Open Source Meeting/2008-05-29

Agenda

Update from the Lindens (standing item) - Rob Linden
If this is an appropriate forum, I would like to talk about ways we might utilize the JIRA to communicate refactoring, decoupling, cleaning of code mostly to prevent overlap with linden internal targeting. I feel these smaller "housekeeping chores" are very valuable for everyone moving forward and they are within the reach of contributors with other jobs but can find some time to help. It is important to prevent overlap in this work for both parties, and also to prevent the possibility of work being done but the implementation of said work does not account for directions the code may take in the future because those goals remain uncommunicated or are difficult to discern. JB Kraft 11:10, 29 May 2008 (PDT)

jira isn't the best way to do this, but it might be possible to add a "project started" status to a suggestion so on-going projects could be sorted. A separate webpage outside jira itself could be fed the list of jiras with "project started" status, so people could scan those without having to log into jira, which is kinda overkill just to see if someone is already working on a project idea. I won't be able to make the meeting today, so just tossing out ideas... Saijanai 11:51, 29 May 2008 (PDT)

Next item? Your name? (sign with "~~~~" which will be converted to your name + timestamp)

Default agenda is issue triage of issues listed here: http://jira.secondlife.com/secure/IssueNavigator.jspa?mode=hide&requestId=11240

Transcript

[13:58] Liana Linden: Howdy, all.
[13:58] Jei Desoto: Hello : )
[13:58] Carjay McGinnis: Hello Liana
[13:58] Wilton Lundquist: Hi there.
[14:00] Rob Linden: hi folks
[14:01] Jei Desoto: Hello : )
[14:01] Wilton Lundquist: Hello.
[14:02] Rob Linden: we plan to post the transcript of this on wiki.secondlife.com, so please don't say anything you don't want posted there.
[14:02] Rob Linden: Agenda: http://wiki.secondlife.com/wiki/Open_Source_Meeting/Agenda
[14:03] Rob Linden: my update: Google SoC began this week
[14:03] Soft Linden: / ( Love JB's suggestion on there )
[14:04] Rob Linden: I've been doing some infrastructure babysitting this week (PJIRA/PWIKI stuff)
[14:04] Carjay McGinnis: yes, it sounds a good idea
[14:05] Rob Linden: I think we're still in good news/bad news territory with those from infrastructure perspective.
[14:05] Rob Linden: good news: it seems to be up to the load these days
[14:05] Rob Linden: bad news: some connection oriented flakiness still persists
[14:05] Carjay McGinnis: meaning?
[14:06] Rob Linden: good news: sounds like it's narrowed down to a firewall issue.....I'm a little fuzzy on the details
[14:06] Rob Linden: well, power users still notice microoutages
[14:06] Soft Linden: Cases where you just get a connection error straight away - a 500 or similar
[14:06] JB Kraft: is it tomcat you are using?
[14:06] Soft Linden: Really frustrating if it happens on a form submit, ie - after entering a long comment
[14:06] Rob Linden: JB: yeah, I think so
[14:07] Carjay McGinnis: ah, ok, not happened to me yet
[14:07] Rob Linden: my babysitting has been more on a management level than hands on
[14:07] Carjay McGinnis: yes, I tend to copy thins I enter into a web form
[14:07] Carjay McGinnis: things
[14:08] Rob Linden: Liana: wanna talk about CMake status?
[14:08] Poppy Linden: o hai
[14:08] Poppy Linden: yeah, i was just about to grab foods but realized what time it is
[14:09] Liana Linden: CMake is being made. There've been some writnkles in the merge process, but we're hoping to see everything ironed out soon.
[14:09] Liana Linden: Heroic efforts there on the part of Sardonyx, CG and Mani Linden.
[14:10] Carjay McGinnis: somebody mentioned it might be on the shelf for the next release cycle?
[14:10] Liana Linden: (and many others -- this isn't an Oscar speech)
[14:11] Rob Linden: Carjay: yeah, assuming it successfully makes it into release, I don't believe Viewer-1-21 has been branched off yet, so it should be part of that
[14:11] Carjay McGinnis: that would be cool
[14:11] Carjay McGinnis: thanks for the effort
[14:11] Rob Linden: (where "release"=="trunk"...we've dabbled with renaming it)
[14:11] Soft Linden: The QA and release teams at this point are pushing to hold any further release merges until cmake goes in. It's definitely coming if nothing fundamental proves broken.
[14:12] Soft Linden: Be really surprised if it's not the norm for 1.21
[14:13] Rob Linden: any other Linden updates, or should we move to JB's item?
[14:13] Liana Linden: Let's hit the agenda.
[14:13] Rob Linden: If this is an appropriate forum, I would like to talk about ways we might utilize the JIRA to communicate refactoring, decoupling, cleaning of code mostly to prevent overlap with linden internal targeting. I feel these smaller "housekeeping chores" are very valuable for everyone moving forward and they are within the reach of contributors with other jobs but can find some time to help. It is important to prevent overlap in this work for both parties, and also to prevent the possibility of work being done but the implementation of said work does not account for directions the code may take in the future because those goals remain uncommunicated or are difficult to discern. JB Kraft 11:10, 29 May 2008 (PDT)
[14:14] Rob Linden: (copied from the agenda)
[14:14] JB Kraft: rob: just to back up a sec and just a thought, perhaps jetty might be worth a look wrt pjira. ive had more luck with it then tomcat
[14:15] Soft Linden: JB - maybe you could write up your findings to that effect as a WEB- JIRA, calling it something like "Evaluate jetty as a tomcat replacement on pjira"?
[14:15] JB Kraft: sure, i'll give it a whirl :)
[14:15] Rob Linden: good idea both JB and Soft. yeah, that'd be good to have
[14:16] Rob Linden: I'm not sure that tomcat is the problem here, but who knows
[14:16] JB Kraft: well, jst a thought, tomcat has given me weird giref in thepast
[14:16] JB Kraft: grief
[14:16] Rob Linden: anyway, with respect to the suggestion of coordination via PJIRA....
[14:16] Q Linden: jb, I agree with the premise on the overlapping work issue...but honestly, that sort of thing just happens from time to time
[14:16] Q Linden: even within the organization
[14:17] Q Linden: hell, i've seen it once in a 6-person company
[14:17] Soft Linden: So, this cleanup issue - what needs to be worked out? Generally, we'd want to make sure cleanup was happening with the coordination of someone familiar with that code to verify that the cleanup benefit is better than the risk of touching that code
[14:17] JB Kraft: yes, indeed Q :)
[14:17] Soft Linden: And also to make sure it wasn't code actively under development, creating merge confusion
[14:17] Carjay McGinnis: I like this especially as a chance to improve communication
[14:18] JB Kraft: soft: yes, this is what i was hoping too, and to be able to easliy identify and implement those things when i, and like me, can find a window
[14:18] Carjay McGinnis: but at the moment even little tiny patches seem to get lost so I wonder how this is going to work on LL side
[14:18] Soft Linden: Right.
[14:18] Carjay McGinnis: it requires a lot more commitment than just looking through a list of possible patches
[14:18] Soft Linden: So the first question might be how much interest there would be from Linden devs if there were some kind of a streamlined process
[14:19] Q Linden: streamlined is the problem, soft
[14:19] Rob Linden: we've talked about having a patch mailing list which gets PJIRA automails associated with patches
[14:19] Soft Linden: Something like a proposal step, a Linden adoption step, an implementation step, and a review step. So a Linden owns it before the work happens, and promises to review it afterward.
[14:19] Rob Linden: the problem is that I'm not sure the best way of configuring JIRA to do that
[14:20] Carjay McGinnis: hm, most people don't like additional work if they don't see a benefit from it
[14:20] Carjay McGinnis: not sure how Linden devs actually think about such an idea or if they are even aware of "outside" contributors
[14:20] JB Kraft: what we do a lot is grep TODO's right out of the code and put them in a list, but in this case, that does facilitate 2 way communication back into the linden ship
[14:21] Rob Linden: Carjay, you realize there's a bunch of Linden developers sitting right here, right? :)
[14:21] Carjay McGinnis: sure, I am always talking about those that we never see here :)
[14:21] Poppy Linden: i think everyone is "aware" of OSS dev
[14:22] Carjay McGinnis: what I meant was awareness that people might want to contribute by doing what JB proposed
[14:23] Carjay McGinnis: not simply just hack at the code for themselves
[14:24] Rob Linden: k....well, I think you should consider the "sldev" mailing list your list for exactly that activity
[14:24] Q Linden: I think it might be valuable for us to suggest internally that when a dev feels a pain that could use some cleanup attention, instead of either forgetting it or doing it themselves, we might consider asking os devs to tackle it
[14:24] Rob Linden would much rather see traffic on this topic than the traffic that often dominates the list
[14:24] Carjay McGinnis: sure and there have been some interesting discussions lately, that's true
[14:25] Soft Linden: The string cleanup work that ate a week of Steve's time recently would have been a great example of this.
[14:25] Rob Linden: Putting something in PJIRA plus sending mail to sldev is a *great* way of advertising "hey, I"m working on this"
[14:25] Soft Linden: I'm sure he'd have much rather spent a day reviewing patches!
[14:26] Q Linden: sometimes that sort of task is actually a relief
[14:26] Q Linden: but that's also one where we'd be really unlikely to take a patch if it touches 300 files
[14:26] Rob Linden: I think a wiki page or a metaissue in pjira would be good as well
[14:26] Rob Linden: so, here's how I see something like this working:
[14:26] Poppy Linden: right now we just do a back and forth with patches on a specific jira
[14:27] Poppy Linden: that's not what we do internally - we use screen or VNC
[14:27] Rob Linden: someone looking for cleanup projects goes to this wiki page/jira metaissue to find something to do (or has an idea themselves)
[14:27] JB Kraft: soft: yes, this came to me as i was in the midst of creating a class to handle dynamic grid lookups for the login panel, appended to by a grids.xml file in user settings. between 1.20.5 and 1.20.8 someone internally started it too and kind of toasted my work :/ its was only a couple of hours work, but begat the question
[14:28] Rob Linden: (that was "1.")
[14:28] Rob Linden: 2. if the issue isn't already in JIRA, they file the issue
[14:28] Poppy Linden: JB, i think that was mani
[14:28] Rob Linden: 3. They send out mail to sldev saying "hey, I'm working on this", with a link to the JIRA ticket and a synopsis
[14:28] Poppy Linden: ping him and ask about it
[14:29] Rob Linden: 4. Discssion might ensue
[14:29] Q Linden: s/might/will/
[14:29] Rob Linden: 5. They assign it to themselves (that's probably 2a rather than 5)
[14:30] Soft Linden: With these though, again - it would be great to hook up the proper Linden early on in the cleanup. Especially in the CAKE era, where someone (or some group) is more likely to "own" that section of code.
[14:30] Rob Linden: 6. If someone else is curious about progress, they comment on the issue
[14:30] Carjay McGinnis: CAKE?
[14:30] Q Linden: yes, in our new organization we should be less likely to drop things
[14:30] Carjay McGinnis: sounds tasty
[14:31] Soft Linden: om nom nom nom.
[14:31] Soft Linden: Internal structuring to better section off parts of Second Life development - it's getting a bit large for everyone to be a jack of all trades 24/7.
[14:32] Carjay McGinnis: sure, thought that would have been already the case
[14:32] JB Kraft: do you folks (devs) check the JIRA before starting these janitorial type things tho? i cant imagine you do for something smallish
[14:32] Carjay McGinnis: special groups for certain areas of the code I mean
[14:33] Liana Linden: We've had different grouping/sectioning strategies in the past. It's important to keep evolving.
[14:33] Poppy Linden: jb, depends on the person
[14:33] JB Kraft: ic
[14:33] Rob Linden: there is a separate discussion to have about Linden's use of JIRA, which we'll also try to figure out how to be more transparent here
[14:34] Rob Linden: ...but I'm trying to stay focused on things that everyone here can do right now to help prevent duplication of effort
[14:34] Poppy Linden: i think it helps to be able to ask
[14:36] Rob Linden: I know it'd be preferable if everyone knew what every Linden dev was working on in perfect coordination, but I think the proposal I made above isn't much different than what I'd tell a Linden developer to do
[14:37] JB Kraft: is there any sort of cordinated internal TODO or just jira and code comments?
[14:37] Q Linden: just jira and higher-level personal coordination within project teams
[14:37] Rob Linden: JB: it's in flux right now...we're shifting around our org to accomodate this stuff
[14:38] JB Kraft: ic, plus the oft used cubicle/watercooler access ;)
[14:39] Liana Linden: Cubicle? What is cubicle? ;-)
[14:39] Q Linden: Yeah, except our watercooler is 40 feet high and located in a private region. :-)
[14:39] JB Kraft: hehe
[14:39] Poppy Linden: our internal IRC is our watercooler, perhaps?
[14:39] Q Linden: liana, cubicle is like a prim for companies
[14:40] Q Linden: yes, poppy -- more seriously, IRC ftw
[14:40] Poppy Linden: these are legit things to wonder, our external comm channels are not close to our internal ones
[14:40] Rob Linden: we know we've got to do much more to break down the internal/external barrier of communication, and I think some org chagnes we're in the middle of will help us get to that point
[14:40] Poppy Linden: but at the same time, our external ones are full of rants and vitriol ;_;
[14:41] JB Kraft: rob: thats great to hear from out here on the ocean :)
[14:41] Carjay McGinnis: at least the sun is shining
[14:41] Soft Linden: The short of all this, though - it should be much easier to find the "right guy" for a specific area you want to patch against.
[14:41] Rob Linden: so....I think you all have an opportunity to lead by example here
[14:42] Rob Linden: if you can start following the steps above, there's a couple things that I and the people here can commit to help out on
[14:42] JB Kraft: soft: absolutely
[14:43] Poppy Linden is melting his CPU rn, gtg
[14:43] Rob Linden: 1. we can interject on behalf of community members when we identify internal/external overlaps
[14:43] Rob Linden: (I should have said "the Lindens here can commit to"
[14:44] Rob Linden: 2. we can advocate following the process ourselves.
[14:44] Rob Linden doesn't know what the devs here would think about following this process for themselves
[14:45] Q Linden: if we made a cleanup master jira and cycled it every quarter, I think people would probably try to use it
[14:46] Rob Linden: we can certainly do #1, which I think is of most interest to everyone here
[14:47] Rob Linden: JB: I'm assuming that gets to the heart of the problem, right? You want to be able to say "I'm working on X", and have a reasonable amount of confidence that if you have a track record of following through, you'll have something that we'll incoroproate at the end
[14:47] Rob Linden: right?
[14:49] Carjay McGinnis: also in the other direction, not spending time to look at things that are being worked on
[14:49] JB Kraft: yes, but not so much that. from out here, if i put some time into the code, the reward really is that i have made it better somehow. to have it not even in contention for that possibility can be rather frustrating
[14:50] Q Linden: rob, if such announcements of "hey, an OS dev is working on cleaning up the String class" went out to an internal list, it might at least get noticed by someone who had it on their plate
[14:50] Rob Linden: Q: exactly
[14:51] Carjay McGinnis: hm, an undertaking like cleaning up the String class that would need very good coordination though
[14:51] Rob Linden: Carjay: that's a great example of the ensuing discussion
[14:51] Rob Linden: :)
[14:51] Carjay McGinnis: at least it's probably not something I'd just start on my own as a OS guy
[14:51] Q Linden: righto...bad example, but the right premise
[14:52] Rob Linden: setting the bar a little lower, you want someone to stop you when you're about to waste your time
[14:52] JB Kraft: rob: to put it in the other direction, to have an internal devnote the issue, even if it's not being worked at the moment, i can identify it and offer a solution perhaps
[14:52] Carjay McGinnis: indeed, just an exchange of ideas would be great, too
[14:53] Rob Linden: JB & Carjay: absolutely. I agree.
[14:53] Rob Linden will push for that, trust me
[14:54] Rob Linden: I just want to make it clear that no one here is in the position to decree that this shall be
[14:54] Rob Linden: so everyone here should think about:
[14:54] Rob Linden: 1. how to lead by example
[14:55] JB Kraft: rob: no worries. thank you, and all, for the consideration of it at least :)
[14:55] Rob Linden: 2. how to make it fun and worthwhile
[14:55] JB Kraft: #2 +1 :D
[14:56] Rob Linden: even if you all don't go through the overhead of putting it in JIRA, just a note to sldev is VERY welcome
[14:56] Rob Linden: we love it when people talk about code they are actally planning to write on sldev
[14:57] Carjay McGinnis: absolutely
[14:58] Rob Linden: there's a time and place for constructive criticism of Linden Lab, but it seems sometimes that the community consensus view is "all the time, in great volumes, on sldev" ;-)
[14:59] Rob Linden: not that the people here do that (/me looks around)
[14:59] Liana Linden: I think the group here today is the exception to that habit (won't call it a rule). All the more reason for this group to raise the level of the dialog.
[15:00] Jei Desoto: There's so much to do, so little time, and not always the most help compared to the size of things that need doing... Unfortunately the masses won't see this, and the masses are average people unknowing of how complicated the situations can be...
[15:00] Rob Linden: in fact, several of you already do a tremendous job of leading by example.....we're trying to figure out how to reinforce that
[15:00] JB Kraft: id rather write code then talk about writing code myself ;)
[15:01] Carjay McGinnis: communication is never easy, sure
[15:01] Soft Linden: understood - just know it doesn't need to be all or nothing. A three line summary is great - you don't have to plan every line of code in public.
[15:01] Rob Linden: JB: so would our devs :)
[15:01] Carjay McGinnis: sometimes it helps to think about ideas by just giving a short sketch
[15:01] Q Linden is doing it right now on the other machine... :-)
[15:02] Rob Linden: we're running a little over now..... anything else before we wrap up?
[15:03] JB Kraft: just once again, thanks all for your consideration :)
[15:03] Rob Linden: thanks for bringing it up JB
[15:04] Rob Linden: thanks everyone!
[15:04] Wilton Lundquist: I'm new to the SLDev list, and this certainly helped bring more of what folks are talking about into context for me. Thanks.
[15:04] Soft Linden: Is that an Iron Man power pack, Jei?
[15:04] Q Linden: welcome, wilton
[15:04] Jei Desoto: i believe so... a friend made it for me, so i wear it <: )
[15:04] Soft Linden: Ha - very cool
[15:04] Liana Linden: Welcome aboard, Wilton.
[15:05] Rob Linden: welcome!
[15:05] Carjay McGinnis: oh, one thing, what do you use that anchor for, Sot?
[15:05] Rob Linden: bye everyone!
[15:05] Soft Linden: To hold the rabbit down *sage nod*
[15:05] JB Kraft: ciao!
[15:05] Jei Desoto: Have a good one : )
[15:05] Carjay McGinnis: thanks and bye Rob
[15:05] Carjay McGinnis: lol
[15:05] Soft Linden: Ever seen a Linden in a strong breeze?
[15:05] Carjay McGinnis: thought so
[15:05] Liana Linden: Soft?!
[15:05] Jei Desoto: Heh Bye...
[15:05] Liana Linden: Am I the only one who doesn't get that?
[15:05] Q Linden: bye
[15:06] Jei Desoto: <: )
[15:06] Carjay McGinnis: lot of harsh winds out there
[15:06] Wilton Lundquist: I didn't get it, if that helps.
[15:06] Teleport completed from http://slurl.com/secondlife/Hippotropolis/248/18/27
[15:06] The region you have entered is running a different simulator version. Click this message for details.

Open Source Meeting/2008-05-29

Agenda

Transcript

Navigation menu

Search