Some interesting stuff this week.
There is little change in the viewer pipeline. There is a new maintenance version, a new Alex Ivy version, and a new Voice version. All these versions are basically a folding in of recent main viewer update.
Alex Ivy, the 64-bit viewer, has 1 or 2 more bugs to be fixed before it is a candidate for promotion to the main viewer.
The Voice version has an elevated crash rate, but it is better with this last fold in.
All viewers have been tending to crash more over last few months than previously.
Third-Party Dev’s are not getting Linden reports on their viewer performance. Oz will be getting reports out Monday… events permitting.
All the Linden viewer crash rates are higher than the Lindens like. They are and have been working on it.
The new maintenance version with a collection bug fixes is doing pretty well. So, I suppose it may be the next candidate for promotion.
Will see the 360-Viewer updating as they catch it up to the main version.
There is work going on to revise the viewer’s splash screen. CEF will be used more for the new screen.
Phronimos Linden, a new Linden. Is the one changing the splash screen. He says the RSS feeds are not changing. Oz says they are improving data in traffic on those… widgets… those being updated. Not sure what that means beyond somehow better.
A way to see the new widgets will be released in the S3 bucket. His is for those testing them, soon. Watch the Open Source Dev channel for details.
The big difference users will see is the look and feel. The Lindens thing they have a better prioritizing of the display and it will all be in responsive design, meaning screen size independent.
Jessica Lyon – The Firestorm team has lost Nicky, a programmer in the team. They are working to replace her and integrate her recent work. The result is the departure is the Release date for the next version of Firestorm sliding.
So, for viewer developers the big question is when will the Lab turn off the old UDP asset fetching API’s? The Lab has not yet set a date. Guesstimating the it may be after the Valentine’s holiday.
Grumpity Linden was asked about Estate Tools… Seems Alexa Linden is a new product person… and that challenge is hers. But, while the Estate Tools have had no work done on them, they are next in line.
Talking about asset- HTTP stalls. This is a problem that seem to be popping up here and there.
Kitty is planning to remove asset-HTTP from Catznip, which will be a problem but is likely a workable short-term fix. Kitty plans to have it back in by the following release of Catznip. Which is when Kitty expects the problem to have been fixed on the Linden side.
So, those suffering from the problem, may want to use Catznip after its next release. At least until the asset-HTTP stall problem is resolved.
Lindens are aware of the asset-HTTP fetching problem where something causes HTTP pipeline to get out of sync. For some people, it is a frequent problem and for others it never happens. The Lindens are convinced it happens and are actively working to figure out why and fix it.
Since the Lab is having problems reproducing the problem, it is going to take some time to find it.
Moving to the Cloud
Currently there is ongoing work that is the foundation for moving service to the cloud.
The plan is to move the whole thing. They will be carefully moving in stages. You may not know asset storage has been in the cloud for years. Now additional services are moving to cloud. It will be a slow process. In the first steps users will not likely notice changes as backend servers move first.
Will regions be cheaper to lease? Honestly, the Lindens are not sure. They think and hope it will be. But, since costs vary tremendously by instance-type used, the Lindens have to experiment to figure out the costs.
Will region size change? Grumpity says they currently don’t know. Want larger regions, faster connections for Australia, instance regions…
Since region size is pretty much baked into the server-viewer software, the shift to the cloud doesn’t change that code. But, with less server restrictions and more server flexibility, the hope is regions can carry MORE load. I’ll take that to mean more people in a region.
See Pantera Północy’s video at 20:00 for the discussion. You need to hear it to get the feel of what the Lindens are trying and wanting to do. Their evasive answers at this point are based on numerous uncertainties. The Lindens are being pretty good about telling us what is going on, once they are certain they know.
The Firestorm anniversary party is Saturday 9/9/2017 – Firestorm Viewer 7th Anniversary – The free gift is a cute ass… donkey. So, brace for the jokes.
Talking about asset- HTTP stalls. What are HTTP stalls?
Assets are all the things we see in-world. They download via the HTTP protocol from the Amazon S3 cloud servers. When that6 download stream stops, stalls, things stop downloading.
It can require a restart to correct or some times just 5 minutes or so then it magically restarts.
They lost Nicky? What happened to Nicky? Not in the mood for depevelopment anymore? Health issues?
At Third-Party Dev meeting Jess said they lost Nicky…
I see Nicky just posted about the Kokua viewer. So…
So i asked a bit around. I see where the confusion came from. The Nicky they lost is NickyD whereas the Nicky we are always referring to is NickyP (thats Nicky who does Kokua and did some stuff for open source in the past too, including the havok alternative to upload meshes). I never heard of NickyD before, i assumed that they lost THE Nicky i know -> NickyP. I may sound blunt here but i don’t care about NickyD, infact i thought NickyP was the only Nicky and was working on FS still, the only Nicky i care about is NickyP.
Thanks for the information 🙂
Nicky want to stop development of Kokua in October 2017 because he will turn 75 years old.
He wrote about this here: http://blog.kokuaviewer.org/2017/07/25/path-forward/