[AccessD] "Faked" replication

Dear Group,

I've been round and round on this for a couple of months with no clear answer. It appears that for Indirect Synchronization to work one must have an IIS web server in place with a hub replica on it and a bunch of FTP drop boxes. Then one has to wrangle with the broadband company to ensure that one may hang a server on the line without the broadband company getting into a snit. Then there's either the training/documentation issue or the programming issue to ensure that this Indirect Synchronization works flawlessly.

I tried the Experts Exchange and paid $10 for a month's worth of unlimited access. I've even answered a bunch of questions and have almost reached the first Expert level of 10,000 points...but both of my questions about Indirect Synchronization have received exactly zero responses.

The Microsoft Replication newsgroup is very disappointing. Michael Kaplan, the widely acknowledged guru of Replication, shoots sarcastic one-liners most of the time. I wonder why he bothers. None of the questions there dealing with Internet/Indirect Synchronization (they're different names for the same thing, I'm beginning to believe) have answers. This is nuts.

So, I'm back to considering what I can do in my particular environment. I will admit that I thought that Indirect Synchronization was appropriate for at least the startup phase of this project; the reason being that a mound of data needed to be transferred manually from Word to Access and then split up by Customer onto a number of disconnected laptops scattered around the state. The people with the laptops would also be doing data entry and would need to send their additions/corrections back to the home office. Once all of the old data is transferred to Access then there shouldn't be any need for the home office to distribute any updates into the field.

Therefore, I'm going to bite the bullet (and forego additional fees) to backtrack on my database design. It will end up where the home office will do no data entry at all, just collect information from the field units. Since each field unit is at a different Customer site, that may simplify the assigning of keys. I'm not real thrilled about using "random" AutoNumbers. Regular AutoNumbers aren't any good either because of the various data entry locations all needing to be merged at the home office. It appears that the natural hierarchical key would be: Corporation, Division, Location, Department, JobPosition...that is, five key fields for each record. The records hold ergonomic information about an industrial employee's work environment. Yes, there'd be all that extra baggage being hauled around in each record, but no chance that there'd be any duplication. 15 different Departments could have a JobPosition like "Hod Carrier", but as long as they were in different Departments !
 there'd be no duplication.

Is this sensible? Should I snap out of it and use random AutoNumbers?
 
Regards,
 
Steve Erbach
Scientific Marketing
Neenah, WI
920-969-0504

Message created with Bloomba
 
Disclaimer: No tree was killed in the transmission of this message. However, several coulombs of electrons were temporarily inconvenienced.