TMG-L Archives

Archiver > TMG > 2001-08 > 0997450783


From:
Subject: RE: [TMG] Suggestion for TMG - tag order
Date: Fri, 10 Aug 2001 09:39:43 -0400


I'm several days behind, but I find it interesting that the issue of "implicit" date sorting of tags is getting a lot of attention.

If the computer system gives you few options (e.g. types of "events"), then it may be fairly easy to define rules for how they are related. As you increase the number of options, those rules would become more complex. If we open the floodgates and allow an infinite number of options, defining the relationships of the options becomes much more difficult.

People talk about a birth event (tag), marriage event (or wedding), and a death event. However, TMG provides many types of tags (some with no dates), and gives the user the ability to define their own tags. Users can specify multiple birth, death, and marriage tags for the same person(s), possibly reflecting conflicting evidence. These can be added and removed at any time, in any sequence. There can be several different types of tags to record each of these events. Tags can tie together lots of people (witnesses), which have their own set of events/tags.

To try to program this complexity in a way that will never mislead the user is not easy. Bob Velke has decided that it is not worth it, both from a practical and a theoretical point of view (as Lee Hoffman has reminded us).

Sort tags provides a way for the user to make his/her own decisions. When you see a display of tags that seems to be in the "wrong" order, fix it. If you are willing to let John Cardinal's utility do that for your whole dataset, use that, and decide which assumptions you want. If you think you will never get your database in the shape you are satisfied with, you are right. And sort dates are one of the least of your problems.

I think that there are a lot of other features we would like to see in TMG that we can't fix ourselves, and we should hope that Bob and crew are working on those first.

Pierce Reid

This thread: