, Institut de Police Scientifique, École des Sciences Criminelles, Université de Lausanne
David Charlton, PhD, Surrey and Sussex Police Forensic Services, UK
Itiel E. Dror, PhD, University College London and Cognitive Consultants International
Jules Epstein, JD, Associate Professor of Law, Widener University School of Law
Robert J. Garrett, Past President and Chairman of the Board, International Association for Identification
Max M. Houck, PhD, Co-Chair and Principal Analyst, Anser
Edward J. Imwinkelried, JD, Edward L. Barrett, Jr. Professor of Law, Director of Trial Advocacy, University of California, Davis
Ralph Keaton, Executive Director, American Society of Crime Laboratory Directors/Laboratory Accreditation Board
Glenn Langenburg, MS, Forensic Scientist, Minnesota Bureau of Criminal Apprehension
Deborah A. Leben, MS, Lead Fingerprint Specialist, United States Secret Service, Department of Homeland Security
Alice Maceo, Forensic Lab Manager, Las Vegas Metropolitan Police Department Forensic Laboratory
Kenneth F. Martin, Detective Lieutenant (Retired), Crime Scene Services Section, Massachusetts State Police
Jennifer L. Mnookin, JD, PhD, Professor of Law, University of California, Los Angeles
Cedric Neumann, PhD, Assistant Professor, Department of Statistics and Forensic Science Program, Eberly College of Science, Pennsylvania State
Joe Polski, Forensics Committee Member, International Association of Chiefs of Police
Maria Antonia Roberts, MS, Research Program Manager, Latent Print Support Unit, Federal Bureau of Investigation Laboratory
Scott A. Shappell, PhD, Professor, Department of Industrial Engineering, Clemson University
Lyle Shaver, Forensic Scientist Supervisor, Virginia Department of Forensic Science
Sargur N. Srihari, PhD, SUNY Distinguished Professor, Department of Computer Science and Engineering, State University of New York at Buffalo
Hal S. Stern, PhD, Professor, Department of Statistics, University of California, Irvine
David Stoney, PhD, Chief Scientist, Stoney Forensic, Inc.
Anjali Swienton, MFS, JD, Director of Outreach, National Clearinghouse for Science, Technology and the Law (NCSTL), Stetson University College of Law
Mary Theofanos, MS, Computer Scientist, Information Access Division, Information Technology Laboratory, NIST
Robert M. Thompson, Program Manager, Forensic Data Systems, Law Enforcement Standards Office, National Institute of Standards and Technology
John Vanderkolk, Laboratory Manager, Indiana State Police Laboratory
Maria Weir, MA, Supervising Forensic Identification Specialist, Los Angeles County Sheriff’s Department
Kasey Wertheim, MBA, Co-Chair, President, and CEO, Complete Consultants Worldwide, LLC
Melissa Taylor, Study Director, Law Enforcement Standards Office, National Institute of Standards and Technology
Mark D. Stolorow, Director, Law Enforcement Standards Office, National Institute of Standards and Technology
Jennifer L. Smither, Lead Editor, Science Applications International Corporation
Kathi E. Hanna, PhD, Consultant Writer and Editor
Shannan Williams, MPP, Program Assistant, Booz Allen Hamilton

Thanks for replying, Glenn.

My point, for clarification, is that the way meaning and value are ascribed through sign relation has been studied for over 100 years. My personal area of reading these days is semiology, which goes back to 1903 with C.S. Pierce, along with the emerging topics Neuroeconomics and Neuroaesthetics. Through those fields, I've found direct parallels between visual representation, preference/value and decision making from a conventional, utility and aesthetic perspective with social/biological/cognitive underpinnings.

The way that relates to this thread is that most current research lacks foundation. Is Dror's research illuminating? Not without knowing the practical difference of what we would expect to see given the specific manipulation of a different variable. There's an ignorance of basic historical perspective of Mill's Methods or the whole philosophy of Pragmatism and the notion of practical difference. (What would we expect to see if things were different, and then holding some variables constant while altering others) I could recreate his study (if it were transparent) and add the additional task of evaluating each Examiner's diet. I could then come up with a statistically signficant ratio of protein to carbs eaten by the examiners and cite that as the source of their mistakes. Sub a nebulous 'bias' attribution for my nutritional ratio and Dror has done the same thing. Neither example shows correlation as to why one variable is impactful on the other.


Allow me to demonstrate a very abbreviated theory of value:

To give one example:
A delta is called a delta because it looks like a delta (the most profound thing you'll hear me say on a monday). Another way of saying this is that the naming convention of this structure is iconic - it looks like what it is called. It's also why a lot of literature and examiners still call features, lakes, dots, hooks, etc. The delta (such as in the instance of a triradius) is an example of this. It is also the phenomenon I was illustrating to you in the question mark print that statistics are blind to, but is integral to what we do.

There are other ways signs gain meaning as well (Note: they aren't mutually exclusive)

Indexical signs are signs that point to other signs (Eg, where'es there's smoke there's fire). This is why when you see three ridges run into an area of low clarity and two run out, you infer a ridge event. It's also why multiple v's (vvv or ^^^) represent a double tap. The last example represents a vertical nature of representation, as the notion of double tap is a meta representation. That is to say, it is a level of abstraction above just the meaning of the features themselves.

Lastly symbolic signs are signifiers that have no direct resemblance to that which they signify. They are agreed upon by convention. In the field of fingerprints a core can be thought about as a sign of convention.


Signs also have modalities, aesthetic cues and temporal aspects. It's the reason when people go see a horror movie, 911 is not flooded with calls reporting murders by all 125 people in the theatre. We can distinguish that what we are seeing is not real by the cues we've picked up on. Purchasing a ticket, the 2 dimensionality of the screen, the purchasing of snack foods. In very much the same way, we utilize similar modalities when we orient a piece of palm based upon the ridge flow.


Temporally speaking Examination has a temporal relativity. It is both synchronic - (temporally simultaneous) and diachronic (across time). Going back to the example of the delta, you are not evaluating the corners of the triradius separately across time (diachronically), but rather seeing them as a singular entity (synchronically). We do this most frequently with 'target groups'. Synchronic value and Diachronic value will be different.

So, right away we see we deal with signs, what is signified by those signs, the way in which signs get signficance, and a temporal aspect (synchronic/diachronic) Sounds a lot like an elaboration of the spatial/sequential narrative, huh?

So there are combinations of all these effects in the print. Signs combine to form bigger signs, which act as the signs for bigger signs, etc. which snowball until we reach a tipping point (threshold of sufficiency anyone?). Pierce described the process as 'infinite semiosis'. That current signs flow from past signs and into future signs reaching a domino like effect of meaning.http://plato.stanford.edu/entries/peirce-semiotics/ It's akin to how letters turn into words which turn into sentences which turn into paragraphs which turn into a treatise which allows for meaning to be conveyed and ideas to be shared. It's information theory.

It's also how a sheep can conceptually be represented as a collection of cells to a specific animal (eg Dolly) to an asset of a farmer to an abstract representation of wealth. Applied to fingerprints, it's how a collection of organic and inorganic molecules collect along a ridge, and are transfered to an object to form intrinsic ridge characteristics, which form ridge events, which form a ridge flow which comprises an impression which is iconically contrasted against other impressions to reach an abstract concept we know as 'Individualization'.

All of that to say this: A publication on the study of human factors that doesn't touch on some of the most basic elements of how humans make/perceive/agree upon meaning doesn't say very much at all.

Authoritative texts that support what I am saying include these:

C.S. Pierce
Peirce, Charles Sanders (1931-58): Collected Writings (8 Vols.). (Ed. Charles Hartshorne, Paul Weiss & Arthur W Burks). Cambridge, MA: Harvard University Press

Ferdinand de Saussaure
Saussure, Ferdinand de ([1916] 1974): Course in General Linguistics (trans. Wade Baskin). London: Fontana/Collins
Saussure, Ferdinand de ([1916] 1983): Course in General Linguistics (trans. Roy Harris). London: Duckworth


Jaques Derrida
Derrida, Jacques (1974): 'White Mythology: Metaphor in the Text of Philosophy', New Literary History 6(1): 5-74
Derrida, Jacques (1976): Of Grammatology (trans. Gayatri Chakravorty Spivak). Baltimore, MD: Johns Hopkins University Press
Derrida, Jacques (1978): Writing and Difference (trans. Alan Bass). London: Routledge & Kegan Paul
Derrida, Jacques (1981): Positions (trans. Alan Bass). London: Athlone Press


Claude Levi-Strauss
Lévi-Strauss, Claude (1961): Tristes Tropiques (trans. John Russell). New York: Criterion
Lévi-Strauss, Claude (1964): Totemism (trans. Rodney Needham). Harmondsworth: Penguin
Lévi-Strauss, Claude ([1949] 1969): The Elementary Structures of Kinship (trans. James Harle Bell, John Richard von Sturmer & Rodney Needham). London: Eyre & Spottiswoode
Lévi-Strauss, Claude (1970): The Raw and the Cooked (trans. John & Doreen Weightman). London: Jonathan Cape
Lévi-Strauss, Claude (1972): Structural Anthropology (trans. Claire Jacobson & Brooke Grundfest Schoepf). Harmondsworth: Penguin
Lévi-Strauss, Claude ([1962] 1974): The Savage Mind. London: Weidenfeld & Nicolson


Umberto Eco
Eco, Umberto (1965): 'Towards a Semiotic Enquiry into the Television Message', In Corner & Hawthorn (Eds.) (1980), pp. 131-50
Eco, Umberto (1966): 'Narrative Structure in Fleming'. In E. del Buono & Umberto Eco (Eds.): The Bond Affair. London: Macdonald
Eco, Umberto (1976): A Theory of Semiotics. Bloomington, IN: Indiana University Press/London: Macmillan
Eco, Umberto ([1980] 1984): The Name of the Rose. London: Picador/Secker & Warburg
Eco, Umberto (1981): The Role of the Reader. London: Hutchinson
Eco, Umberto (1982): 'Critique of the Image'. In Burgin (Ed.), op. cit.
Eco, Umberto (1984): Semiotics and the Philosophy of Language. Bloomington, IN: Indiana University Press

Roland Barthes
Barthes, Roland ([1953] 1967): Writing Degree Zero (trans. Annette Lavers & Colin Smith). London: Cape
Barthes, Roland ([1957] 1987): Mythologies. New York: Hill & Wang
Barthes, Roland ([1964] 1967). Elements of Semiology (trans. Annette Lavers & Colin Smith). London: Jonathan Cape
Barthes, Roland ([1973] 1974): S/Z. London: Cape

Jacques Lacan
Lacan, Jacques (1977): Écrits (trans. Alan Sheridan). London: Routledge


Roman Osipovich Jakobson
Jakobson, Roman (1960): 'Closing Statement: Linguistics and Poetics'. In Sebeok (Ed.), op.cit., pp. 350-77
Jakobson, Roman (1971): 'Language in Relation to Other Communication Systems'. In Roman Jakobson (Ed.): Selected Writings, Vol. 2. Mouton: The Hague, pp. 570-79
Jakobson, Roman & Morris Halle (1956): Fundamentals of Language. The Hague: Mouton

My personal area of reading these days is semiology, which goes back to 1903 with C.S. Pierce, along with the emerging topics Neuroeconomics and Neuroaesthetics. Through those fields, I've found direct parallels between visual representation, preference/value and decision making from a conventional, utility and aesthetic perspective with social/biological/cognitive underpinnings.

Just because it's what you're reading right now, doesn't mean that it's pertinent to every bit of cognitive research going on around the world. From the rest of your post, it doesn't sound like it fits as well as you think it does. Trying to adapt semiotics, and it's ideas on ideology, myth, and culture, to fit friction ridge comparisons is not a given. Your ideas might make a good article for VS. However, it's not a given that a book on Latent Print Examination and Human Factors or even a chapter on Interpreting Latent Prints MUST include semiotics.

In any case, semiology would be a whole lot more interesting as the basis for your own set of experiments, instead of as a complaint that it wasn't a part of book that wasn't about semiology.

ER wrote:Just because it's what you're reading right now, doesn't mean that it's pertinent to every bit of cognitive research going on around the world.

Careful, straw men burn easy. Those are your words, not mine.

Perhaps you haven't noticed Eric, but for a while I've found the research being sold to us a little oversold on its meaning and value. If it were as relativistic to what I was reading at the time as you'd like to believe, you'd see a little more frenetic response.

As you said in this post:http://clpex.com/phpBB/viewtopic.php?f=2&t=1770&start=0

ER wrote:As Boyd states, there are qualitative aspects to fingerprint comparison that cannot be accounted for in the model..

So, let's hear your theory of value. Since in your own words you agree at least with the notion of value. What qualitative things do statistics leave out and how do you ascribe them value. I've given you insight into what I think, now lets hear how yours differs.

Virtually every time I've heard the pitch about the value of a statistical model, it has been with the caveat that it only deals with 2nd level points and the physical distances between them. Intrinsic in that description is that the models do not as yet have any way to deal with 3rd level detail, ridge counts, or quality. Whenever these flaws have been pointed out, the researchers acknowledge them. Speaking of straw men: how can the concept fail because they don't address these issues when they acknowledge that they don't address these issues.

But I didn't raise any of these points. I posted about your criticisms of the recent NIST paper, which seem to be:
Glenn was involved, so it's wrong.
Zeki and Ramachandran aren't mentioned, so it's wrong.
It claims that examiners can keep track of frequencies of combinatory features in their heads, so it's wrong.
(BTW, who said this? I've never heard anyone claim that examiners do, except in a vague, 'that's pretty rare' kind of way, or a 'I see that pretty often' kind of way.)
Semiosis isn't mentioned, so it's wrong.

My last post was to say that I disagree with the logic behind your criticism. Failing to mention semiosis does not disqualify the points made in the paper. Now, I'm sure I can find a list of things in this paper that I disagree with, and I'll write about them soon, but it's failure to mention semiosis is not one of them.

To answer your specific question, some of the research has begun to account for quality by only using the features that are seen by a number of different examiners. Other than this binary accounting, statistics currently seems to take no account for quality. But again, no one is claiming that it does, so I don't see what your concern is. I've heard the researchers making the same points that you do on this forum with one exception: your point is followed by, "...so they're headed in the wrong direction," while theirs is followed by, "... so we need more work in this direction."

All I'm saying is, criticize this paper for what's in there (there's probably plenty), not for failing to include some esoteric subjects. And, focus your criticisms on the limitations that the author's missed, not the ones that they acknowledge.

It would be silly to refute your criticisms since I never said any of those things. If I did, perhaps you can quote me. It would looks something like this:

Since when asked for your theory of value:

b.. wrote:So, let's hear your theory of value. Since in your own words you agree at least with the notion of value. What qualitative things do statistics leave out and how do you ascribe them value. I've given you insight into what I think, now lets hear how yours differs.

you reply with:

ER wrote:To answer your specific question, some of the research has begun to account for quality by only using the features that are seen by a number of different examiners

Apparently, citing vaguely other's research and not giving an explication is what you interpret 'let's hear your' to mean. Since specifically asked and lack of response, I can only think at this point you don't have a theory of value.... It must make case work and testimony difficult.

While it's fun to philosophize, as the familiar phrase goes "the proof is in the pudding" Let's explore the three theories of value that have been put out there on this oh so favorite print of mine. Eric's theory of Pornography "I won't/can't define it, but I know it when I see it", Glenn's statistical foundation, and my system of signs.

The print to analyze:



How do you give the core value and what specific normative methodology does this value dictate?


If you even call this print a whorl, or say that the core looks like a question mark, you've stepped into the theory of value that I've put out because you are using iconic representation to describe them.

The prescription is to either look in the 10P card for a whorl with a question mark in the center and begin to diachronically(sequence) evaluate the agreement of other information spatially. The other information can also be represented and valued making it comparable. Modally speaking the ridge shape 3 ridges away from the bottom of the question mark at about 4 O'clock that looks like an open mouth is more valuable due to the degree of asymmetry relative to the surrounding areas. This valuation of symmetry is modal to the valuation of information in the print. Statistics is blind to such modality, which through a simple reading of my posts is one of the things I say all along. Relative frequency is also a modal element of consideration in valuing the question mark. If this print had a lot of question marks in it, it would not have the same value.

Discussion:

You will also note that I am not saying value is absolute (eg ridge ending = 9). Value is context dependant and differential. In part, you know value by knowing what it is not in the surrounding areas. You also know value positively, as in the case of a question mark (it IS a question mark). This is the same way something like a stop light gains value. Red = Stop. Part of the way you know this is because it is not yellow or green. Illustrating the influence of time on meaning such as in sequencing agreement in an impression, we can look at the stoplight again. Green at time X = Go. Green at time X+n can equal 'caution', the same as yellow, because we would describe it as being 'stale'. Value is differential, contextual and time dependant. Another analogy would be that of a pawn at the beginning of a game of chess and a pawn later in the game that has a king in check. Relational, differential, temporal, context dependant. The notion, unlike the coin analogy that was used to describe value is that in the coin analogy the value is absolute, a penny is always less than a nickel and sometimes you have more nickels and sometimes you have more pennies. But it does not tell you how a penny becomes a penny nor does it tell you that not all pennies are the same. Because even in real life that' the case, if I have a 1922 No "D" (there is no mint mark below the date) penny, that's worth one thing, if I have a bronze/copper penny, the price of wholesale copper may be what dictates the value, or if I have a 2012 penny, it might just be a penny. You will also note I am not saying I disavow statistics, frequency does matter, it just matters more on a relative local level than the 'did you look at every fingerprint in human history' level. This issue here is the tension between these elements. If any element has too much prominence, you get perversion of valuation. (ask any economist). Again, this is a problem I have with trying to push statistics on everyone.


Ok, so now that I've gone first. Let's hear what you think Eric. Rather than put words in your mouth, I'll let you speak for yourself, but since you think my logic is flawed, I would expect your explanation to be radically different than mine.

It would be pointless to post my analysis of the value of the above print because I agree with you. We value some configurations and shapes more than others. That value is not absolute. The value depends on the feature's location within the print. etc. etc.

How is this analysis of the value of the ? print in conflict with a statistical representation of the uniqueness of the 2nd level detail that acknowledges its limitations? How can this print be used to indict statistics when statistics doesn't claim to represent the value of that feature? Doesn't statistics give a relatively accurate representation of the uniqueness of 2nd level feature sets given that it's using distances and not ridge counts?

Why should a report be criticized for not including a discussion on semiotics and neuroaesthetics? You have yet to convince me that these fields are relative enough to be included in the report.


On another topic...

It would be silly to refute your criticisms since I never said any of those things. If I did, perhaps you can quote me.

Glenn was involved, so it's wrong.

I think you should've named this 'What Glenn thinks' and stapled all your published works together and called it a day.

Or did you mean something different?

Zeki and Ramachandran aren't mentioned, so it's wrong.

Anyone familiar with Semir Zeki's Laws of the Visual Brain and V.S.Ramachandran's Eight Laws of Artistic Experience with regards to neuroaesthetics can then at least begin to speculate about the true nature of bias...

I may have misinterpreted your post, but I thought your point was that any paper of value on bias would include a discussion of neuroaesthetics.

It claims that examiners can keep track of frequencies of combinatory features in their heads, so it's wrong.

But if you intend to tell me that examiners can keep track of frequencies of combinatory features, I call B.S.

Again, you set up a proposition that no one has made, and then shoot it down.

Semiosis isn't mentioned, so it's wrong.

All of that to say this: A publication on the study of human factors that doesn't touch on some of the most basic elements of how humans make/perceive/agree upon meaning doesn't say very much at all.

You seemed to define semiosis as the study of how humans make/perceive/agree upon meaning.

Your above description of the valuation of unique features makes sense. I don't understand these criticisms of the NIST report.

Let me break it down a little bit more since the first round didn't exactly sink in. If you still don't understand after this, we'll chalk it up to the fact that you're suffering from a organizational hueristic bias which fits in nicely with the whole meta topic.

1) A publication that deals with a systems approach to latent print examination that doesn't implicity or explicity deal with acknowledging value rating in the interpretation section, misses the point.

Such a publication is akin to an automotive repair book that doesn't mention the engine, but tells how to roll up the windows and check the air pressure on the tires. My desire to see semiotics/neuroaesthetic or even neuroeconomic discussions is a direct acknowledgement of A)the history of friction ridge examination and the history of Philosophy of Science B)the lack of leadership trajectory and ignorance of the ramifications of their publications.


A)--------------

We can see from history that from the beginning, semiotic sign recognition and valuation has taken place. The sourcebook tacitly admits it in the discussion of classification systems. See any resemblance of the letters selected to represent clusters of features in the system? Iconic representation.

https://www.ncjrs.gov/pdffiles1/nij/225325.pdf

[picture] Historically speaking, we can next see semiosis pop up in statistical calculations. This excerpt (again from the sourcebook) shows that the minutiae type was used. An explicit probabilistic valuation based upon ability to standardize and perceive minutiae types. (aesthetics + semiosis)

https://ncjrs.gov/pdffiles1/nij/225334.pdf
I already discussed how the use of a delta was iconic representation in an earlier post, so I'll just reference it.

Lastly, on this very site, we have a repository called what? The Smiley Files. For what reason?? The configurations look like smiley faces (and other iconic representations).

http://www.clpex.com/Smileys/SmileyFiles.htm
So, over the history of our discipline, it can be shown that Examiner's job is to perceive and extract the sublime, which is your tie in to aesthetics. The sublime is defined as the quality of greatness (in this sense aesthetics) especially refering to greatness beyond imitation. Which is exactly what you're saying when you say you've both analyzed and evaluated the impression. In the analysis, you're saying that the impression contains sublime information and that information will allow you to evaluate the quality that this impression is beyond imitation. (you can tie it specifically to a rolled impression if it's put in front of you) And in the Evaluation, you're actually demonstrating that the information contained in the latent impressionis in holistic agreement/disagreement with the exemplar beyond the possibility of imitation by another print.

And that's just wikipedia's definition
http://en.wikipedia.org/wiki/Sublime_(philosophy)

The most renounced philosopher's of science discuss the role of aesthetics in science, but somehow you've put yourself at a place to deny any such role or need for any. That's pretty damn arrogant.

http://plato.stanford.edu/entries/kant-aesthetics/#2


B)--------------

With regards to lack of leadership trajectory and ignorance of the ramifications, let me again refocus to the topic of the publication. Human factors. So Kasey published his HFCS and posted it back in issue 459 of the Weekly Detail.

http://www.clpex.com/TheDetail.htm?issue=459

it's already recognized that organizational climate and culture matter as listed explicity here:

1.2. Organizational Climate

1.2.3. Culture
1.2.3.1. Norms and rules
1.2.3.2. Values and beliefs
1.2.3.3. Organizational justice

And with regards to direct aesthetic judgement by the examiner:

4.1. Errors
4.1.1. Skill-Based Errors
4.1.1.1. Improper analysis
4.1.1.2. Improper comparison
4.1.1.3. Incomplete comparison
4.1.1.4. Improper evaluation
4.1.1.5. Failure to follow the method
4.1.1.6. Failure to prioritize attention
4.1.1.7. Poor technique
4.1.1.8. Failure to reconcile distortion
4.1.1.9. Failure to recognize:
4.1.1.9.1. Intra-sample uniqueness
4.1.1.9.2. Value of the sample
4.1.1.9.3. Distortion artifacts
4.1.1.9.4. Agreement between samples
4.1.2. Decision Errors
4.1.2.1. Conclusion exceeded ability
4.1.2.2. Concluded value in error
4.1.2.3. Concluded no-value in error
4.1.2.4. Concluded distortion in error
4.1.2.5. Concluded disagreement in error
4.1.2.6. Concluded agreement in error
4.1.2.7. Concluded intra-sample uniqueness in error
4.1.3. Perceptual Errors due to
4.1.3.1. Misunderstanding of agreement
4.1.3.2. Misunderstanding of distortion
4.1.3.3. Visual illusion

So we can see that from this, both the values, beliefs and culture set forth by publications like this, that seek to speak for the industry and through the reliance on aesthetic judgements by examiners we've come up with a real deficit in both knowledge and direction (what I would call trajectory)

For instance the ACE-V flowchart does not show the process to be recursive, then the text goes on to say that it can be but that can cause errors. There's no definitive prescription, the document basically says it's not (the flow chart), but it can be, but watch out because there's error on the horizon. That's helpful.......

Secondly on this note, we get the swiss cheese model and all these taxonomies of error which I'm going to predict do nothing but give the next OIG report some buzzwords to use. Something acknowledged by people who we stole it from in the first place. People don't consider that the wishy washy stance and supposed air of authority that this document receives is really only a foothold for attacks in court. What's the answer, well the people who wrote it will come and train you for 700 bucks a pop and tell you a bunch of things, when actually considered against the backdrop of history don't hold up or are just bureaucratic processes that hinder your ability to get cases worked.

http://www.nifc.gov/fireInfo/fireInfo_d ... ssAnly.pdf

The system, which is based upon Reason’s (1990) model of latent and active failures (Shappell & Wiegmann, 1997a), encompasses all aspects of human error, including the conditions of operators and organizational failure. Still, HFACS and any other framework only contributes to an already burgeoning list of human error taxonomies if it does not prove useful in the operational setting.

Lastly, have you notice how I like to make use of the thinking of a lot of dead guys? Do you think that maybe this is because this stuff has been pondered as a philosophy of science for thousands of years now. Yet in our discipline the wheel has to be reinvented in every shape just to find out they already made one million's of years ago and it works just fine.


All of this to say: I have a reason for saying what I say and it doesn't include 'I don't like Glenn'.

...ok.

Besides the subjects that you think they missed, do you have any comments on the content that is in the report?