Linked tables problem - mapped drive vs. UNC

[dima69]

I have an application with a backend database connected to front end via mapped drive. It is installed at my client's office. Occasionally they report a problem while opening an application first time, after turning computer on. An application cannot access linked table. But, after entering the build-in browser and relinking the tables, the problem disapears. This problem is not persistant, so I couldn't reproduce it. Also, drive mapping has not changed !!!
So my question is: Have anybody seen that kind of problem ?
Is there any REAL disadvantage of having tables linked by mapped drives instead of UNC ?

[dima69]

Any suggestions ?

[NeoPa]

The Windows NOS will often lose a connection, but will recreate it when it detects the situation. In most software that doesn't cause problems, but in some it will report errors even though they are generally recoverable.
Strangely, the apps that generally handle this the least well are M$ Office apps.
A UNC may work better, you might try it out. Sometimes that's not appropriate though.
It is unusual for the reconnection to take long enough to cause a problem.

[dima69]

Thank for the reply. However, I havn't quite understood what you mean by

Sometimes that's not appropriate though.
It is unusual for the reconnection to take long enough to cause a problem.

i.e. why UNC connection could be unappropriate ?

[NeoPa]

Thank for the reply. However, I havn't quite understood what you mean by

i.e. why UNC connection could be unappropriate ?

Simply that in some circumstances you need to connect to a drive letter. This may be because different users (or maybe even PCs) may have the same drive letter, but mapped to different UNCs. Your Home share for instance is often mapped to a single drive letter, but will be different for each login on the network.
I just didn't want to make assumtions about your setup.
I can't see any technical reason for you not to use a UNC. Only you know your own circumstances though.