Jump to content

Lucia Nightfire

Resident
  • Posts

    3,477
  • Joined

Posts posted by Lucia Nightfire

  1. 14 hours ago, Prokofy Neva said:

    I don't know if you could easily program for all those factors.

    But it seems to me there are obvious metrics to track to make traffic real:

    o number of avatars who make purchases

    o number of avatars who make multiple purchases

    o number of avatars who return to make additional purchases

    o avatars making purchases over $10

    etc.

    Many of those can be gamed as well and the amount of money transactions one has should determine if they are real or not?

    Why not just have additional Premium account traffic metric if you want to only count people actually giving LL money?

    I would at least like to see a live traffic value in legacy search in the same way they can show parcels for sale instantly.

    Why can't they update a live agent count value that replaces traffic or is an alternate searchable/sortable/choosable filter/metric?

    I don't want to know how popular destinations were 24+ hours ago, I want to know what destinations all the people are at presently.

  2. 13 minutes ago, Extrude Ragu said:

    There is an additional regression that OP doesn't mention

    Rezzed objects now have ~3 second delay before scripts start working. Was about 1/4 of a second before, it sounds like a small thing but it makes systems that rez objects such as pose systems etc very clunky, borks scripted weapons etc. There are multiple bug reports about it on the SL JIRA so it's not just me.

    That has nothing to do with the move to the cloud. That was explicit introduced behavior done back in November 2019.

    The developer that made that change has stated in the user groups that they would like to fix that at some point in the future.

    • Like 2
  3. On 1/27/2021 at 6:25 AM, ChinRey said:

    According to OZ, the LL developers used the cloud migration as an excuse to do some serious code cleaning. That should improve the performance even thous the migration itself didn't.

    When did he say this? He and other developers have stated in the user groups that the focus was to just get the code into the cloud and not waste time on cleanup.

    They did state during this process they took notes or areas of code that needed cleanup/overhaul once the move was done.

  4. That implementation of anim2bvh has major flaws, such as not being able to handle 0.0 duration, single frame animations plus the creator added stupidly large bone sizes and mandatory 3 minimum frame injection that overwrites/replaces data for animations using lesser frames.

  5. 12 hours ago, Nika Talaj said:

    the map hasn't worked since September,  EEP needs a shot in the arm to stimulate uptake, and the opportunities to grow the grid, now that it's on AWS, are endless.  Friggin' Roblox has 830 employees, what's LL at, 250 now? 

    Map tiles haven't worked since November 18th.

    Why does uptake hinge on EEP and not other things like giving creators tools to make content more efficient or to make completely new content altogether?

    With the two rounds of layoffs prior to finally selling Sansar, LL shed 70 employees from their 200 total.

     

    • Thanks 1
  6. For clarity, Ebbe said in the following interview that out of 200 Linden Lab employees, 130 were working on Secondlife.

    Also, from 17:30 to 20:42 he talks about LL's information campaign mindset towards unoptimized content which ArcTan phase 1 will be all about.

    At 20:42 he also thinks the move to AWS would be done around early 2020.

  7. 7 minutes ago, Aishagain said:

    And it has been a bone of contention with non-US customers for as long as I have been using SL.  But that's not the Grist in my mill: that is the revelation, if you can call it such,  that SL's servers are no longer under Linden Lab's direct control.  We have seen numerous recent indications (map tiles, Aditi inacessibility, lack of new Regions, TP routing issues and regions being offline or inaccessible for hours) of what that means in the last couple of months.  To me, it does NOT bode well for the future operation of SL.

    Map tiles was due to LL prioritizing it last, thus being worked on post Uplift.

    Aditi accessiblility has been a problem before Uplift, same with lack of new regions.

    TP routing seems to have been introduced after Uplift.

    Regions showing online, but not allowing entry has been caused by grief vectors for years. Did support explain in recent instances that AWS was to blame?

    • Like 1
  8. On 1/23/2021 at 8:38 PM, Naiman Broome said:

    I think that this technique could help improve the game...

     

     

    This is procedural animation. It scares Lindens who can only envision a world of hard animation assets to death.

    Also, there is one and only one Linden that would be able to improve the animation protocol, but management has had them working things that should have been delegated to other employees for years now.

    This is why SL's evolution is stunningly slow. Because management will not hire people to take the workload off key developers nor will they stop tying up specialists with tasks someone else should be doing.

    I can't believe that, accordingly to Ebbe's own admission, that they were able to afford 70+ people to work on Sansar, yet they can't hire one developer to take to workload off of key developers.

    • Like 1
    • Thanks 1
  9. Did you sit on an object belonging to said person at some point in the past?

    Camera tracking permissions can be granted and stored indefinitely like animation permissions through sit vectors.

    See https://jira.secondlife.com/browse/VWR-13228

    IMO, maintaining permissions of someone other than object owner should have been disallowed (intentionally broken) years ago.

    There are better/legit/efficient ways of granting and using other user's permissions that don't depend on harvesting vectors.

    But griefers, "attack HUD" creators and owners/historians/archivers of ancient/inefficiently scripted content will cry if it is ever broken.

    • Like 1
  10. 5 hours ago, Gabriele Graves said:

    I never once said they were all you said.  I am not accusing you of anything either.

    I am aware of avatar freezing.  I don't like it much but the owner has to be present for that, it cannot be scripted - that makes a big difference.  If the land owner could script freezes on an ongoing basis then that would be far worse.

    There is this feature request. I don't think you'll like the prospect much either.

    1 hour ago, Wulfie Reanimator said:

    The function was not released until 24th of July 2012.

    Technically, it was available the last week of May 2012, but suffered release setbacks when LL only closed a permissions loophole on one RC channel instead of grid-wide and during that weekend and especially on the following Monday, June 4th, 2012, a day that will live in infamy, people all over the grid were mass teleported all over the place and to the cornfields. There was no defense unless the destination region was full. After that, the release suffered major bugs for over two years such as breaking scripts every time you teleported.

     

    • Like 1
  11. LL may or may not offer their own official crypto currency and/or exchange in the near future.

    If they do, they will shut down anyone that would be competing with them even if it is an in-world service that allows purchasing crypto with L$ and no other currency/token exchange.

  12. Governance will NOT do anything against a cluster of noob avatars standing around on tiny platforms at high altitude.

    I've AR'd this in many places for years.

    Nothing has been done, nothing will be done.

    If LL was even 1% serious about combatting such abuse they would change the taffic calculation to exclude "inactive" scenarios.

    Such scenarios would be

    • no # meter(s) change in user position while not sitting
    • no # meter(s) change in user position while sitting
    • no # meter(s) change in sit object position while sitting
    • no controls usage after # minute(s)
    • no open chat usage after # minute(s)
    • no IM chat usage after # minute(s)
    • no group chat usage after # minute(s)

    Yes, the chat usage can be gamed far easier than movement.

    Also, # meter(s) change would be dictated by parcel size. The larger the parcel the more distance change required.

    Is any of this practical? Most of it probably isn't on its own, but they are some interesting points to consider in some combination or another.

    And I'm aware not all environments are equal.

    What factors would you consider when trying to make traffic more fair?

    • Like 4
    • Haha 1
  13. 1 hour ago, animats said:

    A big limitation of animesh is that they can't "sit". That is, they can't  use the animations of another object. They can only use their own built-in animations. So it's hard to do this in a general way.

    (I've given some thought to a pseudo-sit system where messages are exchanged between animesh NPC and suitably equipped objects. Something along the lines of "NPC to chair - want to sit there, send sit target info and animation", and the chair sends back "Sit at <x,y,z> rot <x,y,z,w>, run anim with UUID nnn." Unfortunately, while you can apply to an object any texture for which you know the UUID, you can't do the same thing for animations. So loaning an animesh NPC a texture so it can sit like an avatar is not currently possible.)

    We need remote animating capability, possibly facilitated through the Mod Keys proposal.

    Said temp animations don't survive taking into inventory, owner transfer or region restart.

  14. 10 hours ago, Wulfie Reanimator said:

    You fly over the parcel at 4000 meters, or cross a bit of the corner at ground level, and a script teleports you into the lobby of a skybox.

    Or a more malicious example:

    
    
    
    
    while (TRUE) { llTeleportAgent(agent, "", trap, ZERO_VECTOR); }

     

     

    9 hours ago, Kyrah Abattoir said:

    An improvement on llTeleportAgent could be to silently discard a teleport that won't move you more than 1m.

    (If anything, to avoid people using it as a really bad move2target)

    Just a small FYI.

    llTeleportAgent() and llTeleportAgentGlobalCoords() have a low repeat usage threshold, like a burst of 3 executions in 3 seconds then repeated failure until another 3 seconds of no execution attempts before another tp is allowed. This goes for all same-owner scripts in multiple objects in the region.

    There is a 2 meter minimum distance requirement.

    • Thanks 3
  15. Yes, rain, rain clouds & snow always cut out my connection. During the 2006 - 2009 years SL was such an addiction, I'd take the chance of falling off the roof and dying to wipe the lense on the satellite LNB of any rain droplets or snow build up until I made an access hole through a window an A/C unit occupied and used a long pole with a rag attached to the end to do it safetly. Those were the days...

    Yeah, Hughes used to allow unlimited downloading between 2am and 5am (can't remember which TZ I was offsetting), but then when they started catering to other markets, that went away too.

  16. I used Hughes internet with SL from 2006 - 2009.

    It used a 2 watt LNB and latency was about 0.25 seconds on avg.

    I used it because at the time, dial-up was my only other "broadband".

    I had to deal with a 250MB daily download limit and later, low speeds during prime time hours because Hughes started catering to other markets and shafted their internet users.

    Today, I would hope latency has improved, but the nature of satellite tells me it won't be by much.

×
×
  • Create New...