SL Insiders on Alignment Code Controversy: No Personal Animus, Karl's Code Will Be Integrated into the SL Viewer

I talked with some SL insiders on background about the recent controversy over the alignment code from ex-Linden Karl "Qarl" Stiefvater, which Karl donated to the official SL viewer, but implementation of which has been delayed, provoking accusations of personal animus. The insider perspective, unsurprisingly, was quite different:

• The request for improvements to the alignment code was not motivated by any personal animus against Karl.
• Rewriting the code as-is to Linden specs would require significant company resources.
• A Linden engineer who submitted the same code would have received the same response.
• It's very rare for open source contributions more than a few lines of code to be quickly accepted as-is.
• The Lindens do want to integrate the alignment code into the viewer, which is why they assigned Oz Linden to the task. 

So there's that. Without taking sides, my sense is the Lindens do want to work with Karl's alignment code and Karl's mesh deformer, and that personality clashes like these are fairly inevitable when an open source project intersects with the needs of a for-profit corporation, even if both sides ultimately mean well. In any case, I hope the conflicts (technical and personal) are resolved, and Karl's considerable brilliance is made part of the official viewer as soon as feasible.

Tweet

 

Comments

[qarl]

well - that is EXTREMELY nice to hear.

but guys - i'm not the only one who misinterpreted the rejection. i received about a 100 emails the day it was posted, and all of them had the same interpretation i did.

maybe next time go with something more like: "hey, we like this a lot and want to use it, but we really want these things changed before it goes live, and we don't have time to do it right now - could you?"

also: [email protected].

[Pussycat Catnap]

The lindens use 'corp speak' in that JIRA response.

Qarl, since you were formerly on the inside, but are now out here with us on the outside, hearing people you know spit corp speak at you can sound downright hostile, at best. Where-as to those of us always on the outside, it just comes across as 'monolithicly cold'.

That is of course, stated by someone who has -no knowledge- of the personal stories involved. I just know that its rather unnerving to be chatting with folks around the 'water cooler' one day, then after a lay-off, call em up on the phone for some piece of paper and get a scripted response.
- A statement on our culture's acquiescence to 'the corporate masters' that we all so readily fall into that pattern.

[Bethiny]

"Rewriting the code as-is to Linden specs would require significant company resources."

This is highly subjective especially when considering the amount of Lab resources that would be needed to "fix" (AKA Lindenise) the alignment code which imho would not be that much since the heavy lifting has already been done thanks to Karl. This code is a solid foundation that many users are already familiar with from using TPV's. If LL would put 1/100th if not less of the resources the Lab put into Display Names into polishing up the alignment code the Official LL Client would greatly benefit especially builders who scream for such enhanced tools.

Assigning the task to Oz really means nothing other than Oz is looking for a unpaid non-Linden open source developer with a LL Contribution Agreement to pass any work off to which is far from demonstrating this is a priority or ever will be adopted by LL in any form.

Regardless, if the alignment code ever gets into official LL Viewer, I suspect it will be due to very little Lab resources or engineer time and more so due to free work to Lindenise the code by unpaid developer who realizes the usefulness of tool as is and puts efforts into making it meet Oz & LL additional demands. LL resources are better and more often put into some rather useless features.

[Shug Maitland]

It is not so much they "rejected" Quar's code contribution, it is the socially inept way they went about it (as Qarl says above) and the inexplicable excuses they gave in the guise of reasons.
-- Align sculpts to prims? Really??? Does anyone have a theory about how that might work?
Perhaps Builders Brewery should hold some Lindens only classes :)

[Arcadia Codesmith]

If I could have the tool sans the drama, that would be ideal. kthx.

[qarl]

@Pussycat -

yeah, i understand that. but i don't know of any organization that uses 'corp speak' on its volunteers. in the context of free labor - it comes off as dismissive and superior.

[Ezra]

That's the thing. It was odd to see Charlar react that way. Read the last mesh user group meeting from the 9th and he's joking about penises: http://wiki.secondlife.com/wiki/Mesh/Archive/2011-01-09

Why should Jira attitudes be so much more stoic?

I think that's why folks read into this as a particularly negative response. Really, it still is. It takes an unnamed 'insider' to mention Qarl by name and address the issue humanly?

And it still seems like the patch is held at standards different than other Linden Lab features that affect us, regardless of their claim that Linden engineers are held to the same standard. I -wish- they were given the many buggy and half-baked features that released last year alone.

Hopefully they'll merge in the patch, and then improve it themselves or let others. At present, its 100% value and 0% harm, and already thoroughly tested by half the grid. That's much, much more than can be said about typical new viewer features that go on to be improved more live.

So Linden Lab can continue proclaiming supposed standards and reject the patch, 'lest someone bend to their demands beforehand, or simply just include it.

[Jus]

Actually in some circumstances it seems the standards set for LL engineers on certain issues or JIRA's are much lower and some rather obnoxious obvious problems (too many to list, but there are plenty) get labeled "expected behavior" by Charlar or others. Many of these cases in fact are not or should not be considered "expected behavior" but it becomes a fallback category to dump an issue with LL own work and development into a OK fine we will stick with this for now category rather than holding themselves to higher standards or expectations within LL.

[Tateru Nino]

I was contacted yesterday by a couple of Labbers - which was unusual - who gave me exactly the opposite story: That this was personal, and that certain people within the company would work hard to avoid the prim alignment patch being incorporated into the official viewer.

Now I wonder if the whole thing is a joke, or if they're two entirely opposing (but equally valid) perspectives at the Lab.

[qarl]

ug. Tat - i hope to god they were messing with you. if LL really is that petty, or has a petty faction like that - then SL truly is in trouble.

[Tateru Nino]

Personally, I think that both opinions might be equally right (or equally wrong). Any two people or groups can see the same things and interpret them differently.

One group clearly thinks that Qarl is getting the boilerplate standard treatment, while another group clearly thinks that Qarl is being singled out and stonewalled. Either one may be right or wrong. Both may be half-right. Both may be incorrect.

It's largely a matter of perspective. What's *really* interesting is that people at the Lab felt strongly enough about it either way to get in touch with me and Hamlet about it and to express how they personally see things.

[James Cook]

As someone late to the party... Meh.

I used to work on the viewer (as James Linden). I wrote many of the original prim manipulation tools. I also have long wished for a prim alignment tool. It's probably my fault we never had one.

Qarl's patch looks good to me. (Although... QAlignTool? It hurts that much to use the same LL prefix as every other class? :-) In addition to dealing with the viewer-specific boilerplate there's some good, tricky 3D math in there. I was never good at that stuff, and Qarl is. Kudos for writing it.

That said, I now work on another company-backed open-source project, the Chromium browser. Those Jira comments are exactly the sort of comments any other engineer on Chromium would have made on a patch I submitted. I think it's reasonable to expect that a top-level tool in the object editing panel would work similarly to other top-level tools. I know all the wacky coordinate systems suck to deal with... and all the wacky prim-like objects we have. Boy do I know. Does it work when you have grass selected? How about a tree? But that's the difference between a top-level tool and something you put in a submenu.

I don't know Charlar. He must have been hired after I left. It's unlikely Charlar had any overlap with Qarl or really knows how talented Qarl is. This smacks of the typical kinds of misunderstandings that crop up when people can't talk face-to-face. I hope this can be worked out -- the viewer really could use some better alignment tools.

[qarl]

oh my god James - it is SO good to hear from you. i hope all is well in your world.

yes, i completely hear what you are saying. and i do largely agree.

but may i suggest that google's engineers (on a top-tier product no less) might also suffer a bit from inflated-ego complex? i'd hate to forgive bad behavior simply because others have it too.

just recently i had an opportunity to volunteer at my daughter's school. open house is next week - they need to get the place in shape - and they don't have the largest budget. so i joined the cleaning patrol.

i have to say, i probably didn't do the best job mopping. but that's ok - the teacher was *very* happy i was helping out.

now imagine if she'd said to me "the cleaner did not sufficiently clean these corners - his work here is unacceptable."

[Metacam Oh]

If the Lab was pumping out features, tools, and fixes all year it would be a different story. They just hyped Mesh for 2 years released it broken and now Qarl is cleaning up the mess and not to Linden labs expense. Maybe its time the Lab stops trying to be Fort Knox when it comes to their code and how many versions of their server software are able to be run and start opening up to the customers and volunteers who would immediately start making your product better for free the second you implemented this new policy. Look at Minecraft they have thousands of people programming mods, add ons, features, and at any time that company can officially take that code and use it to their own liking. Let's speed up the development a bit beyond snails pace before we start knitpicking. Joe Farmville who signs into SL for free may not want the prim alignment tool, but most of the people making your current sack of change do. A change of policy is due.

[FunnyOne]

"Rewriting the code as-is to Linden specs would require significant company resources."

Opensim runs better than SecondLife. Qarl's code should have been something that was included many years ago and Linden didn't get the job done. Just like poor search code etc.. and we can go on and on about the issues as witnessed in SL for many years.

Fact is, Linden doesn't like the idea of using code they don't write, and the code they do write is not all that super much of the time. What else are they going to say when they cannot achieve it, cannot get it done?

Qarl's code rocks. I've used it since he put it online. And it should have been part of the viewer version 1 IMO.