> That may be a good point, but what does it have to do with > the original topic? Sean originally stated that performance > gains measured in milliseconds were meaningless. I pointed > out that the way CF7 was coded allowed for performance > degradations which were larger than mere milliseconds.
.... if you're doing large enough numbers of iterations. In any case, I don't really care about the "original topic". I was responding to your response to Sean, which I thought was off-base. I've worked with many, many CF applications, and the vast majority of their problems are one of two things: 1. Poorly-designed database interaction, and 2. doing unnecessary work at runtime. The number of times that a specific way of doing things in CF has made a significant difference is very close to zero. I'm not going to say that these problems never arise, but they're clearly the exceptions rather than the rule. And yet, I see so many CF programmers focused on "which is faster" questions about CF tags, etc, that they lose sight of the big picture - the two points mentioned above. So, you'll have to forgive me, but I tend to respond to those sorts of questions a certain way. In the same vein, I see iterative tests used ALL THE TIME by CF programmers, and frankly I think they're horrible. They're misleading, and again they distract from the real performance issues you're likely to encounter. > Besides my iterative test, you keep ignoring the reports of > real live applications who suffered from the performance of > cfswitch tags and strings. If you have a real live application which suffers from some specific performance problem, you should address that performance problem. My point is simply that trying to solve the performance problem by premature optimization is usually a mistake. Having worked now with hundreds of applications that have had performance problems, I feel comfortable saying that any feature of the language, misused, can lead to performance problems. But that by itself isn't especially useful in telling someone how to build an application. And, of course, the specific numbers you quoted to Sean were from your iterative test. I think it's perfectly OK to critique the process by which you gained those numbers, if you're going to use them as justification for doing something one way or another. > And what makes you an expert in my code base? Have you seen > it? Fact of the matter is, I was just working this week in > some code which performed some conditional processing of > large result sets looping once per column and row. Based on > the data available it has the potential to run thousands of times. Your test ran a million iterations, according to you. You said that this doesn't represent what people actually do. I don't need to be an expert in your code base, I just have to have fundamental reading skills. And, frankly, based on what little I now know about your code base, I would recommend that you investigate doing that conditional processing in a stored procedure if you can, because that kind of operation is usually better done in an SP than in CF. So, where you might see switching from one decision structure to another as an optimization, I would instead see moving the decision structure out of CF altogether as the optimization. > You can beat the what-if horse all you want, but this thread > has clearly been about differences between CF7 and CF8. To > tell me people haven't coded in such a way that optimized for > their version of the compiler would be a lie. People do all sorts of dumb things. That's hardly a justification for those things. Dave Watts, CTO, Fig Leaf Software http://www.figleaf.com/ Fig Leaf Software provides the highest caliber vendor-authorized instruction at our training centers in Washington DC, Atlanta, Chicago, Baltimore, Northern Virginia, or on-site at your location. Visit http://training.figleaf.com/ for more information! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Check out the new features and enhancements in the latest product release - download the "What's New PDF" now http://download.macromedia.com/pub/labs/coldfusion/cf8_beta_whatsnew_052907.pdf Archive: http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:293670 Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
