Brett - Unix is designed to work this way when open files are deleted or
another file is mv'd on top of them. *ALL* Unix machines will work this
way, so you are not exploiting some weirdness of Linux. The only problem
I see with it is that if a transmission is halted and a client sends a
request to restart in the middle, there would be an abrupt skip in the
video rather than just a pause. Seems like an insignificant issue to
me...
Doing it their way, you also need a mechanism to track obsolete clips
and delete them from the system based on their age and/or the time of
the last request, possibly taking into account the maximum download
time of the slowest client.
Tell your Windoze friends to read a book on Unix and get with the program. :)
Jim
>
> Sorry, this is for Unix/Linux.
>
> What you and Jim said is what I thought was possible...I even tested it,
> and it worked. However, when I proposed this to others in my company,
> they basically said that I was nuts. They said the proper way to do this
> was to:
>
> - put the new updated video clip on the server under a different name
> - have a database table with the names of the video clips in them,
> associated with the links in the web pages. So that once a video clip is
> updated, then you would update the database to reflect the new name of
> the video clip. Then the web pages would be dynamically generated by
> reading the database, and would point to the new name of the video clip.
>
>
> Anyways, this seems to me more work then needed, unless you specifically
