On 31 Aug 2005, at 16:16, Matthew Toseland wrote: > Would you please explain what the practical difference is between the > following: > > 1. Put everything in FCP. Including queueing and big FEC downloads. > Binary metadata, hierarchical metadata, etc, dealt with by the node. > Third party clients interface to this. Clients can range from very > simple (but still able to download any size file) to very complex. > > 2. All of the above plus: > - Fproxy can queue a file, so the user doesn't have to copy a link > from > the page into a separate application just to download a file > posted on > a web site. Or are you saying that files shouldn't be posted on web > sites??? THAT I strongly object to. > - Fproxy offers a minimal status interface to show the current > status of > the global queue.
I have never argued that FProxy shouldn't provide status information. Other than that, the main difference between 1 and 2 is that 1 is less work, while providing all the functionality the vast majority of our users are likely to want in practice, and without wasting time on "white elephant" functionality that is cumbersome to implement, and which doesn't work well. Ian.
