One-offs are almost always trial-runs for “doing it right” later.  They are quick because we aren’t too worried about precision, nor about documenting what we did.  We’ll use an approximate conversion factor, and maybe just make up provisional values for parameters we don’t have in front of us right now.  Nothing wrong with that, it has stood us in good stead for ages and helps us get a handle on things.

But what about the “dirty” part?  Obviously, doing it quick now, and right later, spells extra work.  Being light on precision and documentation means we’ll have to figure it all out again.  Or worse, we might find that our quick-and-dirty results find there way into downstream work  (Maybe we’ll never even know when that happens).  In the end, we’re pretty sure that one-off’s are dirty, and wonder if they are really quick after all.

The tradeoff is between meticulous care and agility, and we have always needed to opt for agility even it adds overall cost and uncertainty downstream where we need to verify and validate our way to engineering quality.  The goal of truenumbers technology, simply put, is to break this tradeoff by making it easier to do “quick-but-clean” one-offs than it is to do the old “quick-and-dirty”.

Jim Hendler’s famous semantic web mantra “a little bit of semantics goes a long way” applies generally to the problem of information dilution in our desktop engineering environment. The term “semantic desktop” came on the scene regarding improved personal information management but it is a deep and general problem for all kinds of information, especially in engineering.

Much of PLM’s value seems to be simply aggregating and making available all of the documents associated with engineering projects, but very little progress has been made in addressing the lossiness of the communications those documents contain. We see this as an unneccesssary deficit of our desktop workplace. There is really no reason that when information is documented and communicated by people that precision and context have to be jettisoned. It is “simply” a matter of evolving our desktop environment toward the preservation of meaning and context as we move it from points of origin to documents, and from documents into our CAE and EDA tools. Truenumbers are one step in that direction, we expect that this evolution toward adding little bits of semantics will indeed go a long way

We’ve written about the limitations of the SI unit system before.  One kind of number units can’t touch are known as dimensionless numbers.  A 45° angle, 27%, π, and yes, 284,947^th member, are all dimensionless, but obviously not colorless, numbers.  Mathematical physics says that they need to stay dimensionless, and that there is little that can be said about them unit-wise, because they are not lengths, times, or amounts of electrical current.  But PayPal tells us instantly when a credit-card number we entered is not one, so why can’t we offer engineers a little help in telling an angle from a LinkedIn member?  OK, an engineer will be more interested in distinguishing an angular velocity from a frequency, a common cause of confusion in engineering calculations. 

In the dimensionless world, reckoning angles and counting members are pretty much the same as measuring angular velocity and frequency.  As it happens, software can in fact help distinguish them (one has to be careful not to get physics mixed up in it or numerical hell beaks loose) and this approach to categorizing dimensionless quantities will be part of our upcoming truenumbers product release later this spring.   We look forward to telling you more about it.  We’re thinking about audaciously calling our approach to representing engineering numbers SI++.  Let us know what you think, I hope NIST doesn’t mind!

It is a measure of importance to humanity when something gets its own units of measure designed to quantify its effect on human life.  Hurrcanes, earthquakes and nuclear radiation are among the few things in our world that merit such treatment.  We’ve become familiar with the Safir-Simpson scale of 1 to 5 for the destructive power of hurricanes, and the Richter scale for earthquakes, but the Sievert, used to measure radiation dosage has thrown us for a loop. 

The problem lies with the Sievert itself.  While most of us don’t know a lot about radiation, we do know that it involves little particles streaming at us and messing up our molcules as they zoom through.    Thus, the natural measure we expect to use would reflect the intensity of that stream.  Like the sun and sunburn, if its really bright, you don’t stay out long.  The Sievert is instead a measurement of how much radiation you have already absorbed.  It is a measure of dosage.  So if you want a measure of intensity, that would be Sieverts per hour.   The following sound-bite from a news article:

“The radioactive radiation at Japanese nuclear power plant Fukushima 1 reached new record levels. It is the value of 1000 mSv (1 sievert) was measured, says government spokesman Edano.”

doesn’t tell us what we want it to.  Does it mean that a person standing near the plant would by now have received 1 Sv dose?  More likely it means a person standing there now would receive 1Sv every hour.  When you hear “it’s no more than a chest X-ray”, that has the same problem.  Is the amount hitting a person near the plant like a chest X-ray every hour?  Every minute?  Like one long contiuous chest X-ray as long as you are standing there?

For us readers thousands of miles away, it really doesn’t matter.  I guess the pace of today’s reporting precludes running such facts by the science editor, and I can’t help thinking that this is only a reflection on the editorial process news lacks in general.  But from the Numerator’s point of view, lets at least get the units right!

Where we take issue is this: now that computers are our desktops and tools, we no longer need to directly use the SI system whenever we measure something.   When an American machinist gauges a part, he certainly will use a micrometer that is calibrated in inches.  If he writes down 3.022″, we understand that this is measured to the accuracy of that micrometer. Converting this to SI, we end up with 0.0767588 m which appears to be a more accurate measurement.  If diligent, we would present the result rounded to a precision similar to that of the original measurement, but you get the idea:  something is lost in the translation. 

Physics, the world we live in, and the tools we use, measure things in units appropriate to the situation, and we believe that such measurements are primary data that should not be lost.   In our view, the SI should, like the floating point standard inside our computers, be the gold standard for measurement, but not its only expression.  In that spirit, truenumbers  preserve whatever units of measure a number is originally expressed in, as well as its SI representation.  We can think of this approach to units as operational, because it supports the operations performed by technical people, using the SI as means for accurate internal representation.

When you pass a sign that says “55 mph”, that’s a number.  A number with units we commonly understand to measure velocity.  Furthermore, we understand that it refers to us: our cars moving past on the road.   More precisely, it is saying “the maximum allowed speed of your car = 55 mph” and because it’s on the highway, we assume it was put there by proper authorities, adding another fact to what that sign implies.

Using our computers in our daily work, we encounter and produce numbers all the time.   As with the speed limit sign, there is a lot of information surrounding these numbers that make them meaningful, and which isn’t always expressed clearly, or even expressed at all.    We might have to read a whole paragraph, or scan for heading of rows and columns in spreadsheets to interpret a string of digits we’re looking at.  Even then we often have questions about a number that aren’t easily answered.  Is it the right one?  Who came up with it?  Is it still the right number?

Our simple notion is that by making a more formal package, kind of like a URL, that still looks like a number, can be used almost anywhere, but that carries with it everything we need to understand a number clearly, we can make numbers much more useful and powerful.

We call these “truenumbers”.  Making numbers better in this way is a small thing and a simple idea in some ways, but on the other hand, because numbers are so important to our work and daily communication, a little bit of improvement can go a very long way.

“It’s called Dark Room and its a minimalist fullscreen word processor which forces you to focus on the writing process and nothing else.”

Turning your screen into a 1970′s ”glass typewriter” does indeed improve concentration.  The pace and density of personal multi-tasking across desktop and mobile platforms is bound to cause some reflection and backlash, driven not by tech-nostalgia but by genuine need of a respite.  The illusion that your computer is doing only one thing helps you to do the same, and we find that it feels nice.   On the hypermodern end of the timeline, today’s electronic communications are compact, and tend to be free of the need to emulate printed media.  Fonts and margins don’t mean much in a Tweet or an email, and the WYSIWYG document metaphor our dual-cores strain to run realtime begins finally to wear on us when we’re feeling task-oriented.   Our post about computers as language machines sympathizes with Darkroom’s focus on words as enough, maybe because people are language machines themselves.  Buttons and menus are “easier than remembering and typing commands” but move us away from the use of language, so maybe Darkroom is a step forward more than it is backward.

But searching on 3.14 to find PI , or building the world’s largest data bank of numbers, while they might concievably happen on a site like ours, are not really why we’re here.   Truenumbers are more akin to RSS than they are to Bing or Alpha.  As RSS does for a news item, a truenumber is  intended to provide just enough tagging to make numerical facts more amenable to processing than are raw text and digits. 

A numberspace is a special repository and associated services for creating, storing, organizing and using truenumbers.  We really hope that organizations from publishers to academic departments to engineering enterprises deploy numberspaces for keeping, and keeping track of the many numbers that their work entails and produces. 

Truenumbers and numberspaces help find and keep track on what versions of which numbers are floating around and being used, correctly or incorrectly.  Just formatting the darn things properly, and having the right, validated units are valuable benefits over plain-text numbers. So, while we are excited to be a part of the new wave of interest in new kinds of search, and the role of numbers in the semantic web, the marketplace will judge us by how well we improve engineering and other number-rich activities, one number at a time.

 Today, HTML, XML, RSS, the semantic web and the merging of applications and content seem to be moving us back toward use of language in computers.  The tendency for use to use smaller more focused media like blogs, wikis, SMS and Twitter makes tags and other metadata more pervasive, and the computer more of a partner in our communications than it was when it merely filed traditional monolithic  documents.  In short, computers as language machines are coming back, and I think we will see many useful things result.