Re: [Gnu-arch-users] Automatic archive discovery, take 1

[Aaron Bentley]

On Mon, 2003-12-15 at 15:15, Tom Lord wrote:
> 
>     > From: Aaron Bentley <address@hidden>
> 
>     > I also agree that namespace issues are always subject to the possiblity
>     > of corruption/dispute.  Adding another layer of indirection cannot fix
>     > this, only user intervention.
> 
> On the contrary.   A level of procedural (and perhaps naming)
> indirection is invaluable here.

Err-- but I differentiated between "indirection" and "user
intervention".  I prefer the "user intervention" form of "procedural
indirection", and actually suggested using it for the case of multiple
(conflicting) matches.

I was referring solely to the idea of adding yet another namespace on
top of address@hidden

> One idea is that whenever tla sees an archive name it doesn't
> recognize that it calls a single procedure to resolve that name.
> 
> Another idea is that whenever tla sees an archive name it doesn't
> recognize an error is signaled.   Users have to define archive names
> before using them, possibly using distinct procedures to define
> different parts of the namespace.
> 
> Either approach can emulate the other but the latter approach avoids
> having tla itself encourage a situations where users wind up taking
> the former approach simply because they copied some of their
> ~/.arch-params/hooks from a random post or web site.

Intermediate approaches are possible too:

archive not registered: address@hidden
Do you wish to register
ftp://ftp2.example.com/{archives}/address@hidden as
address@hidden

> The latter approach is just as vulnerable to corruption as the former
> -- but doesn't itself invite that corruption.   Indeed, it invites 
> a recitation of what the issues are.

I'm not sure that going to the user buys you a lot.  If corruption
happens, recent information will probably start using a new archive
name.  In those cases where a user is affected by corruption, it's more
likely that corruption will happen after the user registers the archive,
not before.

But I don't have strong feelings about it.  If you feel any kind of
default-registration mechanism is too risky, it won't affect me much. 
I'd just prefer that if such a mechanism *was* implemented, it could
avoid adding new central authorities.

Aaron

-- 
Aaron Bentley
Director of Technology
PanoMetrics, Inc.