March 2017

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31  

« Dad's on the roof | Main | Big media Marcus »

Aug 26, 2010

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Joe Guarino

Of course, this diminishes the value of all those calls we heard to get more data before making a decision to require filters.

But we have to presume there are a significant number of attempts to access porn in the libraries, and at least some of these attempts are successful. Unfortunately, the Cymphonix system apparently will not give everyone the data everyone said they wanted.

Ed Cone

Triangulating based on various sources of info may be the best we can do, but that's better than acting on emotion.

Even knowing exactly how many sites are actually accessed might not be enough information in and of itself. Some people believe that even one accessed page necessitates further action. Some people believe that a handful of accessed pages represents a manageable problem.

cheripickr

Isn't it great the way Ed sees points on both sides of the issue? Both the points that argue aginst putting in filters, as well as those that argue in favor of not putting in filters. Curious approach from a person who has said porn has no place in the public library and he's only interested in the best way to achieve that. And the Just Seems Reasonable Tour and accompanying narrative rolls ever onward.

Sue

Pesky facts. Porn filters are just a non-issue. If you went to yesterday's State of the Community luncheon (after the invocation stuff), you'd learn we have actual problems in this city that need fixing.

Spag

I have noticed that Ed continues to talk about the need for more "information" but never settles on any solution. That is by design so whatever happens, he can always link back to something he wrote earlier and say "See how smart I am. I told you so." The opening act of the Just Seems Reasonable Tour is Fonzie. The crowd of Dwarfs always goes wild and demands an encore.

Joe Guarino

Sue, the fact that the community has other problems does not mean that permitting porn surfing in the library is a non-issue.

Ed, the fact is that we do not know that we are dealing with a "handful of accessed pages". We do not know the precise numbers, and it appears we will not know the precise numbers.

No one is proposing that we act based on emotion. We are proposing that we act based on what a moral society composed of responsible adults would ordinarily do. We would act to create an environment for the safe, healthy and moral upbringing of young people. We would make some very modest adult sacrifices to make that happen.

In order to do so, we would need to correct the moral error instigated by Ms. Neerman and the library board.

Ed Cone

I think we can use this info (understood accurately) in combination with other info available to get a reasonable understanding of the scope of the problem.

I agree with the Council's decision to allow library staff to set policy, and I thought Mary Rakestraw spoke well when she reminded staff and the GPL board of their responsibility to respond to any information that might challenge their current approach to this issue.

FlaPhyllis

I agree with Dr. Guarino, but I do not know why he insists on moral squishiness. Filters will not stop all porn, so they are an insufficient solution to the demands that creating a safe, healthy and moral upbringing of children places upon us.

We cannot capitulate to an incomplete solution on such a moral imperative. I am sure a little soul searching will have Dr. Guarino agree with me that the only way we can meet our responsibilities to children is to abandon the accommodation of any adult interests and disconnect the library computers completely from the internet. Otherwise we are just talking a matter of degrees which may as well allow adults the option of turning filters off. You can't be mostly moral on this one.

Dave Ribar

Joe:

You wrote that "we have to presume there are a significant number of attempts to access porn in the libraries." No such presumption is actually warranted.

Let's stick with the 700 daily attempted hit figure. First, according to the library staff, that figure represents only 0.15% of all attempted hits. This means that of all of the attempted hits generated by library patrons, only one request in 667 goes to a site that is identified as pornography. If (and this is a big if) page requests are proportional to file requests (hits), porn requests are a miniscule and insignificant portion of web requests. If porn pages are "busier" (have more content) than other pages (someone more familiar with those pages might be able to tell us), porn page requests become an even smaller proportion of all page requests.

Second, while the 700 figure includes all computer attempts at accessing material, it doesn't necessary reflect user requests or intent. If someone accessed an otherwise appropriate site that contained ads from porn sites, those ads would count even though the user had no control over whether they came up (the ads might not even appear on the screen). Open the wrong e-mail or click on the wrong, innocent-looking link, and the same thing can happen.

Third, while we know that some attempts were successful, we also know that the 700 daily hits were identified by the same software that does the throttling, so all of those attempts would have been subject to throttling. We don't know how many file requests (hits) actually got through.

Fourth, we have more evidence than just the Cymphonix report--the incident reports give us some indication of the effectiveness of the throttling software. The library staff reported 89 incidents of inappropriate viewing during the first half of 2009; with new software and procedures the number of reports had decreased 80 percent by this year to less than one incident per week. The staff have instituted a number of procedures to reduce inappropriate incidents.

As you have stated, no filtering technology is going to be perfect, so incidents will continue. The policing question therefore becomes a matter of how many of the remaining 3 monthly inappropriate viewing incidents (or if you prefer, how much of the successful portion of one out of every 667 requests) can be stopped, at what cost to the library, and at what loss of convenience to the VAST majority of other patrons who are using the library for appropriate purposes.

JC

I think filters should be installed with no opt-out capabilities. If you find a legitimate site you want to view blocked, you should be able to ask a librarian to unblock. The librarian should then send that information to whomever maintains the filter. The library patron could also directly contact the filter company. Also, it would be good for more libraries to use the same filters, so that information on what should/should not be blocked can be shared and the filters improved. Having to ask a librarian to have a website unblocked is not an unreasonable expectation. No one would be completely blocked from a legitimate website, and as to if a website is legit or not, isn't the Supreme Court standard, "you know it when you see it"?

Most of what I would go to a library to get would be unavailable at a city's public library, and probably not even at any of UNCG's (I'd have to go to Raleigh). I understand that because it is not in that library's core mission and the costs do not outweigh the benefits. We really need to think what the library's mission is, and how activities there can detract from that.

Spag

"I think we can use this info (understood accurately) in combination with other info available to get a reasonable understanding of the scope of the problem."

Okay, so what else do we need to understand the scope of the problem, Ed? Give us a list instead of some insistence on a never ending litany of "reasonable" things in response to everything else that comes out.

What would it take for your "triangulation" to be complete such that you in your infinite wisdom could actually offer an opinion as to what should be done, if anything?

glenwoodobserver

JC - what about a teenager who may be researching erectile dysfunction, or curious about the changes his body is going through during adolescence? Do you really think that teenager is going to go to the librarian to ask that it be unblocked? My guess is that they will already be embarrassed without having to jump through that hoop.

I just think it odd that Joe is positing his solution as the only moral one. I don't see the library's job as one of being the morals police. That really ought to be left to parents.

Ed Cone

As the post says, another source of information comes from those incident reports. "Combined with other data, including accurate software-tracking numbers, they help us triangulate to reach a reasonable estimate of the problem's scope, and the efficiency of solutions in place."

I'd also include staff observation and user feedback in that mix.

Right now, based on what we know, I agree with Council that staff is pursuing a solid program. I trust staff to continue to tweak its procedures along the way.

When is the info gathering and analysis complete? Never, it's an ongoing process, or at least an iterative one. Changes in technology, or evidence of an increased problem, for example, could lead to changes in policy.

Roch101

" If you find a legitimate site you want to view blocked, you should be able to ask a librarian to unblock." -- JC

Which of course demands the absurdity that someone would know that a site they cannot see is legitimate.

Shared white lists/black lists is a common way to make such lists better, but are we likely to agree with the decisions of the librarians of Berkley, California or Provo, Utah?

I think having to ask a librarian to "unblock" a site is unreasonable -- and problematic. Unreasonable in that patrons should not have to suffer revealing their health concerns, interests or personal proclivities to librarians; problematic in that a librarian then becomes a referee. Joe Guarino has already suggested that staff be arrested because someone viewed illegal content on a library computer. Can you imagine the jeopardy an individual librarian would find himself in if he got an unblock request wrong?

Joe Guarino

A few points, and many of these issues have been discussed elsewhere:

1. The Cymphonix software, as currently utilized, will severely slow people who are browsing for information on breast cancer and STD's and the like. This will severely inconvenience and frustrate customers. Is this truly what the anti-filter camp wants?

2. Dave, we cannot make any presumptions about how many succesful attempts are being made to access porn. The data Cymphonix can provide is inadequate for that task. That exposes for all to see the foolishness of those who were demanding data before making further decisions. The data simply is not there.

3. The attempts at porn surfing as a percentage of attempts at browsing is a meaningless figure. If people are able to surf for porn, and are successful in doing so with some frequency, it is a major problem. Is is a profound moral error to allow minors to do this or to be exposed to it. It is a profound civic error to require taxpayers to pay for it.

4. There is no basis for trusting staff on this issue. Sandy Neerman did nearly nothing about this problem for an entire decade. We ought not feel as if we need to leave the current policy in place. In many cases, trusting staff means bowing before their progressive/relativistic sensibilities. The City Council needs to assert leadership, and take responsibility, and do the right thing.

5. There is no problem with expecting library staff to handle unblock requests. Other libraries around the country have figured out a way to address this issue, and I do not think it is beyond GPL's capability to do so also.

Roch101

"That exposes for all to see the foolishness of those who were demanding data before making further decisions." -- Joe

LOL! If those who asked for the data were foolish, what does that expose for all to see those who touted the data?

Joe, seriously, if it is a profound moral error to risk allowing minors seeing porn, how can you argue in favor of imperfect filters? The perfect solution is to disconnect the library computers. I don't know how you can argue that this is an matter of absolute morality then favor a relativist solution. Can you help me to understand the discrepancy?

cheripickr

It is remarkable the sudden shift of the fiasco/odious coalition’s focus back to manually reported incident report data now that the computer generated hit list didn't yield the desired results, which they now discount, not without some justification. But how many times have they "brandished" the 85% detection rate for filters as an inadequate to justify their use, yet they have no such lack of faith in the "catch" rate of human watchdog-dependent incident reports to detect more accurately the true incidence of porn-seeking computer activity on publicly-funded computers. Do they seriously think it would even come close to the magic 85% number which is still not good enough when it comes to filters?. Or I guess they figure that what we don't know won't hurt us, right?

Dave, I'm surprised that you would gloss right over that variable in your fourth point above and move straight to heralding the marked improvement based on such a crudely insensitive measure. Even more flawed is using the same understating figures to compare performance of throttling software to no antiporn program at all.

If you take the number of incidence reports over the last 6 months that the throttles have been in place and compare them to the previous term when there was NO anti-porn software in place, What did you expect to find? Of course there will be a drop in incident reports, compared to when there was no mechanism in place, and everyone can say “See, it’s working!” If you do not measure the same performance with filters in place (which are already installed and ready to go), the data is meaningless for comparing the relative performance of throttling vs blocking, which is what we are talking about here, isn’t it?
Don’t we all sort of concede that throttling is probably more effective than nothing?

I don’t know how this applies to the study of economics, but the deliberate avoidance of head-to-head comparison is practiced routinely in the pharmaceutical industry and everyone in medicine knows it’s a joke. When a new drug is introduced it is rarely compared in trials to comparable existing drugs to see if it has any advantage over them, but to a control group of patients given a placebo (sugar pill). This sets a very low bar for establishing effectiveness. The drug company has little incentive to compare its new drug to existing ones because if it does not show a particular advantage over an established (especially off-patent, cheaper) drug it will be unlikely to be approved. With so much already invested in the drug, why would they voluntarily expose themselves to that risk?
In the library scenario, the study period prior to the institution of throttling is analogous to the sugar pill. This is the “control” group, so to speak. If there is no true comparison group (filtering) then no meaningful comparison of throttling to filtering can be made. I think that those who have been so hell bent on resisting filters similarly have invested too much into their opposition that they do not really want to see this take place. I hope I am proven wrong.

This is actually a beautifully simple set-up for a comparison study because of the simplicity of using the same already-installed program for both arms of the study. This eliminates all sorts of confounding variables, biases and other apples-to-oranges issues that often erode the quality and validity of the results. It would be easy to measure apples- to-apples both the percent of undesirable sites recognized (sensitivity) and the proportion of desired sites erroneously detected (specificity). You could also measure the respective number of “hits” and incidence reports in each group, number of requests to unblock (or unslow), etc. The library IT guy said they can control the program down to the level of the individual computer. Why not randomly assign all the computers to either filter or throttle mode and do a double-blinded prospective study, both running simultaneously over a 3 to 6 month period. It would greatly mitigate the potential for deliberate bias that is creating distrust on both sides of the issue. This is really the only right way to compare the two and should settle the matter once and for all.
Perhaps Kathy Hartsell knows if there would be any logistical considerations that would preclude such a study.

Dave Ribar

John:

You are right that there are issues with all of these numbers. The incident reports reflect actions that were both detected AND written up. The incidence reports surely reflect a lower bound on questionable activity. We don't even know if the incidence reports are comparable from one period to the next -- if library procedures changed, the definition of inappropriate sites changes, staff vigilence, etc. -- the reports become non-comparable.

We also don't know exactly how the 700 figure relates to inappropriate viewing. The figure depends on criteria that were provided to the Cymphonix software; errors one way or the other would lead to undercounts or overcounts of activities. As has been pointed out, computer-generated attempted hits don't equal page views, especially when they have come through a throttling system.

Assuming that the incidence reports are comparable over time, the library's procedures (which include more than the throttling software) have made a tremendous (80%) dent in the problem. While a direct comparison with filtering software would be informative, it seems safe to assume that the margin for improvement is the remaining 20 percent.

Two educated guesses are that there is tremendous overlap in what the throttling software and filtering software would be able to accomplish and that there are also overlaps in what they would miss. Based on that, we should expect some further reduction in inappropriate downloads, but perhaps not a whole lot relative to the throttling software. I think (and may be wrong) that we are all in agreement that filtering will not be 100 percent effective. The throttling software has set a pretty high bar at 80 percent.

A full-blown comparison study would be more difficult than first appears. Filtering some computers and throttling others would not be the same as applying these approaches to all computers. In technical-speak, you want to conduct saturation experiments so that you can look at general equilibrium effects (for example, will library staff become more careless about this issue if filters are used; will patrons avoid certain libraries altogether). There are also possible Hawthorne effects because the staff are likely to know which method is being used and will certainly know that their behavior is under the microscope.

cheripickr

Dave, I know this is mainly an academic rather than practical exercise as far as the likelihood of this ever being done but if both the patrons and library staff (at least those "patrolling") are blinded to which computers were blocked and which were slowed wouldn't that mitigate most intentional bias and wouldn't the Hawthorne effect apply equally to both arms? It's got to be more valid than comparing something to nothing.

Actually I don't even think these studies are necessary but I don't seem to be having any luck getting anyone to be very specific about the purported advantages of slowing porn vs blocking the same porn with the same program, which is already bought and in place, either theoretically or practically. And at least it's good to inject some intellectual honesty into the discussion every now and then.

David Hoggard

"We also don't know exactly how the 700 figure relates to inappropriate viewing."

Nor do we know, as I mentioned over a Joe's, how many of the 700 'whatevers' were generated during the highly publicized period (7/23 - 8/24) by the curious wanting to see for themselves exactly what they could or couldn't get to - and how long it would take to get there.

I was tempted to go and try but haven't yet had the chance. I will though - thus skewing next month's report towards heavy attempts at viewing porn.

sean coon

hoggard, i thought you were a family man! ;-)

Spag

"Actually I don't even think these studies are necessary but I don't seem to be having any luck getting anyone to be very specific about the purported advantages of slowing porn vs blocking the same porn with the same program, which is already bought and in place, either theoretically or practically. And at least it's good to inject some intellectual honesty into the discussion every now and then."

Excellent point, John. The problem is that would be the most common sense, simplest solution and we are dealing with bureaucrats and the type of people who empower them along with a few basket cases.

Dave Ribar

John:

Blinding the library staff (and patrons) isn't a realistic test of the systems. The library uses a layered approach in which staff are an integral part. Also, staff need to be able to unblock access for adults upon request. That is less of an issue for throttling than for filtering.

Also, part of the "appeal" of filtering systems is that they will deter pervy surfers from entering the library in the first place. A mixed approach, where patrons don't know whether they would be throttled or blocked, wouldn't have the same deterence effect.

Regardless of the experiment that was run and how carefully it was run, we would still have to extrapolate from the results.

Sam,

The bureaucrats that you criticize still managed to generate an 80 percent reduction in reported inappropriate-viewing incidents. What insensitive, unresponsive morons they are. We certainly wouldn't want to empower workers like these.

Joe Guarino

Roch, you can argue in favor of imperfect filters because they represent a good faith effort to maximally reduce the visiting of porn sites that is taking place-- acknowledging that we are likely to continue having computers in the libraries. The band shaping option does not represent a good faith effort to maximally reduce these visits.

cheripickr

And off he runs, would love to chat longer, but must rush off to spread the gospel of the superiority of the status quo over placebo some more, no curiosity to test-drive the filters we've bought. Sigh. I give up. Their last hope--gone. I think I'm going to go talk to a wall now.

Jim Langer

Okay, boy am I getting weary of this. Let's see; if one or two or three guys are doing all 700 mostly unsuccessful attempts, rapidly, a few times a month, this isn't as big a deal. But if several dozen guys are this stupid, I think we should put some limits on allowable stupidity and let the library try to attain some of the worthier goals of educating the populace to get smarter. Pending some more numbers, I am not wholeheartedly but certainly more than ever before embracing the idea of using filters that can be shut off easily with legit requests for research, assuming the librarians are trained to accept some pretty intense art, for instance, but not anywhere near the little kids area. Parents should be made aware they should not allow small kids wandering about the library for this and many other safety concerns. But it does make a difference how many idiots we are dealing
with here.

Jim Langer

Dave's point (hadn't read it yet) about the period being publicized and the possibility of people purposely messing with the system to generate numbers is worth considering. Still, if we make the presumption that libraries and their computers are really there to try to make people smarter and not just for entertainment (and never for "adult entertainment", as the euphemism goes) I am getting myself some religion here: what's the Gulf Oil Spill term? "Bottom Kill"? "Cap and Tirade"?

Ron

Filters dont block out everything like a speaker said at the council meeting. The school where she worked had filters but when a young girl googled daffodils, a picture of a naked man in a field of daffodils popped up.

Jim Langer

I always suspected daffodils were phallic symbols. Those trumpety blossoms are extra-aggressive. Georgia O'Keeffe didn't corner the market on that imagery.

cheripickr

I am replacing all the deadbolts on my doors with new special doorknobs which delay opening the door for 10 minutes when you turn them. My thinking is that me or my family might accidentally lock ourselves out of the house and I would hate the inconvenience of having to go get another key just to preserve this imperfect system of locking doors. I figure a burglar, on the other hand, will get frustrated or nervous waiting that long and will go away.

JC

Roch - what I meant was if you were doing a legitimate search, and something came up blocked, you could go to the librarian and ask to have it unblocked. I would assume someone searching for porn wouldn't even ask.

The health issue seems a red herring to me. You should set up the filter so that anything from mayoclinic.com was safe to access (maybe outside the kid's section). It would not be difficult to differentiate between accredited medical sites and porn. Independent or alternative medicine sites might get blocked and then you'd have to ask, but again, the library doesn't stock Bubba's Home Remedies Monthly.

You would also have the option of submitting directly to the filter owners via the web. With more filter users, legitimate sites would quickly get unblocked.

Working at the library carries no expectation of privacy - it is a public library. That is the reason there aren't stricter filters! If it were private, it could be censored to any extent.

You could also use the existing safe-search functions in google and somehow set it up so that you couldn't take it off moderate filter.

Roch101

JC, here are some sites blocked by the Charlotte library's filter. They are not porn. Should a librarian unblock them if asked? Should a patron have to ask?


Spag

I also Googled "daffodils" and guess what? All I got were images of daffodils, and that was with the safe search off. I'm willing to bet you get the same results.

Mick

Cheri,
Who's point are you making there. The hamburgler would "most likely" get frustrated and go away. Wait, who's point am I making!

I am glad that the city folks are actually getting more info perhaps even all available data PRIOR to making a decision. Good luck to them trying to get teenagers and folks who lead a rather chaotic lifestyles to stop surfing for porn.

Joe Guarino

The burglar might sit and wait in the event he had nothing better to do. Or he might run to several houses to have them primed simultaneously.

You see, the porn browser at the library can open up a number of windows simultaneously, in the hopes of maximizing his "gain" after he waits for images to materialize.

Ed Cone

The multiple windows/multiple attempts by one person scenario further reduces the number of individuals trying to access porn.

A handful of people routinely trying to access porn might prove manageable by staff.

I think one of the selling propositions for burglar alarms is that they frustrate burglars so that they give up and move on.

glenwoodobserver

Well, of course we know that dogs are better at scaring of burglars than alarms...so maybe that's an option we haven't discussed. porn-sniffing dogs anyone?

cheripickr

Whoops

From someone who rarely offers his own solutions or ideas, I think you've just triggered a brilliant one here--a program that produces a loud alarm every time someone tries to access a porn site--seriously. Works for buildings and cars.

cheripickr

Whoops again.

@CP -- Hah! Now that's a great solution that I would think everyone could get behind. Maybe add a giant flashing neon sign that says "PERV!!!!" with a pointing arrow.

Posted by: eric | Aug 27, 2010 at 11:25 AM

Eric, I wouldn't even kick 'em out the, but make them sit next to the librarian's desk on a tall stool for at least an hour with their pointy perv hat on.

Ed Cone

In all seriousness, stiffer penalties for people accessing porn might be part of the solution.

cheripickr

Better choice of adjectives maybe?

Roch101

JC,

I wrote a comment to you that got snagged by this blog's filter. I had to contact Ed to override the filter and publish it, but by now there are several posts after it and it may do unnoticed by you. If you scroll up a few comments, you will see it now or click here.

Spag

"Stiffer penalties...porn" Rephrase that?

P.S. Y'all don't know burglars too well.

JC

Roch, while your sites may not meet the definition of porn, they are either inappropriate or very close to it. My question would be, would the library stock books or magazines on those topics? If those were deemed appropriate, asking for them is not an excessive step to take.

Roch101

Should a librarian unblock them if asked? Should a patron have to ask?

cheripickr

Elbow room must be getting very tight in the shrinking box of the fiasco/odious coalition's dogged resistance to library porn filters. I hope you guys used your deodorant.

"Why does it show multiple hits for a Web page I only went to one time?

The system counts each item loaded on a Web page as a hit so some pages may have several hits each time the page is loaded. With the release of v.5.1.1, the hit counter does not count each image as a hit, which dramatically reduces the number of hits per page, making it much more useful as an indicator of Web surfing activity.

15. How do I upgrade my Network Composer?

See the Upgrade Guide."

JC

If the pages are text only then they are probably OK. Asking for them, OR using a via-the-filter request, might be necessary for the first user. What to filter is a judgment for many things, so what I might or might not want to filter is not so much the question, but whether the filters would be used at all.

Roch101

JC, I feel like you are dancing around my question. I'll rephrase. I gave you examples of three sites that filters block. Do you think it would be appropriate for a filter at the Greensboro library to block those specific sites? If a filter blocked them, do you think a librarian should unblock any of THOSE sites if asked -- not some hypothetical sites, but any of those specific sites? Do you think a patron should have to ask?

Spag

Didn't they used to keep some books "behind the counter" at the library?

Ed Cone

I think GPL used to take the print edition Playboy many years ago, but someone will have to factcheck that.

The comments to this entry are closed.