I usually prefer to read others predictions rather than make them. The Dolphin Poll on users’ preference got me thinking and I have not seen many predicting a lot about where viewers are going. Of course that may be because it is a foregone conclusion. But, since this is my blog, I get to voice my opinion and ramble or rant on about it until you click away from boredom. Hopefully not. In this case I think I can put things in terms you’ll find interesting.
I think the viewer development path using series 1 viewer code is a dead end. At some point viewers based on series one code are going to go away. I think this year.
I think the series 1 User Interface (UI) will last longer may be even into next year, but it too will disappear. Simple, end prediction. Following are my reasons and thinking for why I believe this will happen.
All the sources I have access to say the Second Life Viewer 1.23 (SLV1) code is a mess. It is not modular. To me that means there is lots of spaghetti code, code that has interdependencies and sections that are so poorly documented no one is sure what they do or what depends on them. Such code is notoriously difficult and expensive to maintain or modify.
Part of the reason Linden Lab (LL) paid to have SLV2 developed was to get clean, modular, and documented code. Regardless of how well the SLV2 UI succeeded or didn’t we have some information from Linden comments that the SLV2 code still was not as modular as it could have been. The Lab is working on making it modular enough the UI and render code can be separated. Separating the UI and code is something that will not be possible, at least within rational limits, with SLV1 code.
The user interface is a controversial topic. I think all users will agree both SLV1 and SLV2 user interfaces have problems. Changes need to be made. Many of us are hoping Third Party Viewers (TPV’s) will make brilliant changes to the UI and make something easy to learn and use. I even have hopes for LL to do better.
To enable significant UI changes the code needs to be modular. The lack of modularity in SLV1 code prevents that and contributes to SLV1’s ccoming demise.
People are notorious for hating change, not everyone, but enough. I think Emerald was popular because it was an extension of the SLV1 UI. It was mostly the same UI as SLV1. A resident could switch and have a small learning curve. I think that is a huge factor contributing to the popularity of Emerald (now Phoenix), Imprudence, and other TPV’s.
New residents coming into Second Life now learn the SLV2 UI first. A change to Phoenix, Imprudence, Dolphin, and other series 1 viewers with the old interface presents them a large learning curve. We already see posts in the SL blogs and forum about people giving up on SLV1 style viewers because they can’t figure out how to do things they do in the SLV2, too much brain damage.
So, the learning curve wall is going to be a major growing obstacle to adoption of any viewer continuing to use the SLV1 UI.
We have some number of players leaving SL every year and moving on to other games and activities. As time passes there will be fewer residents that understand the SLV1 UI. This means decreasing user base for any viewer using the SLV1 UI.
Features & Speed
TPV’s grew in popularity because they offered features not available in the SLV1. TPV’s were more stable than SLV1 and presented fewer problems. That was and is two good reasons to adopt a TPV. Now that is turning around. There are new features coming out in SLV2 that are not available in TPV’s. The cutting edge crowd made up of early adopters has to use SLV2, at least enough to understand new features. Think mesh clothes, new skins, new styles of makeup, and outfits. TPV’s are having problems with the new texture compression and download processes, appearance editing, Texture access, and inventory access. Textures upload with the alternate OpenJPEG process are probably a cause of all sorts of issues. Whatever the case, stability is now a problem and there is no clear best solution.
It takes considerable effort to back port SLV2 code to SLV1’s. We see a delay in getting new features in from SLV2 to SLV1 type TPV’s. Conversely, we see few of the neat features in SLV1 viewers making it into SLV2 based viewers. LL is not adopting that many of the new TPV’s features. So, there is opportunity for TPV’s to make feature rich viewers.
The render pipeline in SLV2 viewers is being improved. The SLV2 code is modular enough that KirstenLee is developing a new render pipeline that allows excellent performance and shadows. I doubt it is going to be possible to port such improvements to SLV1 code. This means as mesh begins to appear as a basic component of region building old SLV1’s will become unusable. They do not render mesh.
Until one is using SLV2 code and developing new features for that code, the workload to fit pieces together and create stability and compatibility in SLV1 code is so challenging the process is slow. SLV1 viewers are going to lag farther and farther behind. I think that means fewer and fewer people will use them.
I suspect many of the non-technical people don’t really care about the code base. Their interest is in what they can do, how easy the viewer is to use and learn, and if it is stable and works. The UI is paramount to them. If they learned the SLV1 UI then they mostly prefer that UI and would rather not change. The same is true of those that learned the SLV2 UI first. This means the SL community is dividing in their viewer preference. Time is against the SLV1 UI.
I think we will see all the TPV’s change over to SLV2 code this year. It is just too much work not to do so. Those TPV’s that try to stay with SLV1/Snowglobe code will fade from popular use. Viewer problems and slow addition of new SL features will make them less and less popular.
I think we will see most TPV’s adopt more of a SLV2 style UI. I expect a radically different UI from at least one of the TPV’s. But, any such UI will have to have major advantages to attract users and hold them. The SLV1 UI is going to fade away over the course of the year. It may take longer than a year. It will also likely hang around longer in the OpenSim arena.
Users faced with a radically different new UI will have to decide whether it is worth learning or not. Many have stayed with the SLV1 IU to avoid learning the SLV2 UI. I think faced with such a choice it is likely the users will split with some part choosing to learn the new UI and others deciding it is time to learn the more common SLV2 style UI. The SLV1 to SLV2 experience is showing us what is likely to happen with any new UI. Without the forced introduction of new users to the UI, ala SL new sign ups, the advantages to users will have to be massive enough to generate strong user referral and obvious incentives.
Time will tell how well I’ve predicted the future.