[Opensim-dev] Alternatively Identifying Assets using URL's

[Hurliman, John]

The Cable Beach AssetInventory server in OpenSim should support this. mikem would know the exact status of fetching assets over HTTP, but (IIRC) the format is: http://www.assetserver.com/[uuid]/[data|metadata] to get either the raw asset data or metadata about the asset including a link to the asset. The normal procedure would be to grab the metadata first, since that contains information such as layer boundary ends for JPEG2000 files and a link to the asset that may point to a content distribution network or some other accelerator.

John

>-----Original Message-----
>From: opensim-dev-bounces at lists.berlios.de [mailto:opensim-dev-
>bounces at lists.berlios.de] On Behalf Of Mystical Demina
>Sent: Friday, May 01, 2009 8:00 AM
>To: opensim-dev at lists.berlios.de
>Subject: Re: [Opensim-dev] Alternatively Identifying Assets using URL's
>
>I have started working on the Xenki viewer which is a web base viewer
>and I can see wanting to support this.  Right now Xenki is based on XAML
>and XBAP.  I have been playing with Silverlight based web client as well
>but that is a bigger effort but where I like to get long term.  But both
>of these I can see getting images using a URL so I can use the browser's
>image caching and simply texture management.  I suspect we can create
>URLs based on http://www.assetserver.com/AvatarUUID/AssetUUID or some
>kind of CAPS concept, so the server can see if this avatar is logged in
>and has permissions to retrieve that image.  But I can also see if we
>can move away from UUID to file names then that opens up to using every
>image in the internet not just images that are uploaded into the grid.
>
>
>
>If we have something working let me know I would be happy to play with
>it.
>
>
>
>Kevin Tweedy
>
>IRC: Mystical
>
>
>
>
>
>________________________________
>
>From: opensim-dev-bounces at lists.berlios.de [mailto:opensim-dev-
>bounces at lists.berlios.de] On Behalf Of Stefan Andersson
>Sent: Friday, May 01, 2009 8:36 AM
>To: opensim-dev at lists.berlios.de
>Subject: Re: [Opensim-dev] Alternatively Identifying Assets using URL's
>
>
>
>It is my impression that adressing by uri/url has been a goal for the
>last 1.5 years or so.
>
>I'd be interested in hearing if there's actually anybody _not_ wanting
>us to work towards that.
>
>The only issue that I can see would be in regards to SL compatibility. I
>have started to poc an approach where guid-based adressing of resources
>is considered a thing purely between the viewer and the content endpoint
>(aka the region) - that the region acts as a 'proxy' translating guids
>to uris and back as needed. So far the code is merely stubs, so I
>haven't got to the point of proofing it, but I believe you all
>understand what I'm saying. Have a look at what uses the fairly new
>ContainsReferences method on AssetBase to see where I was going.
>
>Best regards,
>Stefan Andersson
>
>
>
>
>> Date: Fri, 1 May 2009 08:56:57 +0300
>> From: tommi.s.e.laukkanen at gmail.com
>> To: opensim-dev at lists.berlios.de
>> Subject: [Opensim-dev] Alternatively Identifying Assets using URL's
>>
>> If we do not consider SL restrictions for a second and look to the
>> future it would be nice to discuss the following which popped out from
>> dahlias idea of supporting loading textures from websites directly and
>> Stefans suggestion of using strings as ids in the standard format
>> thread:
>>
>> It is good to consider supporting identification of assets by
>> traditional urls. Some public asset servers could be www servers or
>> svn http servers. As all 3d modeling programs are refering to
>> aggregate assets using file paths, it would be easy to upload the
>> model to SVN http server and then using viewer pull out the model and
>> aggregate assets for viewing. This could be alternative asset
>> distribution model. it could work through region asset cache or
>> directly from viewer to the http server. This could be much more
>> scalable than the current model. I realize this is a threat to the
>> data consistency as models and aggregate assets are not saved to
>> relational database which could safeguard relations. It is also viable
>> only for completely public content but I assume considerable portion
>> of the vw content is likely to be freely distributable.
>>
>> This would be in practice support for "importing" content from web on
>the fly.
>>
>> -tommi
>> _______________________________________________
>> Opensim-dev mailing list
>> Opensim-dev at lists.berlios.de
>> https://lists.berlios.de/mailman/listinfo/opensim-dev

>