Training 2014/01/30 #006

Knife Defense

I greatly enjoy this class. We seem to do it on an annual basis or so.

What do you do when someone comes at you with a knife?

It's probably not going to happen but it's something we all worry about so it's nice to have a strategy.

1. Keep some distance. Yell "He's got a knife!" For good measure, tell the attacker what to do with the knife: "Drop the knife! Put it on the ground!"
2. Keep your hands up. Imagine that your hands are blades. Keep your fingers together and present the backs of your hand to your attacker to protect your wrists.

There are a lot of different attacks with a knife but we only address the most basic -- the straight forward lunge at your gut.

3. Block the strike by making a "Y" shape with your thumbs on each hand, putting your hands together, and trapping the wrist of their knife hand. We started just working on the block. Using two hands and the Y shape is important because you really don't want to miss and expose your middle.

We then moved on to the throw, which is basically shiho nage ni.

4. Grab the attacking wrist with both hands. You will probably naturally step into your attacker.
5. With your back foot, pivot to the outside (i.e., 95-degree pivot), bringing their wrist up to your forehead.
6. Pass under their arm.

It's shiho nage. Basically, you and your attacker are standing side-by-side facing different directions. You're holding their knife wrist and the knife is floating somewhere in the space between your attacker's ear and your ear. Hopefully, you've started to break their balance because their knife-side elbow is up over their shoulder and they're moving backwards.

7. Push forward to drop them. Or down. Just don't let go of that knife hand!
8. When they go down, keep control of the knife hand and get the knife away from them. You can basically put them into knee on belly, wrist lock them, choke them, whatever.

We sped the activity up a bit and put some resistance into it. You quickly realize that the technique doesn't always work the way you want it to. For example, you might not drop your opponent but you can still get the knife away by locking the wrist. Then you're just into ground fighting!

Tanto dori is always interesting because everything gets a lot more scrambly. The preset Aikido moves breakdown (at my lowly level of skill) but the BJJ training gives you the ability to react to the positioning. I also found it interesting that my shiho nage seems to be much better when my partner is holding a knife. I'm far less worried about choreography and, in general, I'm lower and my hips are in a far better position. The pin seemed more natural too. I was training with a pretty big guy so I was much more inclined to control his body using something like kesa gatame and it just felt more natural.

Rolling went well. My closed guard game is getting better. I think I could benefit from more visualization in my sweeps. I hit some sloppy pendulum sweeps but stalled out in the scissor sweeps. I had some success with establishing more of a z-guard with decent grips and then pushing out a knee to get the sweep. It was ugly.

On the submission side of things, I got halfway through a triangle... which is farther than I normally get. I'll take it as a sign of progress!

Training 2014/01/27 #005

Drills and skills

We worked on knee on belly today. I honestly spent most of my drill time working with a heavy bag because my timing was so wretched! Basically, it was knee-on-belly transitions: right to left, step around the head, and drop into mount.

We also played an interesting game. It was 2 on 1. Two attackers attempt to move one defender about 20 feet across a line on the mats. The defender can submit the attackers but the attackers can't attack the defender. The drill is really about positioning.

You quickly learn a few things as the attacker:

1. Keep your grips legal! An inside cuff grip is dangerous when someone can kick out.
2. Get a leg and an arm. If each attacker can get a leg and an arm, the defender is useless.
3. Control and move. Another strategy is for one of the attackers to simply control the body of the defender via mount, guard, etc. The other attacker can then grab a cuff and haul them both across the line.

There is also a few lessons to be had as the defender:

1. Try to keep both of your defenders close. If they get in each other's way they really can't coordinate the attack.
2. Escape decisively. If you're going to move, don't pause or hesitate. You basically have two sets of grips to break and then two tackles to avoid!

Overall, the exercise was quite entertaining and absolutely exhausting. It was very interesting to explore the different interactions of weight and position.

Process definition to initiate document management

It seems that processes are a key component of taxonomy construction and enablement. Generally speaking, we have four different types of processes: operating processes, support processes, management processes, and control processes.

Porter explored these issues in "Competitive advantage" under the general rubric of "value chain".

There are a few different standard process models including: APQC-PCF, SCC-SCOR, Canadian Municipal Reference Model, Exploration and Mining Business Process Reference Model, VCG-VRM, eTOM, and ACORD. In reviewing processes, one can identify a variety of different facets:

• does it exist with the enterprise?
• who is the owner (person or department)?
• what are the support systems?

Lehmann introduces some other categories (slide 29), including:

• value (core, secondary, low, outsourced, NA)
• asset type (operating, support, management, exception control, resolution control)
• use (routine, periodic, occasional, not used, unknown, new)
• complexity (simple, low, moderate, complex, highly complex)
• discipline (formal, informal, rely on 3rd party)
• automation (manual, semi-automated, fully-automated, rely on third party)
• documentation (detailed, satisfactory, unsatisfactory, undocumented, rely on third party)
• training (sufficient, insufficient, not available, not necessary, rely on third party)
• effectiveness (very effective, effective, adequate, not very effective, not at all effective)
• efficiency (very efficient, efficient, adequate, not very efficient, not at all efficient)
• ACTION (improve, consolidate, eliminate, benchmark, re-evaluate, design, outsource, insource, none)
• TIMING (immediately, near-term, long-term, unspecified)

This kind of exercise could be valuable during an early phase of a document management initiative. For example, IT could review these processes while enabling document management as a service. It would also be a valuable exercise when engaging with new business units.

With each business unit we could expand the discussion. Lambe suggests a variety of different facets that could be amenable to knowledge management, for example:

From Ranganathan:

• People and organizations (a controlled schedule)
• Things and parts of things
• Activity cycles
• Locations
• Time or sequence
• Subject matter (generally for a well defined discipline)

From Rosenfeld and Morville:

• Topic
• Product
• Document type
• Audience
• Geography
• Price

From Tiwana's Knowledge management toolkit:

• Activities
• Domain/subject matter
• Form (document, file, tacit, etc.)
• Type (document type)
• Products and services
• Time
• Location

Personally, I like to understand a few things from key business units:

• Document types they use
• Materials they use
• Activities or business processes
• How they describe locations (often linked to processes)
• How they describe times (also linked to processes)

So I guess I'm actually pretty close to Lambe on this one!

Defensible Disposition: An early step in controlling documents

Rich Medina over at Doculabs has been kind enough to provide some guidance on defensible disposition, particularly as it applies to file share cleanups. His approach is very similar to the one that I advocate but I want to provide some additional detail.

Basically, to take control of information you want to break it into day forward activities and historical review.

For day forward, make sure everyone has a personal network drive; make shared drives read only for relevant departments; use ECM -- non-records for 3 years; non-records with value for 7 years; official records, as per retention schedule. Social? No document storage with appropriate retention (e.g., 3 years).

Dealing with historical information requires some attention to your approach for defensible disposal. In particular, you must identify:
1. regulatory retention reqs.
2. hold retention reqs.
3. business retention reqs.
4. cost implications

It's always a tradeoff between requirements and cost, particularly for information that is low-value and low-risk. It's important that the defensible procedure process is documented. Courts don't expect you to be perfect but they do expect some rigour.

Rich provides some interesting numbers noting that manual classification requires about 10 seconds of premium labour per document while automatic classification tools require less than one second, work around the clock, and can route the 5-10% of classifiable docs to offshore labour!

When reviewing files consider:
1. Environmental attributes around the file (location, department, etc.) Use access controls for ownership; location via file path.
2. File attributes about the file (file type, age, author). duplicates, file types, document properties, file name character strings.
3. Attributes within the file (keywords, character strings, word proximity, word density). Specific strings (e.g., client names, works like "guarantee", or "privileged", or CCN

Audits: unnecessary file types 13-15%; duplicates 15-20%; near duplicates 9-30%. Files with PII 10-16%; keywords 3-5%; 10 years or older 7-11%; accessed within the last 18 months 25-35%.

The Janus Nature of Document Management

We have a challenge with document management. In most enterprises it can be two different things: a service, or a project.

A service is something that has to be delivered as a basic business enabler (think help-desk, dial-tone, etc.) There is really no business case for service provision; it's simply a cost of doing business. The only case to be made is to assess the Value-at-Risk for major initiatives (e.g., process improvement, etc.). Document management is different from other investments in that it is unlikely to reduce operational costs. For example, one could replace a phone system based on the costs of the existing system or the possibility of it failing. These criteria aren't applicable to document management.

The other side of document management is projects. There is a clear business case to do something but by definition that project effects only one business unit or business process. The danger is that the strategy will ultimately fragment the business due to our inability to bring common information facets together.

Basically, you need to define document management as a service before you can really initiate projects. Doculabs has played with this idea in the past. In particular, it provides an interesting operational model for what to put in place before expanding into actual process improvement initiatives.

This same white paper provides a compelling breakdown of the ways in which ECM deployments fail: lack of an implementation approach, poor adoption, over customization, user dissatisfaction, poor problem diagnosis, and a lack of metrics and measurement.

Kipple Corollary and Sock Entropy

I've recently written on kipple and its application to modern life, particularly electronic file shares. There is, however, a corollary: sock entropy.

Basically, the order of matched socks dissipates with time due to the loss of individual socks. Note that this phenomenon does not result in an empty sock drawer as one would expect. The space is inevitably occupied by other things: old mobile phones, thai-boxing hand wraps, extra shirt buttons, promotional bottle openers, bow ties, etc.

I have one strategy for battling sock entropy. I only buy one kind of sock so I don't have to worry about making pairs. I simply have to find two socks. Even this task, however, has become difficult due to sock entropy and the drawer itself is no less empty.

File Share Triage: 1

I recently returned from a consulting engagement in which we were attempting to address file share issues, primarily through the introduction of document management. Some thoughts:

1. Why file shares are a mess

1.1 Too much stuff

There is too much stuff for a file share to contain. Humans are, by nature, hoarders. We are the product of eons of evolution in which stuff was very hard to come by so it made sense to hold to what little we could. This behavior has largely been condemned. Dante's fourth circle of Hell, for example, is dedicated to Hoarders and Wasters. Hoarding disorder is now included in the DSM. But where could it have come from?

We have to recognize that stuff has never been cheaper than it is right now. Consider that in the dark ages, a shirt would have cost the equivalent of $3500:

"So, 7 hours for sewing, 72 for weaving, 400 for spinning, or 479 hours total to make one shirt. At minimum wage - $7.25 an hour - that shirt would cost $3,472.75. And that's just a standard shirt. And that's not counting the work that goes into raising sheep or growing cotton and then making the fiber fit for weaving. Or making the thread for the sewing. And you'd still need pants (tights or breeches) or a skirt, a bodice or vest, a jacket or cloak, stockings, and, if at all possible, but a rare luxury, shoes."

The innovations of the industrial revolution addressed some of these issues but the problem persisted into modern times:

"Even in 1800, a farm woman would be lucky to own three dresses - one for best and the other two for daily living. Heck, my mother, in 1930, went to college with that exact number of dresses to her name... This is why old clothing is rare: even the wealthy passed their old clothes on to the next generation or the poorer classes. The poor wore theirs until it could be worn no more, and then it was cut down for their children, and then used for rags of all kinds, and then, finally, sold to the rag and bone man who would transport it off to be made into (among other things) paper."

Things are very different now. A recent BBC magazine piece gives us some numbers:

"Consider the average British woman. According to various surveys, she buys 59 items of clothing each year, she has twice as many things in her wardrobe today as she did in 1980, and she has 22 things in there she has never worn."

It would have been economically impossible to collect clothing like this for most of human history.


This situation is particularly acute with electronic artifacts since they are non-rivalrous (i.e., the use of the object by one individual does not limit another from similarly using and benefiting the object). Furthermore, the cost of storing and managing these objects is ridiculously low. Robert Hooke, for example, paid 35 shillings for Bockler's Theatrum machinarum, or about 2% of his gross income in the late 1600s! Regardless of the accounting, that cost is certainly more expensive than the limitless reproduction we see on file shares.

1.2 Too little organization

File shares are almost always a problem due to a lack of organization. Parts of a file share might be organized, but not all of it. We see a variety of different organizational structures that persist for various lengths of times. Individual users will sort document by file types (e.g., "forms"), or by time, or -- very occasionally -- by process. These different organizational structures might work for a period of time by they are eventually superseded.

A file share is essentially a commons. A variety of different people use its resources and contribute to its ultimate usage and success. But why would some store documents at all? Why do people keep stuff?

Whittaker's ARIST review on personal information management summarizes the literature on personal information keeping. Basically, people keep a lot of information (and have difficulty in culling their collections) because they can't accurately identify the potential value of that information. Furthermore, there is little visible cost associated with keeping it.

That's not to say that there isn't a cost. The costs, however, are typically bourne by someone else (e.g., IT) or represent something non-quantifiable (e.g., ediscovery costs). A described by Prospect Theory, a knowledge worker is driven by loss-aversion so is unwilling to get rid of documents and information.

This "preservation bias" is really unaffected by organizational method. Pilers are generally as inefficient as filers. In reality, most retained information is never accessed again. Personal data generally contain 49% unique data (notes, completely projects, legal documents); 15% completely unread documents; and another 36% that represent copies of publicly available information.

People keep these doucments for a variety of reasons, typically related to:

• Availability -- keeping things on hand
• Reminding -- prompting action associated with a document
• Trust -- addressing worry that the document might get lost or misplaced
• Sentimentality