TableBuilder, Equates and bad dict names (OpenInsight 32-Bit)
At 25 APR 2006 11:44:57AM dsig _at_ sigafoos.org wrote:
Working with an old bad pick dict names .. down around 150 there was something like G/L_INV
I didn't see it and then created the equates (using the handy dandy Equates feature of Table Builder). Then used the equate in a program and on the first compile got error messages about the name .. upon investigation i found the 5 bad dict names .. oh my
Without getting into the 'why does OI even care about dict names' discussion, would it be a good idea that IF an equate record is being build by the system (oi supplied sw) then should it make sure that the names are good .. ??
At 25 APR 2006 01:18PM Gerald Lovel wrote:
Well, the first time I tried to use the Automatic Equates builder, it was with my filename NOTES. I would say that validating dictionary names is not the only issue with this program feature, which is why I maintain my equate lists by hand. But then, my tables do not have 150+ data fields, or I might have programmed my own routine for this, which may be an option, DSig .
I don't expect every option of every design module to trap for every user foobar. I do expect the main program design modules to do their jobs right so I can get my job done. OI works pretty well for that.
At 25 APR 2006 02:14PM dsig _at_ sigafoos.org wrote:
I like to use the tools as delivered .. as a developer i realize the 'drive' to write everything myself BUT i resist that
Since CTO is a new portion of the product and since the idea is to get 'PICKIES' over .. i look at shoring up these bits as ways to make sure all bits follow the rules.
but thanks for the thought
At 25 APR 2006 02:26PM Gerald Lovel wrote:
DSig,
Well, hold that thought. If we are talking about converting the archaic MV applications to OI, then we really do need to focus on the tools which make that translation easier. From that perspective, Table Builder needs to be perfect in converting dictionaries, not just in recognizing when it dislikes about some previous DICT naming convention. I believe you have remarked on this before, and I would agree.
We could worry about whether changing DICT names for OI compatibility is going to work. However, I have heard that Pick-alikes were not good at referring to dictionary fields by name. For example, the braces {} notation in dictionary symbolic code is not a part of other systems, I believe.
But a bigger issue is, what do you do to take advantage of OI once the initial CTO coversion is completed? I would like to start a conversation with you on how to restructure the MV applications for OI and Windows, as I think I have something to contribute there. Interested?
Gerald