Depends how you transfer it: assuming you're downloading using HTTP, FTP, BitTorrent or some other TCP service, all the packets of data will (eventually) arrive intact. It will just take longer on a slow connection.

If you were pulling data down on a V92 modem using kermit or a more primitive mechanism, then transmission errors would be a possibility.

But using internet protocols, the only way you'd have a corrupt file is if the download were incomplete; this could happen if your browser or download-manager is a bit dumb about decided when the transfer has finished. Then you end up with a short file: the data you have is intact, but you don't have all of it.

If your internet connection is flaky, and you often get disconnected, your software should still be able to cope with that and should just pick up where it left off when you're back online.

The underlying TCP protocol uses a checksum to ensure that each packet (actually a "segment" in TCP-speak) is correct, and will retransmit segments which are found to be corrupt.

Having said all that, it is just possible that some stray cosmic rays could flip a bit in the data as it arrives on disk, so where large files are distributed there's usually a checksum published somewhere: typically it's an MD5 or SHA hash, and you can find software to compute these checksums on your own copy of the file. If your nervous about picking up any old (possibly virus-riddled) free software, then Microsoft provides a checksum tool.

BitTorrent clients use checksums implicitly to guarantee the data is intact.

Depends how you transfer it: assuming you're downloading using HTTP, FTP, BitTorrent or some other TCP service, all the packets of data will (eventually) arrive intact. It will just take longer on a slow connection.

If you were pulling data down on a V92 modem using kermit or a more primitive mechanism, then transmission errors would be a possibility.

But using internet protocols, the only way you'd have a corrupt file is if the download were incomplete; this could happen if your browser or download-manager is a bit dumb about deciding when the transfer has finished. Then you end up with a short file: the data you have is intact, but you don't have all of it.

If your internet connection is flaky, and you often get disconnected, your software should still be able to cope with that and should just pick up where it left off when you're back online.

The underlying TCP protocol uses a checksum to ensure that each packet (actually a "segment" in TCP-speak) is correct, and will retransmit segments which are found to be corrupt.

Having said all that, it is just possible that some stray cosmic rays could flip a bit in the data as it arrives on disk, so where large files are distributed there's usually a checksum published somewhere: typically it's an MD5 or SHA hash, and you can find software to compute these checksums on your own copy of the file. If your nervous about picking up any old (possibly virus-riddled) free software, then Microsoft provides a checksum tool.

BitTorrent clients use checksums implicitly to guarantee the data is intact.