They say everyone chooses their own kind of hell. For shipping, it just might be the rules.
By Ryan Skinner (email)
It's total sacrilege, but I'm going to say it. Maybe DNV f*cked up.
Now, I have all the respect in the world for the guys over at Høvik. They're brilliant. Even if I kind of panned their Quantum thingy, they're still leagues ahead of most outfits in this business. They're what passes for bleeding edge in (let's be frank) an oftentimes dull industry.
But consider this ECDIS business. Over and over again, people pushing ECDIS cite the DNV studies, which demonstrated that mandatory ECDIS would reduce groundings by some dizzyingly high percentage. It was more or less on the back of these studies that the IMO went whole-hog and made these chart computers mandatory on all SOLAS ships PDQ.
The rules logic goes awry on the graph where x = some not-too-distant point in the future and y = a metric tonne of rulebooks. Suddenly, a rule requiring every master to have his mother on the bridge to hold his hand, doesn't work. The sheer weight of rules sinks the ship.
Now, I'm a big fan of ECDIS. It's ingenious like the GPS in a car, which I don't use but love on rentals. And perhaps that's the difference. No one mandates a GPS. No one argues that every GPS operates a little differently. No one moans about the need to update a GPS now and then. No one needs to take any special training to use one.
Would it not be better for regulators to get far out of the way, and let technology adoption take its own natural path? If ECDIS is just an example, is the profusion of rules making shipping safer, or more dangerous?
In order to apply a little science to this question, I've created/found a few graphs that help illustrate the problem. The first illustrates the expanding stack of rules that apply to shipping operations:
As we can see from this well-documented study, the amount and detail of rules-making has grown out of control. This chart is anecdotal, yes, but anecdotal evidence from shipping companies suggest it's approximately true. Is it having a beneficial effect on safety? Here are two charts taken at (relative) random regarding maritime casualties.
Granted, increasing volume of accidents and value lost may easily stem from a much larger world fleet. Nonetheless, it doesn't seem as if more rules = more safety. Is it then wise to create, yes, even more rules?
When considering mandatory ECDIS, perhaps it would be best to look at accident figures for high-speed craft (ferries and the like). ECDIS has been mandatory on these for many years; has it resulted in fewer incidents? Anyone have the data on that?
Both shipowners and officers are rightly grumpy about ECDIS mandation. Many people are simply PO'd that this thing's being pushed down their throats. Officers hear that they will get dual ECDIS and paper will be gone, and they worry. What about viruses? What about system crashes? Others rejoice.
I know the ECDIS fan-boy crowd is going to come down on me hard for this. They argue (in many cases, rightly) that without mandation the merchant fleet simply won't do ECDIS. Why? Because it's costly. And that may be true. But costs come down. Applications increase. And pretty soon ECDIS would be mandatory for entirely different reasons, not legal ones.
You can legislate yourself into a thicket, and I fear this may be happening more and more often in shipping and at the IMO.
Provocative as ever, Ryan! I do enjoy a good ShipCrunch blog entry :)
As one of the few navigators in the world who has sailed on a genuine paperless bridge, I guess I'm what one would call an "ECDIS fan-boy", but I agree totally with your point about legislation - there is too much of it and it is unclear what applies to who. I draw the attention of others to www.ecdisregs.com - a project that will probably never be "finished" because the flag-state legislation just keeps shifting and updating!
Sadly, I don't think there's a realistic solution, and there's no doubting that with the right training and implementation, ECDIS makes for a safer bridge with improved spatial awareness and more spare capacity.
Posted by: Mjspuk | February 01, 2011 at 11:00 AM
Good post. A point that is often overlooked however is that paper charts can fail. Without an accurate position and track-line a paper chart is of little use. Not infrequently mates miss-plot the position or, in traffic, neglect the plot altogether.
Posted by: Ken E. Beck | February 04, 2011 at 03:29 PM
You are right that a great deal of regulatory justification has been placed on the DNV study. Remarkable, given that it did not use real navigators, but a neural network model of a navigator. So, there has been a considerable load placed on the validity of such a model.
Posted by: BrianSJ | February 07, 2011 at 11:32 AM
Brian, can you explain those of us who don't know what a neural network model is? And what the drawbacks are, what the risks to validity are?
Posted by: Ryan | February 07, 2011 at 11:55 AM
First a correction; it is a Bayesian network model. It has 67 nodes. The model has had input from experts, but so far as I can see, no validation against human performance e.g. in a simulator. It is easy to build plausible models and to believe in them, but plausible models are often wrong and need validation.
I have been involved with projects where a close examination of the human decision making involved invalidated the whole Bayesian approach. If a Bayesian approach is in fact appropriate, there are difficult issues of prior probabilities, conditional independence, uncertainty propagation to check. The adaptive nature of human behaviour may be accurately captured or it may not.
The wikipedia article on Bayesian networks is not for the beginner (surprisingly), and so far I have not found an article that is not for the enthusiast.
Computer models have a chequered history in public policy; perhaps the worst example in the UK was the modelling of foot and mouth disease. As a general principle, it is wise to be sceptical about them!
Posted by: BrianSJ | February 07, 2011 at 02:45 PM
Ryan,
Even though electronic charts are my bread and butter I tend to agree with you. Sadly.
The light marine industry from where digital charts originated will not touch the official ENCs with a barge pole.
Industries are mostly driven by the so called voice of the customer. After all he is the guy who is paying the bills. Right?
The way ECDIS and ENCs are being mandated in commercial shipping is totally wrong. Charts are produced by HOs who do not talk to the customer because they are government. Ships will be forced to use them or else face costly detentions.
Who enforces these regulations? It's the flag state and the port state control. Ship-owners can side-step strict national laws by registering their ships with the FOCs. But they cannot elude the Port State Controllers.
Not many people are aware of the menace the PSCs can become to ships. PSC inspectors who enforce the IMO regulations ostensibly to uphold their noble objective of safer seas and cleaner oceans sometimes add their own little objective. To derive business for the local ship-repair centers.
Masters are scared that mandatory ECDIS will give yet another weapon to these PSCs. Whether they like it or not ships will have to buy the ECDIS and their poorly designed ENCs.
Posted by: Raj | February 21, 2011 at 06:05 PM
Now that ECIDS has gone mandatory, owners are going to go for the cheapest model that satisfies the requirements. If it had been a "good to have" thing on the bridge, then owners really serious about safe navigation wold have installed it and that too a practically operable one and not something that requires x hours of manual reading to operate and x(squared) hours of manual reading to carry out a weekly update
Posted by: Archie | March 15, 2011 at 12:28 AM
Great analytical post! I like too much. It saws in the chart that holy mother of god is highest. I like to read such analytical post.
Posted by: תאורה לחדרי ילדים | October 01, 2011 at 10:41 PM
C'est vraiment la bonne fa?on, vous discuter de ce genre de sujet.
Good job.
Posted by: tee shirt pas cher | March 28, 2012 at 11:39 AM
Freight rates and shipping costs are low at the moment and a lot of cargo ship and ferries are at present for sale at quite reasonable prices. Great opportunity for renewing the cargo fleet and taking advantage of the low shipping costs.
Posted by: MARETEC Shipbrokers | April 03, 2012 at 03:08 PM
Der Wikipedia-Artikel über Bayes-Netzwerken ist nicht für den Anfänger (überraschend), und bisher habe ich noch nicht einen Artikel, der nicht für den Enthusiasten gefunden.
Computer-Modelle haben eine wechselvolle Geschichte in der öffentlichen Politik, vielleicht das schlimmste Beispiel in Großbritannien war die Modellierung von Maul-und Klauenseuche. Als allgemeiner Grundsatz gilt, ist es ratsam, skeptisch zu sein über sie!
Posted by: abercrombie fitch madrid | May 28, 2012 at 10:19 AM