Hi Bill and co,
I see you've been busy while I've been sleeping!
We use MS Access for a number of use cases:
Surgical intervention when things go wrong
Running scripts (MS Macros etc) to perform actions directly on EA Tables - including adding objects, relationships, attributes, operations and tags
Running scripts to check higher level validity or structure than that obtainable with the Integrity checker.
These are all administrative use cases - not for normal users.
We don't do data entry, per se, in MS Access. But there's no reason not to. But Geert is right, for normal users you should try to do everything from EA in some way.
We do have tables not in the EA Repository, some common to all repositories (usually metadata related) others specific to the target repository (as we change the target repository via a linked table manager, we automatically re-link these external tables to the requisite MDB.
Now, it took a few years to get all this additional framework in place - but now we have a significant amount of freedom to create our own modelling environment.
However, as the others have said, we try to use existing functionality within EA wherever possible - using predefined tagged values against metatypes a case in point. In fact, sometimes we have been "ahead of the game" and when EA has caught up, we've retired our "extension" and switched to the new EA functionality.
We took the view that EA is not just a product, but a framework. We build our own environment /configuration on top of that framework. Don't be afraid of using MS Access, but don't see it as a panacea.
BTW we also interact with EA via MS Excel (we automatically synchronise formal "Catalogues" with the repository, for example).
My advice to you is to look at what you want to do, decide on a design and the put it up for comment on the forum. Over the years, I've donethat where I wasn't sure if the idea was correct.
HTH,
Paolo