dvfs api and toolkits

[Nate Nielsen]

Sean Middleditch wrote:
> Atomic whole-file writes are thus possible on pretty much every file
> system and protocol I know of in major use without needing any form of
> locking.  

That's wonderful, and it's an elegant solution to one particular problem.

> If we try to emulate seeking in HTTP by re-requesting files,
> we make that impossible without adding locking.  If we just tell apps to
> download the files and seek on the local cache then we will still retain
> the atomic write capability. 

This seems like a good idea at first, but it breaks down when files are
large. Let's say I open a movie file (let's say, for example, over a
local network) and seek halfway through it. Does the entire file (ie:
hundreds of megabytes worth of data need to be cached before my the
application stops saying "buffering...".

I don't propose full POSIX semantics in DVFS. All I'm saying is that if
the solution needs to be complete, in addition to the apparent goal of
elegant and conceptually simple.

Cheers,
Nate