Marco,
You said a lot in you statement, too much for me to respond to in one go. Although I agree, in principle, with much of what you said.
First. your examples.
On the surface everyone would agree that A=B until you look at the software that needs to translate from A to B -or- B to A. I've written multiple database translator programs in my consulting job for the largest Business Software and DB company in the world so I have a little experience with this delema.
The delema has to do with two-way data translation. Generating consistantly equal data over multiple exchanges will over time cause issues.
Consider the following DB scruture:
AMT Structure:
Source_Record
-Title
Citation_Detail_Record
-Vol
-Issue
-Sec
-Page
-Text
FTM Structure:
Source_Record
-Title
-Vol
-Issue
-Sec
Citation_Detail_Record
-Page
-Text
AMT Data:
SOURCE: Title="The Daily Newspaper"
Citation1:
-Vol="1"
-Issue"2"
-Sec"3"
-Page="52"
-Text="John Died"
Citation2:
-Vol="1"
-Issue"5"
-Sec"3"
-Page="21"
-Text="Billy Bob Died"
Translating this to FTM would need to build a source record first. The AMT.Source.Title goes very easy into FTM.Source.Title, but what about FTM.Source.Vol (and Issue and Section) what populates them? From AMT.Citation you can only use Citation1 you would need to create a whole new source to support translating data AMT->FTM from all different citations.
So in FTM the outcome of the translation would generate a considerable number of additional sources than what was in AMT. Not a bad thing but could consern users. What complicates the issue even more is when some of the data translated is partly equal to other sources that are already in the target DB. A doable but complicated compare and match would need to be done for every AMT Source/Citation combination on the translation to an FTM source. People think the SYNC takes long now. Wait until this compare/match is done.
Translating FTM->AMT has the same issues in reverse. A FTM.Source (plus any FMT.Citations) would have to be broken into one AMT.Source (matched and merged) and multiple AMT.Citations (Matched and Merged). Think about the possible distruction that could happen if things go wrong!!
I'm not saying that this can't be done, I just do not think that having two different DBs and data entry stratagies work well together over time and that one will eventually have to be eliminated. Bandaids will not work.
You said a lot in you statement, too much for me to respond to in one go. Although I agree, in principle, with much of what you said.
First. your examples.
On the surface everyone would agree that A=B until you look at the software that needs to translate from A to B -or- B to A. I've written multiple database translator programs in my consulting job for the largest Business Software and DB company in the world so I have a little experience with this delema.
The delema has to do with two-way data translation. Generating consistantly equal data over multiple exchanges will over time cause issues.
Consider the following DB scruture:
AMT Structure:
Source_Record
-Title
Citation_Detail_Record
-Vol
-Issue
-Sec
-Page
-Text
FTM Structure:
Source_Record
-Title
-Vol
-Issue
-Sec
Citation_Detail_Record
-Page
-Text
AMT Data:
SOURCE: Title="The Daily Newspaper"
Citation1:
-Vol="1"
-Issue"2"
-Sec"3"
-Page="52"
-Text="John Died"
Citation2:
-Vol="1"
-Issue"5"
-Sec"3"
-Page="21"
-Text="Billy Bob Died"
Translating this to FTM would need to build a source record first. The AMT.Source.Title goes very easy into FTM.Source.Title, but what about FTM.Source.Vol (and Issue and Section) what populates them? From AMT.Citation you can only use Citation1 you would need to create a whole new source to support translating data AMT->FTM from all different citations.
So in FTM the outcome of the translation would generate a considerable number of additional sources than what was in AMT. Not a bad thing but could consern users. What complicates the issue even more is when some of the data translated is partly equal to other sources that are already in the target DB. A doable but complicated compare and match would need to be done for every AMT Source/Citation combination on the translation to an FTM source. People think the SYNC takes long now. Wait until this compare/match is done.
Translating FTM->AMT has the same issues in reverse. A FTM.Source (plus any FMT.Citations) would have to be broken into one AMT.Source (matched and merged) and multiple AMT.Citations (Matched and Merged). Think about the possible distruction that could happen if things go wrong!!
I'm not saying that this can't be done, I just do not think that having two different DBs and data entry stratagies work well together over time and that one will eventually have to be eliminated. Bandaids will not work.