(Here is my quick summary of the first 50 minutes of the meeting that took place on Friday between Oz Linden and the third-party viewer developers. The full meeting is 105 minutes, and I hope to have the second half posted by the end of this weekend. My apologies for parsing and grammar failures along the way, it’s still very early in the morning for me.)
Also: “LL Update Their TPV Policy” by Inara Pey.
Changes to the Third Party Viewer policy and related changes – changes that will effect all viewers connecting to SL, whether they’re on the ‘approved’ list or not:
llRequestAgentData – the LSL command that returns your online status – is being broken. The primary ‘abuse’ of this feature is displaying the online status of other residents (“Avatar Online” HUDs, for example). This will have the effect of breaking the “Show Online Status of Renter” on advertising and employee boards, and also will break inworld inventory delivery and update devices which check for online status before delivering an item to avoid capping. The revision of llRequestAgentData will only return a true online status if the script containing the command belongs to the owner or the creator of the script being run.
The work around for employee boards suggested is to give the employee the “show-online script”, have them cut and paste it into a new script, and then give that new script back to the employer. (My comment: Don’t give scripts with mod permissions to anyone you don’t trust explicitly – but that should be a given anyway.)
There is apparently no good way performance-wise for llRequestAgentData to look at the show-online permissions that people have set for themselves, so this is how they’re working around it.
Oz Linden admits that he’s not familiar with advertising panels that use llRequestAgentData, or what the boards do with that data. (My comment: Nearly every advertising board I’ve ever seen inworld in four years has had that feature – you either have to not spend much time out in the larger grid or be oblivious to your surroundings when you are to not be aware of this. Of course, the purpose for using llRequestAgentData in an advertising board is to notify viewers of the advertisement that the advertiser is online and can be contacted for questions about the product or service on the board.)
To retain the feature, new advertising boards would need to require the renter to create and insert a “show-me-online” script into the board along with the ad texture/notecard/landmark. Not unworkable, except perhaps for residents who are not beyond the learning curve.
Oz says that the need for delivery scripts to check the user’s online status before delivering inventory to avoid capping will be eliminated once the direct-inventory features are implemented. Oz indicates that he’s willing to assist in testing this with others on the beta grid, and that serious efforts are presently being made to fix the problems with inventory delivery.
There has been no date set yet on breaking the existing functionality of llRequestAgentData, but the estimate will be no more than two weeks. The change will not send an error to scripts which use the function, it will simply return a false response of “user not online”.
Client-side radars will not be affected.
TPV creators will not be permitted to work-around displaying profile data that users have requested to be hidden.
Online/Last Online status in group lists will not be affected.
Viewer tags on TPVs are going away. You will not be able to know which viewer another resident is using; Oz: “You can’t display viewer tags, period. End of story.” Existing TPV color tags will also be broken.
RLV will not be affected by this change.
Oz: “You can’t redefine how objects and the world behave. You can’t extend the world itself without doing it with us [Linden Lab]“. It is intended to provide a uniform way for everyone inworld sees things. The example given were the extra attach points which were implemented in Emerald – they looked fine for Emerald viewers, but not on other viewers. (For example, a tail on a secondary Emerald attach point might actually appear hovering over that avatar’s head on a non-Emerald viewer, rather than where it was actually attached.) Bugs are bugs, and can be fixed – this will not change. Parcel windlight, however, is a violation, as it alters the shared experience for all. Oz says that the reason that it’s a problem is because SL’s own parcel information field was “hijacked” in a way in which it was not intended in order to make this possible – but Oz indicates that TPV viewers have a “free pass” on this one until LL does parcel windlight properly. Oz admits that parcel windlight is a good feature, but one that should have been implemented together with LL. (Oz: “It should have been done, it should have been done a long time ago, and it wasn’t.”) When LL implementation of parcel windlight is implemented is uncertain.
The model is “If you want to do a shared-experience feature, that’s good, but you have to do it with us”. It needs to be in LL’s viewer, and then propagated outward.
Someone in the conference mentioned that if it weren’t for third-party viewers, multi-attachments, avatar physics, and parcel windlight would not exist because they were suggested to Linden Lab and rejected. Oz responded that these were all suggestions made before he was present and before the current management was present. An LL implementation of Qarl’s mesh deformer is in the works – but if LL chose not to implement the mesh deformer, then no TPV connecting to the Second Life grid would be able to implement it, either. (My comment: Does this mean, then, that the prim alignment feature, also created by Qarl, is a violation? My guess is that the answer would be no, since the prim alignment taking place would appear the same on all viewers.) The objection was raised that if TPV creators can innovate in a matter of weeks, but have to wait for LL to act on it in order to implement it, then what’s the point in continuing to innovate for Second Life rather than, say, a grid like Inworldz? Oz responds that LL is very serious about innovating within the limitations of staffing.
(My comment: I really do think that Oz makes a very good point here, and one that should not be overlooked – it’s not fair to accuse the present management of the actions (or inactions) of previous management. Sure, things won’t move as quickly as everyone would like them, but that all comes down to “how many people do we have to do how many things and what takes priority given these limitations”. It does seem, though, that some great tools – like Qarl’s prim alignment tool – are being rejected outright, and I hope that that’s not simply due to a clash of personalities, given the awkward ending to Qarl’s employment at the Lab.)
To be continued.