Sunday, January 20, 2013

Zeno, in the modern context

It's common knowledge that the modern world knows things through computers. This is true from the most recent phenomena, of noses to smart phones parading as intelligent behavior, to the wide expanses of cosmology's modeling of the heavens and exploration of multi-universes as an explanation, of sorts. In between, we have IRS's use plus business computing, such as design, planning, and a number of other things.

So, where does one go to look at issues related to such knowledge? The ACM is a good start. Say, their Communications of the ACM. Then, we have a whole lot of other folks, such as IEEE, IJCAI, and such.

---

The following is motivated by a viewpoint, expressed by Phillip G. Armour, in the ACM. His article is titled: The Business of Software: How We build Things (in paper, slightly different on-line). There are two things to mention here, though the article ought to be read.

He uses Zeno in talking about what I had called Earned Value. I only used Zeno once (Fedaerated) in several posts on three blogs. Why? I had talked about this with my colleagues on many occasions. It seemed that referencing the guy was more useful in person as then one could get off on the peripatetic issues.

Zeno, Veritas et Falsitas
Why Zeno? He's the guy of the arrow. Or, as the joke goes, the mathematician who doesn't get the girl. So, Phillip asks: why do people guess that they're 95% (or some such number) complete on a task as if they're monotonically approaching, with no end in sight?

Phillip laughs it off. I don't as it was a regular occurrence as we tried to assess completion of a project with lots of people and oodles of modules. Nowadays, it's not an issue (say, with Zuck's stuff) as they can just push out system changes (with a recovery method, hopefully, to use if things go bad) without regard to testing status. This is not true for other parts of the business world, say like the 787 (even a most-specified test plan will still leave room for judgment calls -- we'll get to that).

---

Phillip's use gets me thinking, that I need to bring the topic forward, again.

First, though, a useful exercise would be to gather all of the posts, for each of the blogs, that dealt with the subject of earned value. For each of the blogs, I have a list of posts that include the term. Then, I provide a list of a few of the important posts and the count of posts with the term.

 Fedaerated (18)     7'oops7 (41)     Truth Engineering (20)

Now, for this blog, all but a couple of the posts were in 2009 and before. That sort of indicates the shift to looking at finance. Engineering worries about things like this. Finance seems to have this short-term view of the content of the current day's pocket. That is idiotic, pure and simple.

So, we'll bring this subject up to date and relate it, as it ought to be, to fair value.

A sampling of posts follows:
  • Earned value II (Jan 26, 2009) -- Looked at some of the factors that relate to the hardness of the problem. 
  • Earned value (Jan 23, 2009) -- Two years in, decided to tackle the definition'l issue that had been ignored from the start in 2007.  
  • Here we go again (Jun 24, 2009) -- Tied earned value with its compatriots in the context of planning and managing a project. 
  • Middle and out (Jan 29, 2008) -- Management, like Jamie, like top-down in their illusion of being in control (Hey, can they even control themselves? That's more of an open issue than you might believe (or want to accept).) 
  • Carts need horses (Nov 2, 2007) -- It looked for a long time as if the tail was wagging the dog. We'll not go into details, at this time, but the whole thing relates to the above bullets (know where you're going, how far you are, etc.), especially the middle out. 
There are a lot more posts to look. We'll get back to that.

---

Phillip used some mathematics to show the problems related to knowing where you were with a project (the managers, like kids, say: are we there, yet?). Nice article.


Remarks:

01/24/2013 -- Article from New York -- The Digital Skeptic: Dreamliner Brings iPhone 'Reliability' to the Skies. With something as complex as the 787, how many operational issues can be expected to crop up (at 35K feet -- rhetorical answer? A bunch.)? A student paper at Uniz of Az looks at 787 outsourcing (see Mahmoodi). Boeing Tech Fellow's cautionary note, on outsourcing (2001 - there was quite a bit of discussion at the 2004 conference in LA, to boot, about matters of technical design -- also, whistle blower?). Related, 787 blog.

01/21/2013 -- WSJ has a nice recap from the beginning. Look at how management kept promising that they would hit a milestone when most figured that they would not. Then, the guy would switch gears. In July 2007, most knew that the shell was empty, yet the claim was for delivery in May 2008 (a mere 11 months later). Then, in October 2007, the message was that they would fly in March 2008. It didn't fly until December 2009. Finally, things quieted. It was obvious that engineers were taking the lead. 

Modified: 01/24/2013

No comments: