[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: SEUL: X/Berlin/db, and reinventing wheels
> Definitely. If we want to get this out this year, Berlin is not the way
> to go.
I think Berlin would be great, but after reading through today's backlog of
messages I have come to realize that there are quite a few roadblocks to
using it:
1) it has to be coded to 100% completion (unknown ETA)
2) it has to be proven/stable
3) it has to be able to run anything we can run today in X
#3 requires either a massive porting effort initially, or over the long haul
if we can get X compatibility libs. Compat libs would be a step in the right
direction, but would likely not be a 1:1 map to Berlin, causing unknown
problems. Besides, they don't exist.
I say stick with X.
> I've mentioned similar things before, but I would support a registry
> only if ALL of the following were true:
>
> 1. It would keep track of exactly which program added exactly which
> entries to the registry, enabling #2 and #3 (so we don't have another
> Doze registry fiasco)
> 2. There would be a provided utility to nuke all entries made by any
> given application (so we don't have another Doze registry fiasco)
> 3. Another utility to copy the entire registry, any part of it, or all
> entries made by any given application to a file, and the ability to
> restore that file to the same or a different computer (so we don't have
> another Doze registry fiasco)
> 4. Has to be FAST (so we don't have another Doze registry fiasco)
> 5. Would need to be somewhat small, only used for things that NEED to
> be in a central registry, and well organized (so we don't have another
> Doze registry fiasco)
I think it'd be relatively easy to code something like this. Use gdbm (since
it exists, works, and is relatively simple), provide a simple set of library
calls to deal with it, and we're set. We can define a namespace for
applications that can guarantee #1, and provide a control-panel component to
control the contents of it, including merges and backups, etc.
Following the Doze idea a little, we'd have a registry for the whole system,
then a registry for each user. Or is that a bad idea and someone has a
better one? I think it'd work, besides, it's kinda what X does with resources
anyway.
Oh, 'nother thought on the gdbm thing - perl (and likely python as well)
has a *direct* interface to it. That would make things significantly easier
to do, as it simply shows up as a hash in perl...
TTYAL,
Omega
Erik Walthinsen - Webmaster and infrastructure for SEUL __
__ / /\
/ \ omega@sequent.com Work: (503)578-5314 / / \
| | M E G A omega@aracnet.com Home: (503)281-4281 / / /\ \
_\ /_ psu12113@odin.cc.pdx.edu Majoring in CS / / /\ \ \
/ /_/__\ \ \
SEUL: Simple End-User Linux - creating a Linux distribution /________\ \ \
http://www.seul.org/ for the average home/office user \___________\/
----------------------------------------------------------------------------
Simple End User Linux Mailing list
To be removed from this mailing list send a message to majordomo@txcc.net
with the line
unsubscribe seul-project
in the body of the letter.
----------------------------------------------------------------------------