very simple datastore reimplementation

[Tomeu Vizoso]

On Fri, May 9, 2008 at 7:34 PM, Marco Pesenti Gritti <mpgritti at gmail.com> wrote:
> On Fri, May 9, 2008 at 6:11 PM, Tomeu Vizoso <tomeu at tomeuvizoso.net> wrote:
>  > On Fri, May 9, 2008 at 2:15 PM, Jim Gettys <jg at laptop.org> wrote:
>  >> FUSE is great, but...
>  >>
>  >> It means interoperability must be an explicit planned-in-advance action:
>  >> if a datastore is already on a removable device in your pocket, and you
>  >> need to access something on a foreign system, you are stuck unless there
>  >> is some minimal level of human interpretability of the file system...
>  >>
>  >> Instead, you have to dig up a system with FUSE/olpcfs installed, and
>  >> then copy the files to a conventional file structure.
>  >>
>  >> This is the use case that's hard to get around.
>  >
>  > Ok, I think I see now where is the misunderstanding.
>  >
>  > In the first post in this thread, I tried to explain that this
>  > proposal would use removable devices in the same way they are used in
>  > other systems and that the DS would have nothing to do with them:
>
>  I think expanding the space available to the DS through usb devices or
>  sd cards is a use case we should take in consideration when designing
>  the DS, even if we don't plan to support it right now.

Sure, but we already know that we need a DS that supports different
backends with different on-disk layouts, right? Unless we can find an
equally efficient, robust and "transparent" layout that works in all
file systems including vfat.

In many areas, we have over-engineered because we tried to achieve
long-term goals in a short period of time with very few resources.
This has been one of the reasons why we have failed to deliver a
satisfactory solution to more basic goals.

I'm not against discussing the future, but I would like to advocate
for doing the basics well before investing too much effort in the
"turing prize features".

Thanks,

Tomeu