Mikus - You should check out the very recent thread entitled "Ideas for Journal: How Epiphany...", because we're discussing just the type of things you bring up here.
On Sat, Sep 20, 2008 at 9:54 AM, Mikus Grinbergs <[EMAIL PROTECTED]> wrote: >> The current plan is to land a rewrite of the datastore early in the >> release cycle, using the same API and the same user interface. That >> will mostly help reliability and performance but it's also a >> prerequisite for the new design and likely for any journal UI >> improvement. It might or might not support versions. >> >> What is not defined is which new features we should aim to develop on >> the top of the new datastore in the 0.84 release cycle. > > The Journal "concept" promised to replace traditional 'hierarchical' > access with 'intelligently filtered' data access. But the existing > 'filters' available in Journal are pathetic. DESPERATELY needed are There should also be a filter for "who" you worked with, which hasn't been added for technical reasons I don't fully understand. However, the new back end mentioned should, I believe, make that possible. > sorting of displayed entries, Yes, this is something that's been in the designs since early on, but we haven't had time to do yet. We even considered the more advanced concept of "sort by, then by, then by.." to allow pseudo-hierarchical structure to emerge. > "filtering" on 'tags', I'm pretty confident that this should work at present. If it doesn't, please open a new ticket for us! There was another recent (as of today) ticket that noted searches don't work correctly on metadata fields (perhaps that's what you meant?), which is definitely a bug we need to fix. There's also a related bug that prevents custom metadata from being preserved across reboots. That's pretty fundamental. I really hope we iron out the tagging and metadata system and search for the next release (even if we don't get to all the "fancy" stuff being discussed in the other thread, yet), but there's a lot of work to be done there. > and "stacking" > of search terms (i.e., to provide the equivalent of "search within > previous result"). This should also work fine, I think. Terms are combined with boolean AND so by typing additional words into the search you should continually narrow your results. If we ever get a "tokenized" (treating completed words which match on existing tags as single, tokenized, objects in the search field) system, it would be easy to then go back up levels, or remove a level, by deleting specific tokens. Thanks for your feedback; I encourage you to check out the other discussion I mentioned! - Eben > mikus > > _______________________________________________ > Sugar mailing list > Sugar@lists.laptop.org > http://lists.laptop.org/listinfo/sugar > _______________________________________________ Sugar mailing list Sugar@lists.laptop.org http://lists.laptop.org/listinfo/sugar