TMG-L Archives

Archiver > TMG > 2004-09 > 1094139673


From: Darrell Martin <>
Subject: RE: [TMG] Entering / Selecting Places
Date: Thu, 2 Sep 2004 11:41:17 -0400 (EDT)


Hi, John:

TMG certainly has a database file (the project_p table) that stores place information. That file is even closer to being relationally "normalized" than some other TMG tables are. However, that is not the same thing as what I meant by, "a true Master Place List stored as a separate file". I might have been more precise if I had said, "TMG's Master Place List display does not necessarily correspond directly to records in a single file", but in general I still believe what I said in my earlier message is an accurate characterization of the situation.

The reason is that TMG's "MPL" is not a display of the _p table, nor of selected records from that table, nor even of selected field values from selected records in that table, but rather (as you say) "a list of the unique combinations of fields that occur in the MPL entries where the fields are limited to the remaining place fields on the data entry screen". My only quibble with your statement is that "in MPL entries" should better read "in _p table records".

Problems with the "MPL" have seldom if ever appeared in my testing when there were no undisplayed fields involved. Using F2 with the cursor in the Address field of the Tag screen is one example where things have always worked as expected for me, since the "MPL" in that case includes all the L1-L9 fields in its selection process. Working with places that use the default Place Style, and have no date range or comments, is another example of a context in which I have not seen problems.

I think sometimes fields other than L1-L9 are the cause of unexpected behavior; at other times, the way the contents of non-L1-L9 fields interact with user input seems to be a symptom of something else. Sometimes the non-L1-L9 fields act as if they were integral to identifying or selecting an "MPL" record ("as though they were part of the 'key' "); TMG will not optimize places that differ only in date range, for example. Other times these fields act as if they are only peripheral to the "MPL"; the way the Place Style can change when F2 is used with the cursor in a place field on the Tag screen is an example of that.

When TMG 5 was first released, I poked my nose into every nook and cranny of its place functions that I could find. I tried changing the labels for the L1-L9 fields on screen and in data manipulation. I tested the way a custom Language could be used to modify place usage. I tried to use place comments in output. I attempted some pretty bizarre things with Place Styles. There is a very great deal more to all this than just how the _p table stores data. It is also true that some of the problems that I recall have been dealt with between 5.0.1 and the current version of TMG.

Anyway, the original issue in this thread was how to enter or select places. I suggested that users should be careful when using F2 from a Tag screen, especially if any of the place fields are edited after the selection is made from the "MPL". I still highly recommend that great care be taken in this area.

Darrell M.



Darrell A. Martin
a native Vermonter in exile in Illinois


This thread: