packet loss effecting building and textures

When I first looked at SL i was taken with the idea that the content was created by the residents and that it was to be an evolving world. Well that was obviously a pipe dream as residents now struggle to even get one prim to stay in the position it is required.

I have struggled for weeks with prims bouncing about the screen after I have positioned them. They have now started to change rotation too and can even move up to two minutes after i have stopped editing them. This makes it almost impossible to create things and perhaps enhance the world for others.

Texturing is just as bad now. Last night i had a group of 8 prims that i wanted to texture with a single texture but i had to select it seventeen times before it stayed as required and did not change back to default on random faces.

If these problems are not fixed pronto then SL wil end up as a barren landscape as builders will not waste their time and effort only to find they have created a pile of random lego block.

These problems have been bug reported several times by me and others. They have also been discussed in the forums, but then i doubt they ever get read by those that can do anything about it.

1. There is also the problem of textures stretching when you stretch the prim, whether you have the "stretch texture" box checked off or not. There is no workaround for it that I know of. Since apply repetitions per meter no longer works at all reliably (and never does, in this case), you have to manually guess at it, which can take as long as 15 minutes to fix one prim.

This texture bug has nothing to do with packet loss. How do I know? Because I have learned to build with one eye on the bars at the top right of the screen. I time my moves - where possible - around those bars. (I say where possible, because in some situations - such as having a box of contents in edit - there may be steady packet loss.)

2. For building in general, slow and judicious moves help with the packet loss problem. Give every move you make extra time. Don't do any two things too fast. Keep one eye on the packet loss bar and time your moves accordingly. (You can actually learn to keep an eye on it out of your peripheral vision, without thinking about it too much.)

3. Whether the other recent texture bug - the one that applied the texture to many faces of one linked prim, rather than just the face you have selected - has been fixed or not, I don't know, because I have stuck to my workarounds there (texture only unlinked prims; always randomly click in other boxes before clicking out of the edit page) and haven't taken time to test it lately.

4. As for the "edit linked prims" bug (can't stretch linked prims) that has existed almost since I joined SL - it seems to have gone through various incarnations: not working at all, working sometimes, working for a while and then not, and now, I'm sorry to say, back to not working at all.

Not only does it not work at all again, it does so with a new twist - rather than simply spring the prim back when you let go of it, it will stretch BOTH sides of a prim, move the prim up or down, and leave it distorted and in the wrong place until you click "position," click on the prim again, and click control-z. Then unlink it which you should have done in the first place.

5. There is one grim consolation. I doubt there will be just a whole lot of competition in the future from new builders, compared to back in the day, when things worked as they should.

coco

I suspect (wich is a way to say "i'm pretty sure" that packet losses affect tetxture rendering. High PL are correlated with fuzzy textures ang gray world. Since 1.13.3, I'm experiencing huge packet losses (10..50%%). Traceroute from Paris to SF show loss of ICMP packets entering Level3 then Internap. The trace never reaches lindenlab.com.

3 * 213.228.20.254 (213.228.20.254) 49.782 ms 41.876 ms
4 lyon-6k-1-v804.intf.routers.proxad.net (212.27.50.102) 45.438 ms * *
5 * * *
6 te-3-2.car1.paris1.level3.net (212.73.207.13) 82.628 ms 51.680 ms 88.101 ms
7 ae-0-55.mp1.paris1.level3.net (4.68.109.129) 74.893 ms ae-0-51.mp1.paris1.level3.net (4.68.109.1) 51.974 ms ae-0-55.mp1.paris1.level3.net (4.68.109.129) 139.849 ms
8 so-3-0-0.mp1.sanfrancisco1.level3.net (209.247.8.89) 269.268 ms as-0-0.mp2.sanfrancisco1.level3.net (64.159.0.217) 233.969 ms so-3-0-0.mp1.sanfrancisco1.level3.net (209.247.8.89) 233.675 ms
9 ge-6-0-0.gar1.sanfrancisco1.level3.net (4.68.124.206) 268.410 ms 300.448 ms 389.129 ms
10 internap-ne.gar1.sanfrancisco1.level3.net (4.71.44.6) 273.771 ms 238.799 ms 215.237 ms
11 border1.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.1) 220.816 ms 316.265 ms 221.491 ms
12 * * *
13 * * *
14 * * *
15 * * *
16 *

After reading the last two entries, I had to check since I hadn't ran into these paticular problems. So, having the window open to SL, I jumped in game and tried each problem. Ok, first I don't have the severe packet loss so maybe its their connection that has things bouncing about & repositioning mins later. I sometimes will get a small offset reposition(usually after rotating), it is irritating though not prohibitive. I linked some prims and both stretched and textured individual faces with absolutely no problems. I have stretch linked prims often since the end of 2005 and only rarely had a problem stretching linked prims, unless of course one prim in a linked set already has one side at .o1, then it can't be reduced or the opposite if one prim has a side of 10m's it will not stretch larger (the set), while linked. The packett loss, is generally client side from what I understand, even though I think sometimes its source comes from SL . However, clearing the cache, defragging the hard drive now & then, deleteing all temporary Internet files & cookies ( they(websites you've visited) hide spyware there). If, on multiple computered network changing the port # in the target helps. Then there is the equipment and the connection. None was mentioned so I am guessing we would find less than desirable conditions in either one or both areas. Its hard enough working in a buggy environment and when the equipment is less than up to snuff I can appreciate your frustrations. I would not be in SL if I couldn't build. Good luck fixing your issues. BTW , I just posted this a few minutes back and its currently, FEB 1 at 12:23 pm., funny it says I originally posted this 8 hours from now, always wanted to be a time traveller

Thank you for your concern, Ray.

A resident named Autumn Heyse determined that the texture face bug (the one where you apply numbers to a single prim face and the numbers apply themselves to other faces either on that prim or in a linked set of prims) does not affect everyone.

She also came up with a way to reliably replicate this bug - that is, it was reliable to those who suffered the bug, though not for those who didn't have it.

With that information, Blue Linden was able to determine that YES THIS BUG DID EXIST! Even though her replication didn't work for him - he spent some time coming up with another replication that did. YAY! So at least that one got on the known issues page.

Therefore it is quite possible that a bug exists which affects some people regularly, but not others, yet comes from the SL side of things, rather than a person's connection or personal computer.

As for this new texture bug, I have stretched linked prims since the beginning of 2005. Although I have had issues with editing linked prims for a long time, I never had the issue with the textures stretching even when the box isn't checked off before, until the update before last, I believe it was, when it arrived suddenly and hideously.

As I said, it happens when there is no packet loss showing.

Combine all those facts together, and I conclude that something has changed with the updates that is causing this bug. That a particular bug may not affect everyone isn't much consolation to those it does affect, and also doesn't mean that the bug isn't on the Linden side of things.

Clearing cache, defragmenting hard drive and all the rest don't improve these bugs at all. Some things in SL used to work, and now they don't. It's that simple, really.

I"ve had days where I have experienced everything the OP describes. And not coincidentally, those were days I had a lot of packet loss. But for the most part, on a good day (which luckily I have many of) I don't have any issues positioning,stretching, texturing, etc. So my own personal experience points to packet loss as being the culprit.

I was going to copy and paste all the quotes but suffice it to say, still no identification of the video card the computor(wireless or not), the connection. All the things experienced I have no doubt exists. SL has bugs, but until we eliminate what may not be causing the issues we can only ass u me. I know, if I worked in support with the myriad of computor architecture and various connection speeds out here, I would have little chance of determining correction of issues or bugs without first eliminating the most obvious issues first. Believe me, there are times I have cursed the computor & sl for bogging down freezing up, jerky movements , sllipping of selections (even buying things I had no intention of, not to mention dealing with griefers and many other things , and I won;t assert that SL and or its bugs have not been the source of some problems, niether will I say that it is all sl especially when this thread all began with the notice of packett loss.

"Clearing cache, defragmenting hard drive and all the rest don't improve these bugs at all. Some things in SL used to work, and now they don't. It's that simple, really. "

Pointing out, packet loss could be a bug and also could be caused by many other conditions,
so though your problems may truley exist because of an SL bug(s), clearing cache, defragmenting hard drive and all the rest, has helped a great many people with varying issues and it may be that simple for you( that it is all sl), but until more is known, esp. about the equipment client side, I respectfully diss-agree.

Well, I do all that, Ray.

Packet loss is at least partly a bug; that has been pretty much determined, from what all I have read. I've given my pointers for working around packet loss.

I am saying that the building issues I mentioned are separate from packet loss issues - though, of course, can be confounded by them.

All those other things - bogging down, freezing, jerky movements, and slipping on selections, not to mention lag - I'm not really talking about them right now. I'm just talking about specific bugs which I do know specifically exist.

The reason I know that is my computer is the same, my connection is the same, all the numbers on the stats page are the same, etc. - but suddenly there comes a brand new bug along with an update.

In other words, chances are that something which worked perfectly well for a year and a half, then suddenly stopped working immediately after an update is very, very likely a bug on the Linden end.

I figure sometimes this happens because they tried to fix a previous bug, say a texture bug, and it had unexpected effects on the rest of the way texturing works.

coco

I have read several replies stating that the problem of packet loss may be client side. This does not hold water as the problems have only started immediately after the so called sl updates. You can only blame your customers so many times before people become aware that the problem is widespread and has to have its roots in your product not the users.

I have been in SL for 18months and my system is high spec dual processor with twin gtfx graphics cards running through 8 meg broadband connection so i hardly feel that it will be my end. I had very few problems with building until the update before last when immediately after i had the problems described. This suggests it is something that changed during the update not that my system suddenly chose that moment to stop working correctly