I know this is going off into a different area but there is also a big overhead that ColdFusion (note not all CFML engines are this "bad") adds to a request. Give a servlet and cfm page that creates the same domain object of the servlet (the java object on both sides and I was not using Mark's Java loader the jar was on the classpath) the servlet consistently returns in <100ms (that's round trip to local host). The cfm upon server startup could take up-to 2000ms but typically at it's fastest I saw 500ms when I ran webLoader on it. Let me put this in perspective though, there was no database interaction no CFC objects and everything was self contained in a single jar. In a real world application you will have CFCs (not a bad thing) and database interactions the network ltency and speed of the DBMS will affect your application as much as the CFML engine will. Some of us get real caught up in these domain model talks and making full blow "correct" OO architecture. Reality is if you want that then CF is not, has not and will never be the appropriate tool to model your business in. Can CFML provide a wonderful UI layer for HTML based or quick service fronting of a complex domain yes it can and does so very nicely in many respects. I'll also disagree with the comment that groovy is magnitudes slower, from my personal tests groovy is not magnitudes slower...noticeable maybe but not magnitudes.
Now back to the training, there is a rather dated book called Java for ColdFusion Developers pick it up pretty good reading, I also recommend Thinking in Java for folks starting out as well. I also think Hal Helms offers training so you might ping him and inquire. The whole XML XSLT thing is a bit odd to me but hey to each thier own, I would recommend looking at something like tapestry or velocity before trying to use straight XML and XSLT (or hell if $ is the issue switch to OpenBD or Railo and use them as the UI layer only). Adam H On Wed, Mar 11, 2009 at 11:37 PM, Jaime Metcher < jaime.metc...@medeserv.com.au> wrote: > > > > > I'd want to see some pretty solid benchmarks before I > > believed that the Java written by CF is 200 times slower than > > the Java written by a random programmer. > > The 200 times difference comes from my own comparison of the same domain > model implemented idiomatically in both ColdFusion/ColdSpring/Transfer and > Java/Spring/Hibernate. Similar experiments conducted by colleagues give > similar results. Mark Drew did some testing with a totally different > scenario, but the results are in the same ballpark. > > Not sure why you'd find this surprising. Java code to solve a given > problem > isn't remotely similar to the bytecode that CF spits out for the same > problem, unless that problem happens to be interpreting a dynamic language. > The fact that they are both "Java" is far less significant than the fact > that one is dynamic and the other static. There have been similar results > from comparisons of Jruby and Java, and Groovy and Java. We're not talking > about 20% or 50% performance differences, we're talking about orders of > magnitude. > > Jaime > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to date Get the Free Trial http://ad.doubleclick.net/clk;207172674;29440083;f Archive: http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:320430 Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
