>>"Cursors" as in the ability to define and create tables on-the-fly in memory which may or may not have a disk presence, create indexes on fields in those tables, add/delete/edit >data in those tables, use them in reports and other purposes and then, when closed, they disappear.
>
>I agree. Python Dabo does exactly as you describe and more. Which one if the reasons I went with Dabo and python (actually love the language).
>
Not enough practice with Python here as it is not a dark, but an almost invisible horse over here for my kind of work, but the bit of stuff I did put it at the top of the line for dimensions of clarity and speed of grokking other peoples sources. Has been sometime since I looked at the Dabo sources, back then Dabo used Python cursors but needed to stuff them back into SQLite to be able to SQL-manipulate them. Piping your remote data directly into a local SQL-manipulatable datastore (which might be memory: targeted for speed reasons by the dev somewhat half-automagically) IMO is a better architecture to base your biz objects on, for example to sub-filter locally and not always call back to remote when using multiple father-child relations on a dataset. But for me the failure of python to get foothold in xPhones and xPads as strong as on desktops across all major Intel OS including Linux (with GUI library...) turned me away from Dabo: portability not only across windows desktop machines but Mac and Linux as well is not good enough after the "personal device" gets redefined to describe a touch-based small device - Dabo would only slightly increase possible reach to those desktops NOT running windows compared to vfp and still miss out the pad/phone market. Any changes done in that area in Dabo?

tia

thomas