Fwd: [sldev] Re: Lookup Tables and Texture Bugs

[Mathew Frank]

I think we are at the point of having hashed out most of the ideas, and 
potentially start forgetting some of the earlier ones.  (such as the LL 
proxy idea, which was one of the first I raised in my list of options)

Does anyone else agree with putting all these options on the wiki so we 
can move to the next step, which would be rating them all in terms of 
usefulness and difficulty, so a proposal can then be produced?

I do think that a "darknet" P2P proxy system ( I would suggest just 
using TOR if it weren't so damn slow - usually 30 seconds for web page 
whenever I used it over a period of a month or so) for some of the 
information is a useful (and transparent) solution to a lot of this - 
without producing more load for LL anyway - for a start surely it would. 

I actually am less interested in having to create textures in LSL 
against external textures on my own server.  One reason is that I can an 
unlimited amount of complex things on my own server without consuming 
any script resources on the SIM server. 

Something else occurs to me, that will allow quite paranoid settings for 
a user but not limit them too much, is to show HTTP based textures in a 
different colour grey, and all the user to click on the texture to 
enable loading of that particular texture.  So this would man that a 
user interface texture for example could still be used by somebody with 
all HTTP textures turned off otherwise - with a mouse click.

Cheers,
Mathew


Zack Geers wrote:
> From: *Zack Geers* <kunnis at gmail.com <mailto:kunnis at gmail.com>>
> Date: Jan 25, 2007 4:57 PM
> Subject: Re: [sldev] Re: Lookup Tables and Texture Bugs
> To: Gigs Taggart <gigstaggart at gmail.com <mailto:gigstaggart at gmail.com>>
>
> There is a certin amount of Privacy that users expect.  We shouldn't 
> do anything that moves away from their current set of expectiatons.  
> IP's don't matter, it's the associations between accounts that do.  I 
> feel we should start this conversation by actually figuring out what 
> our requirements are.
>  
> 1) What level of privacy should users expect?  If it's too high, then 
> you'll never get any features.  If it's too low, then there will be 
> all sorts of things like focused marketing, spyware like prims, 
> association of alts.
>  
> I bet for example, LL employees have their own private accounts so 
> they won't get pestered constantly.  I doubt they will want it well 
> known who their alts are.  There are also some famous people that 
> play, and they would be major targets.  Right now you have to go onto 
> someone's land to even be prompted for a stream download, and even 
> that is not guarenteeded to happen, only if the user starts the media 
> are they even exposed.
>  
> This idea would let someone farm data for almost nothing.  It would be 
> simple to rent space in several popular malls, and farm data that 
> way.  The player density is low, and you could put a couple opaque 
> objects on each side that would obscure the view of the object till 
> they are close.  Place several dozen objects in diffrent malls and 
> you'll probably stumble into nearly everyone on SL, not to mention 
> having some bots float around, randomly switching the object they are 
> wearing to figure out who the person that's in LOS of them.
>  
> Plus a good bit of cross corliation, and you'll learn what everybody's 
> IP address is.  If you now ABC downloaded the image when they were 
> near you, then you wait for one person to leave, and switch the 
> texture, you then know who one of the people is.  Repeat a bunch, and 
> you easily map out a lot of IP addresses.
>  
> I think that prompting the user is NOT the way to go.   Popups are 
> horrible UI design, and users blindly click through popups.  Yes, they 
> are stupid for blindly clicking through popups, but it doesn't matter, 
> that's what they do, don't say they shouldn't, we're talking about 
> reality, not what it should be like, we have no control over how users 
> act.  You can't change the world, you have to work with it.  They 
> confuse users, and will only make the game harder to play.  What do 
> you think it's going to be like when everyone has a dynamic image on 
> their shirt, and you walk into a populated area and get hit by 30 popups?
>  
> For http images hosted on a texture, what about making it so that the 
> textures are fetched by LL, and hosted on their servers?  This would 
> remove most of the privacy issues, but still allow dynamic image 
> generation.  One of my biggest concerns about something like this is 
> that people will be hosting textures all over the internet, and there 
> will be a lot of support issues due to the fact that "Well this should 
> work..." and "..it works for me" because of diffrent people's access 
> to the internet.  If we get a stable connection to LL, then LL could 
> host all images for us, and it'll be less cranky.  What I'd sorta like 
> to see is to be able to create an image programicaly through LSL, 
> because eventually we'll have .net code running in LSL, and you could 
> then do a mixture of a HttpRequest and a convert to image.  I feel 
> this solves some of the problems, but not all of them.
>  
> I think this will be good for some cases, but not nearly all of them, 
> I don't think it's a perfect solution, but it's the best I've heard so 
> far.
>  
> Kunnis Basiat
>
>  
> On 1/25/07, *Gigs Taggart* < gigstaggart at gmail.com 
> <mailto:gigstaggart at gmail.com>> wrote:
>
>     Grr... Why doesn't reply reply to the list??
>
>     ---------- Forwarded message ----------
>     From: *Gigs Taggart * <gigstaggart at gmail.com
>     <mailto:gigstaggart at gmail.com>>
>     Date: Jan 25, 2007 5:08 PM
>     Subject: Re: [sldev] Lookup Tables and Texture Bugs
>     To: Rob Flickenger < rob at hackerfriendly.com
>     <mailto:rob at hackerfriendly.com>>
>
>
>
>         If SL ever expects to transition from a service to a platform, it
>         will have to confront this problem sooner or later, because it
>         doesn't just affect dynamic textures.  Any peer-to-peer
>         service that
>         is likely to be implemented in the open source client (VoIP, file
>         sharing, direct chat, and eventually users hosting their own sims)
>         will have precisely these same issues.
>
>         Personally, I think it's the user's responsibility to protect
>         their
>         own communications (assuming it's important to them) and I have no
>         sympathy for people who "need" to hide behind alts yet don't
>         understand the technical issues.  Update the ToS to mention
>         that your
>         IP is revealed in certain circumstances, let's implement these
>         features, and move on.
>
>
>
>     Yes, exactly.  We need to think big picture here.  The sooner
>     people realize that IP addresses aren't private information, the
>     better.
>
>     It was bad enough that LL didn't make it clear to people that the
>     DRM didn't work on anything but scripts.  Giving people a false
>     sense of security is a terrible thing!  People shouldn't have an
>     expectation that their IP won't be exposed.  The more effort we
>     put into providing that illusion, the more paralysed SL becomes as
>     a platform.
>
>     There are whole classes of features here that we would be throwing
>     out the window if we start down the road that IPs should be private.
>      
>
>     _______________________________________________
>     Click here to unsubscribe or manage your list subscription:
>     /index.html
>     </index.html>
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Click here to unsubscribe or manage your list subscription:
> /index.html
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.secondlife.com/pipermail/sldev/attachments/20070126/d50f13e8/attachment.htm