Slashdot Mirror
Second Life Hype vs. Anti-Hype
The new GigaGamez site, part of the OM network, has a look today at the hype fight over Second Life. It's the new darling of media companies, but is increasingly attracting negative feedback by people who know a thing or two about the industry. James Wagner Au tries to sort out who is saying what, and provide a little context for the discussion. From the article: "Can they really build a fully streamed world comprised of tens of thousands of servers? That's way above my paygrade, but I'll guess that task fits under the rubric of Fricking Hard. Can they fix a profoundly unfriendly user interface and thoroughly disorienting first hour user experience, which are aggressively, almost intentionally unwelcoming to the vast majority of interested users? Both shortcomings are at the heart of Second Life's poor retention rates, but neither have significantly changed in the three years since its commercial release. You have to wonder, whatever their stated intentions, if Linden's tech-centric corporate culture simply puts their improvement at a low priority."
Don't we wail about Newbies everywhere else? There could be a side benefit that only certain people "get it" and stay. Anyone who doesn't ... "doesn't deserve to be there".
In this case, I think the problem is the interface is terrible and clearly designed by someone who doesn't 'get it' - personally I find it easier to create my own objects in OpenGL in my own code than using their interface to do the same thing, it really is that clunky. Even moving around is painful thanks to floaty controls and because the client side collision detection is terrible. That's some pretty basic stuff to have nailed down, I would be embarrassed to have written this software.
Added to that the 'real time streaming' aspect of it is dire too. Either their peering/transit isn't that great, or, I suspect, their systems often just can't handle the load. The client does nothing to help matters but not caching sufficiently, and being very slow to read from cache when it is. It doesn't handle drawing-in objects in the world well either, no attempt to fill them in nicely (when it could draw a construction banner around them while it loads a building or given area, for example).
Lastly, the rendering performance, features and quality of 3D engine are all poor! Even on an AMD-FX CPU, a couple of gigs of DDR RAM and two 7800 GTX cards in SLI it's still slow as hell (even without it really bothering to do things like collision detection). God knows what it's doing, it performs like a developers 'my first DirectX 3D environment' run on an old Pentium 2 Celeron (really, genuinely).
The whole thing is an over hyped marketing exercise of no technical merit IMO, and it's never going to take off because it's so badly broken (unless it's entirely re-worked, to the extent it's a new product - but I figure that's highly unlikely). Star Wars Galaxy with it's player built cities w/ deforming terrain (but fixed 'craftable' items) was far better executed technically, and even it wasn't all that great in that respect (for example, it too suffered from pop up because they didn't bother to put in client side caching for player placed world objects).
It certainly is an exclusive community - essentially just of people all trying to make money selling goods, mostly to no-one and (believe it or not, it's really true) a large furry community, for some reason I can't fathom. I don't think may people who arn't trying to make money on it are interested in using the software beyond a brief trial for a couple of days.
Personally, if they didn't know how to write their own engine (which they clearly didn't) I think they should have done what both World of Warcraft and Lineage 2 (and 3) have done and just licensed the Unreal engine (or used similar). Then content creators could have created models in a decent external modelling tool, allowing them to add sophisticated animations to them. The collision detection would work, the performance would be great and getting the controls right would have been easy (just a case of 'tweak a the example code from the engine vendor'.
The Metaverse concept of Second Life is still attractive, but the Linden Labs implementation is just stupidly bad. Same problem with Active Worlds and a host of other similar offerings. It's like no one in these company has ever tried an MMOG (or indeed used any interface featuring standard click to move / WASD controls). I can understand a lack of polish, but the consistent inability of these companies to get the fundamentals like movement right is something I find mind boggling.