From: Gregg N 

"Geo."  wrote in news:411e9a0d{at}w3.nls.net:

> Oh I understand what you said, but how does it know what files it
> needs from Windows Update and what files to take from the local
> machine? Is that process foolproof/bugfree?

I am only speaking hypothetically, based on my understanding of what Rich
said. I don't know for sure how this actually works.

When it is done taking files from windows update and the local machine it
ends up with the exact same service pack, bit-for-bit, which can then be
installed. I imagine MD5 or similar checksums are used to verify the the
service pack was properly assembled before it is installed. If the checksum
fails, it can tell you you need to obtain the service pack through other
means.

> I mean think about this
> for a minute, verifying 100mb of patches or just assuming none are
> installed and installing everything from a known good source, which
> seems less error prone?

It wouldn't have to verify patches, it just has to see if the files it
needs are present. If they were corrupted, the checksum on the service pack
would fail.

> installed and installing everything from a known good source, which

You are using the term "installing", when you should be using the
term "downloading". I think it is only the "download"
part that is different between the WU route and the complete download
route. Once the download phase is complete, the "install" part is
the same.

Is my understanding correct, Rich?

Gregg

--- BBBS/NT v4.01 Flag-5