SecondLife, Graphics and Usability
Submitted by Taran on Thu, 02/08/2007 - 09:51
Reading through the morning's slew of stuff, I came across Yet Another Blog Entry About SecondLife Graphics - which is actually a really thoughtful entry on the use of the Quake Source Code, and just happened to be where this tangent came from (in other words, its a good post and you should read it - but not in this context). I've seen plenty of posts say that SecondLife graphics suck, this one just triggered this response.
Isn't it odd that people who want to improve graphics don't talk too much about usability? And why is it that people who talk about usability don't talk much about graphics? Are they related? Maybe.
I think it is fair to say that those who point out that SecondLife has poor graphics have higher expectations - and those expectations probably come from applications (probably games) with higher end graphics ability. These folks are pretty computer literate as things go, and probably have no problem adjusting for the SecondLife interface.
I also think that it is fair to say that those who have problems using the interface aren't as computer literate as the people who gripe about graphics. Yes, I know, they might be whizzes at using office software but can't we agree that computer literacy transcends pressing the big button to get some cheese?
What does it mean? I don't know. It does seem to demonstrate a difference in the reasons why SecondLife is criticized. Me? I find the graphics good enough for now, and I find the interface usable.
I think overwhelming people with gamer-like detail is all well and good, but in an environment created by users there simply have to be limitations so that the whole thing runs smoothly. In games and virtual worlds of higher resolution, this is built in since you don't have users leaving a trail of wooden prims as they scamper across a simulator - and when the users figure out how to texture the objects beyond the standard wooden prim, lookout!
There's a reason that there are limitations on the textures. I believe that presently, the largest texture resolution viewable is 1024x1024. 512 x 512 is what is typically suggested. It takes less bandwidth to transmit, occupies less space in cache, and in an environment where users can add textures for 10 Lindens - that is all important.
So yes, maybe the graphics could be better - but that would probably limit who could use SecondLife (more so than it does now). Would better graphics make SecondLife more usable? I don't think so. Are there things which could be done to make SecondLife more usable? The learning curve less steep?
Maybe. But graphics really isn't a part of that problem. A few pixels here and there isn't going to help people walk around better.
Now if you start talking about the physics engine, well... maybe.
thanks for support
Great article
I hope everybody read this article
no so great
So you've come to no conclusions then about the graphics or the UI?
I've played games since Wolfenstein 3D first started the whole craze and I tell you, beyond mouse look, not much has changed with most FPS.
Now I've logged into Second Life twice. Both times for over an hour and I still don't have a grip on WTF is going on with the controls. It doesn't help that its really laggy on that 1st island that you encounter (I guess coz of all the other people there, why not go Matrix and just have a few rooms all to yourself to begin with. Hell, have then as part of the software so you don't even have to be online!). Its hard for people to get a feel for anything if they make a movement and then 10 seconds later they are looking at something completely different and not sure how their input has affected the change. In other words make it faster and it will be easier for people!
Most 3D games that deviate from the standard controls have a orientation mission with a guide that walks you though each step. Each step being designed to teach you a feature of the UI and/or the environment/game play.
As for the graphics, most games also let you adjust the level of detail with the resolution/texture detail/polygon count/gamma.
I think 2ndLife could take a few lessons from the FPS, they have been doing both graphics and UI for a long time before 2ndLife showed up.
Dam, thats almost another post...
Umm....
The post wasn't about coming to conclusions, but about pointing out who gripes about what. As you have indicated, you're pretty experienced with 3D environments. Thus you discuss graphics. You support what I said.
That said, the controls of SecondLife aren't notably different from those in any 3D game. Your arrow keys move you around in the way that 3D environments have always done. Up arrow - forward. Left arrow - rotate left. Right arrow - rotate right. Back arrow - back up.
PageUp and Page Down allow you to fly and land. That might be different from some other 3d controls.
What have you been playing that is so different?
As for the graphics - yes, you can adjust the level of detail. Its in the preferences.
As far as the graphics detail - well, since almost every 3D game already has a texture set which is cached, that's a big advantage. In SecondLife, users can upload their own textures which means that the system has to accomodate everyone. Apples and oranges. ;-)
Yes and No
How can I support what you say if you are not saying anything i.e. no conclusion(s)? But anyway...
My post isn't actually about graphics, thats a side note, its about bringing an open source code base together with an open standard to make a _free_ platform for everyone. One of the supporting facts for doing so was graphics, another was usability. I'm discussing both so I'm not in ether of the "two kinds of people" who grip about precious 2ndLife.
But as for the physics engine, I find this the most boring of all. In a universe where anything is possible, why make it the same as the limiting one you just came from?
Here, Let Me Read for you:
'Isn't it odd that people who want to improve graphics don't talk too much about usability? And why is it that people who talk about usability don't talk much about graphics? Are they related? Maybe.'
You support that with your responses. That you choose now to discuss usability means nothing - I was responding to your comment which did, indeed, support that observation.
And the link to your article comes with text which quite clearly states that it is interesting, and that it caused a tangent.
But anyway...
Second Life: How to Implement an Insanely Great Idea Poorly
I really don't see this as a user type issue at all. The graphics are low quality 15fps max and the interaction is extremely rubber band problematic with severe usability issues in navigation / search / travel history / messaging / etc.. People use the experience in spite of how technically poor it is in every conceivable way. Essentially Linden has created this extremely important metaverse gift for humanity yet the product itself is technically very poor.
To really address the majority of all user type issues Linden should be setting a few simple limitations on object editing capabilities. It is absolutely insane and to real time render tens of thousands of inanimate objects from dirt to houses to plants to light posts. As a developer I really feel their CTO and development staff have done a massive disservice for the company by not making this complete lack of graphic object caching obvious to management.
If Linden really wants to address their graphics and interaction problems they should set a few simple rules (70% of your prims are not editable unless you put the land into design mode). Reducing render calculations by 2/3 would allow extensive improvements in frame rate / graphics richness capability.
As for interaction... if you eliminate even 50% of all that real time GPU network traffic I'm sure a few more movement syncs would be possible. Don't even get me started on their decision to write a low level C++ app which is slow to iterate and dips so far down into proprietary GPU stacks that they can't support Intel GPUs? Thats pretty much the majority of mid-priced laptops being sold with Intel integrated chipsets.
In short, they have themselves to blame and I think your comment assuming it is one type of person or another is really missing the point that the implementation is very poor. They have created the most amazing concept I've ever seen yet their failure to place limitations on anything is leaving plenty of room for competition. Its really sad because I want to buy an island myself! :-)
Post new comment