I guess the way I look at it is that, since we're re-writing all of the views and database queries anyway, why re-write them to match the old system and then spend more time changing them again later, instead of just writing them how we want them in the first place? Similar argument with modularizing, if we're copy-and-pasting algorithms, may as well paste them where they're going to end up, rather than having to shuffle them around in a couple months.
But that's me, always looking for a little extra short-term work to let me be lazier in the long run. And, in this case, I'm not even convinced that it's any extra short-term work... Greg. -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Dave O'Neill Sent: Thursday, December 24, 2009 2:52 PM To: Development discussion for the Leaguerunner league management system Subject: Re: [Leaguerunner] Converting Leaguerunner to CakePHP On Thu, Dec 24, 2009 at 02:36:42PM -0500, Greg Schmidt wrote: >Thoughts? Anyone? I'm looking to start on the groundwork (database >conversion) next week, and would like to make sure I start in a direction >that best serves everyone. I haven't really had much time to look at it in detail, but here's some off-the-cuff thoughts: >a. There are many issues (from small to medium) in the current version of >Leaguerunner. Some of these require database changes to correct. If we >make these database changes in step 1, the code generated by step 2 will be >closer to final and not require as many changes to complete it. On the >other hand, changing functionality during the initial conversion process may >make it harder to find the cause of things that are broken. So the question >is, do we take steps to correct these issues right away, or do we save that >for a later version? It might be best to clean up any minor things that need fixing, but save any major corrections for a later version. >b. How hard do we try to make things look like they do now? For example, >there are places where things could be spruced up by using Ajax, and it >seems to me that a lot of the links that Leaguerunner currently has in the >dynamic areas of the menu are handled by most web apps these days by >embedding them in logical places in the output. We don't want to confuse >the users too much with changes, but if we can simplify typical usage I >think that's worth doing. I'm all for fixing look-and-feel issues, but they're probably best left until after the conversion is complete. >c. How much do we try to modularize existing functionality. I think that >things like scheduling algorithms, spirit questions, etc. may be pulled out >into pluggable modules, as has been done with the online payment provider >modules. Yup, I'm in favour of that too. However, I think that's probably best left until after the CakePHP conversion unless it's trivially done during the conversion. Cheers, Dave _______________________________________________ Leaguerunner mailing list [email protected] http://www.dmo.ca/cgi-bin/mailman/listinfo/leaguerunner _______________________________________________ Leaguerunner mailing list [email protected] http://www.dmo.ca/cgi-bin/mailman/listinfo/leaguerunner
