|
Patch reliability is unclear. Unless you have an immediate, pressing need to install a specific patch, don't do it. |
i am having a ‘funny’ experience with the Scripting.Dicitionary. I have an array of 3 S.Ds, each of which is of the form Keys(string), Items(Long).
The first SD is 1600 elements, the 2nd 8700, the 3rd 14000. Accessing the elements of the smallest SD takes reasonable time, but for the larger ones, the access time is _significant_.
Anyone point me in the direction of an explanation / documentation?
I *am* an engineer and that’s why I talk like this. 
‘O()’, pronounced Big O Notation, is a way to rank various algorithms.
Given an algorithm ‘x’ with a collection of data of size N, possible measures would be expressed:
O(N) = constant e.g., array access which is independent of N
O(N) = log N, e.g. binary lookup (log is to base 2)
O(N) = N linear e.g. start at 0 and access each element in turn till you find it
O(N) = N^2 e.g. the distance formula. This kind of algorithm gets expensive real fast
It sounds like an indexed lookup is close to O(N) = constant, which is *ideal*.
Not only are we boring, but we irk folks when we insist on measuring things, like performance.
May the force be with you (nyuck nyuck)
Unh-uh! 
Not me. I’m allergic to formulas I can’t understand, and I never got to take either calculus or statistics (had to settle for trig and finite math). 
Taking the hint… i will NOT explain those formulae.
So how do VBers measure performance of their code? [That’s why I built my Meter]
Um … by whether we get paid?
Since many (most?) of us are not engineers, we tend to do quick and dirty testing rather than serious, in-depth analysis. Once we find tools and techniques that turn in reasonable performance, that’s what we use. PCs today are so fast that unless the user notices a delay, it may not be worthwhile chasing the “fastest” solution unless the application requires that kind of speed for technical reasons. None of mine do. 
Does your meter also tell you how long it took you to test the speed of various data structures? 
Ah yes, for pay, a crucial point, since that IS the bottom line. In fact, the “solution” that i am working on is something that i would like to sell, and it is such a 
that *I* wouldn’t buy it. That’s my “marketability” test.
Of course, speed is not everything; faster solutions often take more space, and more time to produce. I’m not directly measuring the latter (yet), but the Xybernaut (a wearable computer) is produced & sold here in No. Virginia . If i can get it to monitor my vital signs, … non-invasively, that is…
What I found “interesting” was the Stack implementation in Getz’s book, which is implemented with a linked list. Rewriting it using an array gave me a 5-time speedup. But that one was pretty obvious if you have experience with these things, and took almost no time. Well, a couple of hours.
Examples in books aren’t necessarily the most efficient way to do something. They’re there to show how to use a technique and exposition sometimes takes precedence over efficiency. 
Didn’t I hear something about the Xbernaut being faster than a speeding bullet …? 
One of the benefits is that now you KNOW. The expense is tuition that we programmers continue to pay. You also have to include the time spent posting in this forum.
i would guess by looking at the size of the array (N=7) that it is too small to display ‘difficulties’. ‘Advanced’ algorithms are written to deal with cases when N is ‘large’.
But now i will have to retire the joke i was planning: ‘How many VBers does it take to change a lightbulb?’ Answer: ‘Who knows? No one ever measured it’.
Or at least change the punchline to 
uhh… let me work on that … 
How about to “Who knows? They’re still trying to measure the algorithm to find the fastest method.”
Donations from Plus members keep this site going. You can identify the people who support AskWoody by the Plus badge on their avatars.
AskWoody Plus members not only get access to all of the contents of this site -- including Susan Bradley's frequently updated Patch Watch listing -- they also receive weekly AskWoody Plus Newsletters (formerly Windows Secrets Newsletter) and AskWoody Plus Alerts, emails when there are important breaking developments.
