• Table Occurrence field order
    • Field sorting order within the fields tab are arranged using the View by "custom order" option. When arranging table occurrences on the graph, they are placed based on the "custom order" setting.
  • Table Occurrence Placement
    • Unfinished
  • Color use
    • Unfinished
  • No labels


  1. I'm really attached to alphabetical sorting. Custom orders are only helpful when you already know what that order is. There is nothing more annoying about working with someone else's code than a full-screen-height scrolling list of fields that is in an order I don't (yet) understand. That makes collaboration and maintenance (because we'll all forget what made sense 18 months ago) more difficult. Alphabetical sorting is much more practical to navigate.

    Using the field sort order for arranging table occurrences is reasonable enough, until you bring in (multiple) multi-predicate relationships. It may not be topologically possible to avoid crossing relationship lines. Keeping relationship lines uncluttered and the fields they connect to visible is a nice-to-have, but it just can't always work. Opening the Define Relationship dialog to see what's happening isn't really all that bad, is it?

  2. Anonymous

    I agree, alphabetical sorting for fields

    As far as the relational graph is concerned I have the tables sorted in alphabetical order in my data files, but logical contextual sorting in the interface files. But the children of the relationships in alphabetical order. All TOCs in the same tier the same width, the minimum width that can contain the longest name. 

    Tim Anderson

  3. Anonymous

    I'm curious about the views of the contributors here regarding relationship graph organization. When I started out, I just started connecting one table occurrence  to another and wound up with graphs that looked like a bowl of spaghetti or a close-up picture of malaria germs. Then I found the anchor-buoy method and have been using that for quite a while. It does get to be a challenge maintaining so many sets of relations when I have a database with lots of tables and layouts.

    From looking at the Standards.fp7 file and watching some of the videos over at Filemakermagazine.com, it looks like Matt is using something other than A-B. I found Matt's Graph Tactics video from 2007, but wondered if his (and the filemakerstandards community) practices had changed since then.



  4. Anonymous

    I have to say, the only time that I've found alphabetical sorting to really make sense is when I'm working with an existing, not-so-well documented solution.  If you're developing from scratch, I feel that Custom Sorting is a much better approach.

    1. Custom sorting has advantages for the original developer of a solution, but a professional developer should assume that every solution will eventually become an "existing" solution, and any supporting documentation (including the original developer's memory) will be irretrievably lost. I don't personally find the advantages of custom sorting to be robust under these assumptions.