I find that using the ID works perfectly well... I'm using UUID only on those situations where I need to pass it as vars/args (URLS for example) so that hackers cannot guess the IDs....
what problems have you got moving the DB to another computer ? On Fri, May 11, 2012 at 2:07 PM, Johann Spies <[email protected]>wrote: > I developed my original app to use uuid-crossreferences in stead of the > normal 'id'-field. The reason was that I wanted it to be consistent when I > move the database to another computer. > > Now I am rewriting the app to make it more efficient and to make the code > cleaner - applying some new features that came into web2py recently - e.g. > archiving. > > I am considering it to drop the dependence on uuid-references and move > back to id-based references and to keep the present references in the > database valid I am thinking (as an experiment) to do it in the following > way: > > * add a computed field to each table that use uuid-based references. This > computation will then lookup the id in the other table based on the > uuid-reference. > * after confirming that the references work well, I plan to drop the > uuid-fields and uuid-based reference fields from each table using them. and > change the computed field to normal reference-field. > * remove archiving methods based on uuid which I have developed and use > the new archiving in the trunk. > > Am I overlooking something? > > I would like the opinion of some experts in this list on what I have in > mind please. Is there still a place for uuid-based references? When? > > Regards > Johann > -- > Because experiencing your loyal love is better than life itself, > my lips will praise you. (Psalm 63:3) > > -- Sebastian E. Ovide
