[UK-CONTEST] Redundant contest reports

Andy Swiffin a.l.swiffin at dundee.ac.uk
Mon Feb 9 05:46:18 EST 2009


>>> On 09/02/2009 at 09:57, in message
<BV.WM.2.0.pv.1.0.18.0902090957380.6694 at webm10.global.net.uk>, Peter Hobbs
<peter at tilgate.co.uk> wrote:
> Andy
> 
> I suspect what you are suggesting is probably a bridge too far for the 
> typical logger user. I'd suggest that Paul has gone about as far in terms 
> of flexibility as is reasonable for most of the target users. An 
> arrangement with built-in scripts for existing (known) contests and an 
> alternative DIY deal for other stuff would meet your intent, but I don't 
> know whether the added complexity would be value for effort, or even 
> effective without quite a lot of detailed knowledge of the program.
> 
> We are just talking about trying to avoid the need for long term support by 
> the author(s), which the basic user is always going to need.

Well I'm not suggesting that the "average" users will craft up the script for a new contest - this would usually end up being the "power" users.   But it would mean that when new contests came along that there wouldn't (normally) need to be any change to the actual program code at all making it so much more flexible and able to adapt quickly.   I'm not sure it is added complexity.  It's actually reducing the complexity because the code that exists in many of the existing loggers must be a nightmare to maintain where there are so many conditionals dependent on each different contest.   By stripping the "data" out of the program you have one "lean" program code path which "interprets"  the data for the different contest exchange.    I wrote something similar years ago as an interpreter which ran an ATM (cash machine) mock up used in trials of touch screens (when they were very new!)  and then later went on to use the same interpreter code base to run a totally different application just by changing the data.    I suppose a  contest exchange is a finite state machine?  It has a couple of defined entry points, a number of intermediate states with input dependant on the contest and an exit.

Anyway, just a thought, it always niggles me to see a piece of software with no distinction between the code and the data.

73
Andy
gm8oeg




The University of Dundee is a registered Scottish charity, No: SC015096


More information about the UK-Contest mailing list