If local path is not specified, then the current directory is assumed.
If local path is a directory, then the last segment of the path of the url is appended to form a local filename. If the url path ends with slash the name ``index'' is used. With the -s option pick up the last segment of the filename from server provided sources like the Content-Disposition header or any redirect URLs. A file extension to match the server reported Content-Type might also be appended. If a file with the produced filename already exists, then lwp-download will prompt before it overwrites and will fail if its standard input is not a terminal. This form of invocation will also fail is no acceptable filename can be derived from the sources mentioned above.
If local path is not a directory, then it is simply used as the path to save into. If the file already exists it's overwritten.
The lwp-download program is implemented using the libwww-perl library. It is better suited to down load big files than the lwp-request program because it does not store the file in memory. Another benefit is that it will keep you updated about its progress and that you don't have much options to worry about.
Use the "-a" option to save the file in text (ascii) mode. Might make a difference on dosish systems.
$ lwp-download http://www.perl.com/CPAN/src/latest.tar.gz Saving to 'latest.tar.gz'... 11.4 MB received in 8 seconds (1.43 MB/sec)