dear all, sorry for the lag, i've been at OSCON all week with no
computer...
Daniel this makes a lot of sense to me. Yes, geodata.osgeo.org right now
is our project site / project pages. As a project chair I am 100% not
happy with what we have on CollabNet, and would *like* to redirect
geodata.osgeo as soon as we can find an alternative which also allows
us to start managing mailing list+archive hosting. But this is a
discussion for another time... 
At the last geodata meeting we agreed to use a subdomain of
geodata.osgeo.org to point to what is currently
geodata.telascience.org - dev.geodata.osgeo.org ,
find.geodata.osgeo.org , phat.pile.of.geodata.osgeo.org etc, and i
will submit support ticket requests through https://support.osgeo.org/
to get new DNS subdomains set up - i have been advised by auke that
this is the best thing to do and SAC will be kept in the loop.
cheers,
jo
On Tue, Jul 25, 2006 at 11:28:19AM -0400, Daniel Morissette wrote:
Frank Warmerdam wrote:
>Benjamin Thelen wrote:
>
>>Daniel Morissette schrieb:
>>
>>>I see that our machines at telascience are given names directly under
>>>the root of the telascience.org domain, without any distinction from
>>>any other telascience servers.
>>>
>>>To keep things organized and keep our stuff distinct from other
>>>telascience servers, should we use a separate domain (e.g. osgeo.net
>>>which is currently unused AFAIK) or at least a subdomain of
>>>telascience.org (e.g. osgeo.telascience.org) for all those machines?
>>>
>>>I apologize if this has been discussed before I joined the group.
>>>
>>>Daniel
>>
>>
>>
>>According to a mail between Frank and auke@collab.net, some days ago,
>>it is easy to assign subdomains of osgeo.org, thus I'm a suprised,
>>that for example geodata.telascience.org has been assigned to
>>198.202.74.216, instead of geodate.osgeo.org. Did I miss sth.?
>
>
>Daniel / Benjamin,
>
>I believe that John added geodata.telascience.org as a convenience.
>We can't use geodata.osgeo.org for the machine yet, since that is
>delegated to the CN project pages for the geodata committee.
>
Even if you had full control on osgeo.org I would advise against using
geodata.osgeo.org since the top level under .osgeo.org is already used
for project names so we should not pollute that level with machine
names. My opinion is that we should try to plan a little bit for the
future and put machine names under their own subdomain, either osgeo.net
or osgeo.telascience.org as I previously suggested, or use a sub-domain
of osgeo.org if we want machine names under that domain, for instance
.ts.osgeo.org (ts being short for telascience). A subdomain such as
.ts.osgeo.org could be delegated to a DNS server managed by John or
someone outside of CN to keep things real simple. If we don't start
putting some order then the .osgeo.org domain will be quite a mess in a
few years.
The same way I think we should come up with a standard for assigning
user ids for people accessing the servers (I believe that's done in the
LDAP server?) instead of letting everybody use a nickname, otherwise
we'll end up with a mess there as well. For instance, using first letter
of firstname + full last name seems to be quite common and could do
(e.g. dmorissette, fwarmerdam, etc.). I'd be open to any other
convention as long as we have something.
I don't claim to have a perfect solution, but I'm a bit worried by the
lack of order in the way we do things right now. My experience is that a
naming convention for domain/server infrastructure is as important as a
coding style for code: it will make things much more predictable,
manageable and usable in the long run.
Sorry for throwing another 0.02$ on this, hopefully it's the last time...
Daniel
--
Daniel Morissette
http://www.mapgears.com/
--
ghug is my email archiving bot. if you see it cc'd on this email,
please leave it cc'd, that will help me a lot. http://frot.org/ghug