>Hi Jeff,
>
>>My boss suggested that we do a separate VSS install out there and then use the "synching-up" feature to synchronize both VSS databases and source code. Is there such a thing?
>
>If there is I'd like to know about it. *s*
>
>
>>What are my options for doing an independent installation of SourceSafe and then synchronizing both groups of source code over the WAN?
>
>VSS is a file server system like VFP. Because of that, and they way it stores files, it has roughly to bring down 2.5 times the file size for checking in and 1.25 for checking out. Add the overhead for the locming scheme, the NETBios overhead, and you get a real slow performance.
>
>There are some possibilities within VSS to make it faster to access the VSS database via a WAN. First of all, you should make sure that the database is stored in VSS6 format, because that format requires less file lockups and therefore is significantly faster than the old format.
>
>Second, do not access the VSS database via the project manager. While it's acceptable on a LAN, it really slows down over the WAN as VFP performs a lot of unncessary checks and file lookups in the database. That's one of the situations where it's better to use the VSS Explorer, even if you otherwise use the SourceControl integration.
>
>Third, you can gain a lot if you define how VSS should recognize that a file hasn't changed. There are three options: content, time and checksum. Obviously content is the worst option, because it has to download the entire file before it can compare them and decides based on that whether the file should be actually downloaded. Checksum is more secure than time, because it doesn't depend on any time difference on different computers, while time is the fastest.
>
>Further, VSS 6 uses a file (VSSVER.SCC) in each directory that contains the checksum and other file information from the sourcesafe database. By comparing the actual file entry against what is stored there, and in the VSS database, it can determine whether you or someone else modifed a file and only downloads/uploads a file if that's necessary. For this to work, you obviously shouldn't delete the file, but also should not use more than one database for each directory. That is don't get the file from the remote VSS database and your local VSS database into the same directory, because they overwrite each others entry in that file.
>
>And, if you want to check in/out several files, you can do this via a batch file instead of interactively. This way you can run it through the night and start with a fresh set of files each morning.
>
>But finally, there's a tool out there that let you avoid all this stuff. It's a server that runs on the same network as the VSS database and does all the checking locally. With this tool, only the actual file is transferred and I think that's done with compression, too. I keep forgetting its name, but I'm sure others jump in here.
>
>Christof

Thanks Christof, I was considering that solution already (use VSS outside of VFP project manager) and your input has helped me make up my mind.

Also, thanks Rick and David.

-JT