The Third Party Viewer Directory is updated from time to time. Oz Linden changed the TPV List 4/16.
The viewers with enough users to generate reliable stats and that participate in crash reporting are listed in the first section of the list in the order of stability, least crashes. In this update Catznip is in the first position as the most stable.
The Imprudence Viewer is second, Singularity, Phoenix, Firestorm, and RLV follow.
Congratulations Catznip!
Too bad LL doesn’t provide stats for the Cool VL Viewer… Because then I bet it’d be the stablest viewer (no wonder, after 5 years of constant development, improvement and bug fixes).
When Dolphin v1 was around, it figured once in the crash rate list… and was rated as the stablest. Dolphin v1 was in fact the Cool VL Viewer with a few more patches added (sailor mini-map and sailing related stuff, mostly).
I am pretty sure I’ve never crashed using Cool VL.
Henri, it is a TPV Directory listing…so in order to be included in the statistics, you’d need to be on the TPV Directory.
I understand your reluctance to meet the requirements of that listing, but even so, if you aren’t listed, you can’t be compared.
True Tonya. I think Henri knows I like the Cool VL and his remarkable work. Also that people read this blog for viewer news and opinion. While he is not going to get rated in the TPV List, in all fairness those that blog about SL viewers could make the information available.
There is a problem in that the Lab provides TPV Dev’s stats on the viewers and viewer use and to no one else. The dev’s do let pieces of and hints about the information out. But, I am guessing the stats reports are covered by an NDA agreement because we, general users/residents, never see them. If the information were easily and publicly available, bloggers like myself and Inara would publish it. That could create some competition among dev’s, if there isn’t already is debatable. That might or might not be all that good a thing. I would like the freedom to find out. It would certainly enable the users to better decide which viewer to use.
In many ways I feel hampered in providing reviews of viewers. First the complexity and variety of viewers is overwhelming. When I started looking for reviews of SL viewers in 2008, when Nicholaz Editions were the big thing, things were simple. I wanted a viewer that did not crash. The Nicholaz version was that. Report that a viewer crashed less than the SL viewer and one was done. Things are way more complex now.
At some point I had to decide how objective and comprehensive I was going to be in viewer reviews. I decided to be subjective and completely gave up on comprehensive. Being comprehensive was just too damn much work. So, when Henri or other develops step up to correct me or add information I’m very appreciative and think it helps everyone. I know, try as I might to get things right, I make mistakes. In this case I SHOULD have titled the article differently to be more accurate. Henri called me on it and I appreciate that.
You are correct. We are given the crash rates, and other viewer stats, on condition we not release them. We can say some things about then, but we can’t give exact numbers, and we can’t use them to compare one viewer to another. Further, we only get those numbers for TPV Directory listed viewers with more than a certain number of user session minutes [in] the preceding week, so viewers with low usage don’t appear on the list.
And that’s why I mentioned Dolphin v1 (a Cool VL Viewer v1.25 clone) to validate my guess… Dolphin v1 *was* in the TPVD, it *was* listed once in the among the viewers with sufficient statistics to evaluate the crash rate, and when it was, it *did have* the lowest crash rate…
Since many things improved since Dolphin v1 (with the Cool VL Viewer v1.26 gaining *loads* of more crash fixes and stability improvements, such as memory usage auto-throtlling to prevent out of virtual memory address space crash, etc), the Cool VL Viewer can only be stabler than it was.
In fact, I pretty much never crash during *hours long* sessions of my viewer, and for the rare cases when I crash, I *immediately* track down the crash, fix the bug, and the Cool VL Viewer users get a new fixed release only a few days later… There is simply no other viewer to provide this level of reactivity to newly discovered bugs.
These are just *facts*.