[Opensim-dev] Deprecate OpenSim.Grid.InventoryServer and OpenSim.Grid.AssetServer?

[Justin Clark-Casey]

Melanie wrote:
> Hi,
> 
> this has been spoken about extensively, mostly on IRC. It is NOT 
> about _changing_ OGS1. It's about replacing it.

fyi, as has been said many times in the past, speaking about something on IRC does _not_ mean that everybody knows about 
it.  Peer review must occur either via mailing list or around a wiki page, at the very least.  Wiki page is preferable 
because they provide a summarized, easily understood and referenced proposal.  Mailing lists often provide only an 
incremental picture.  Discussing a large proposal only (or largely) on IRC is not valid and can get overturned or 
questioned.

> 
> OGS1 is a monolithic juggernaut that stubbornly resists evolutionary 
> pressures. It can't be updates any more than the dinosaurs could be. 
> The smaller, faster mammals (services) will push it out of the 
> picture. This has already happened for assets and inventory.
> 
> The basic structure is as Diva describes:
> 
> IN_Connector -> Service -> Out_Connector
> 
> where each is optional.
> 
> So, a region will normally load either a service (standalone mode) 
> or a out connector, which will then connect it to an in connector on 
> a server that then loads the service.
> 
> The interesting part is that the system of connectors allows up to 
> design more sane wire protocols - but not if we have legacy servers 
> around that don't subscribe to the connectors system and lock the 
> present wire protocol in place.
> 
> So, as B.U.S.T. gains implants for a protocol (using a compatible 
> wire protocol plugin at first), the corresponding legacy server 
> needs to go.
> Then new protocol plugins can be developed that are more sane, allow 
> more functionality, have more security, or whatever other 
> improvements. The servers will be able to follow such protocol 
> changes in mix & match fashion by using appropriate connectors.

I don't see any problems with this, providing its documented (as you have now done) and reasonably extensively tested by 
others (which as far as I can tell is not the case yet, hence the concerns of Charles and others).

But the real question was about your statement

"But changes are planned as we are moving to more sane protocols."

source: https://lists.berlios.de/pipermail/opensim-dev/2009-July/006992.html

Who is the 'we' in this?  What are these protocols?  Why are they more sane, etc., etc.?  This is an entirely different 
question to generalizing the OpenSim grid servers.  Perhaps they were not meant to be mixed up in this.

> 
> Melanie
> 
> 
> 
> Justin Clark-Casey wrote:
>> Mike Dickson wrote:
>>> On Wed, 2009-07-08 at 14:48 +0000, Melanie wrote:
>>>> Charles, I said documentation is on it's way. I would like to get 
>>>> people's opinions on moving forward with deprecating the old 
>>>> servers, not people's opinions on the obvious need for documentation.
>>>>
>>>> Further, if the old servers are kept, they will have to be 
>>>> maintained indefinitely and so will the protocols, blocking any real 
>>>> progress. So, could you please explain your -1?
>>> You mentioned (and Justin I think asked for details) on protocols changing. 
>>> Which protocols are you refering to (inter-server protocol, client-server protocol)?
>> Yes, you (Melanie) have mentioned 'new' protocols but afaik this is the first I've heard of this.  Changing OGS1 
>> protocols is a big upheaval and should be properly laid out and peer reviewed before work starts. not after or during. 
>> Going down the wrong path now will be expensive later on.
>>
>> If you can point me towards existing documentation on this (if it's part of BUST) that would be appreciated.
>>
>> Thankyou for writing the existing documentation on BUST.  I'm sure if the new technology is well received then we can 
>> deprecate and remove the old servers.
>>
>> I also don't like the name 'BUST'.  As we've already *very extensively seen*, we should not assume a sense of humour on 
>> the part of others.  A compromise would be to change the server prompt.
>>
>>> In general I'm +1 for this since I'd really like to see the core settle
>>> down into a stable state and getting the refactoring behind us is a
>>> prereq. for that.  I certainly understand the desire for doco so people
>>> know what they're being asked for feedback on however.
>> Absolutely.  Asking for a general vote on progress is like proposing a vote on motherhood and apple pie.  We can only 
>> make meaningful votes on specific proposals.
>>
>> I'm actually now confused on whether you're asking for a vote on eventually deprecating the old servers (but not before 
>> another vote) or on changing basic grid protocols.  I assume it is the former rather than the latter.
>>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
> 


-- 
justincc
Justin Clark-Casey
http://justincc.wordpress.com