JIRA Change Follow Up

Back in September how the JIRA works changed. Now we can’t really know what’s going on with bugs and other problems in Second Life™ because most of us cannot read the JIRA. We cannot browse or search the JIRA to see if a problem we are encountering has been reported. So, we just report our problem.

Nor do we know if a problem is fixed. Unless the problem is something we experience almost daily and one Tuesday or Wednesday it suddenly starts working. For users the JIRA Change is a pain.

SL JIRA 30 Day BUG Stats

SL JIRA 30 Day BUG Stats

Various parts of the JIRA work different ways. Some projects are open and anyone can read and comment on the items filed in those projects. Others are closed, like BUG, and people can only read and comment on the items they file and only until they are triaged. Only a few can read most of the JIRA projects. You can read my initial announcement here: Oh CRAP! JIRA Change (Sept 7, 2012).

Darien Caldwell started reporting SL JIRA’s in the SLUniverse forum so they would be readable by anyone. There we can comment on and search for JIRA’s. But, searching is not all that handy on forums. Also, using Google to search the forum has disadvantages.

Fortunately the SLUniverse forum uses tags and we can search by or with a tag. That seems to show the last JIRA item update was November 11, 2012. From the chart above you can see that about 300 items are filed per month in the SL JIRA. That suggests the SLUniverse forum is missing about 600 JIRA’s. So, the JIRA’s at SLUniverse do not provide a good substitute for a real JIRA search.

I’ve always thought getting people to double file a JIRA (file in the SL JIRA and then again at SLUniverse) was an improbable task… noble but unlikely to work. Far too many people using Second Life are unaware of the SL blogosphere and related forums. Some significant number of users are even unaware of the SL forum.

The Firestorm JIRA is another open JIRA where problems are reported. They tend to be oriented to the Firestorm and Phoenix viewers. But, some are SL related. Whatever, the number of bugs reported there is about half of what we see in the SL JIRA.

FireStorm JIRA Stats

FireStorm JIRA Stats

So, we simply don’t have a good alternative to the SL JIRA.

Reconsideration Likely?

We are now 6 months down the road. The Lindens said they would reconsider the JIRA down the road. None were clear about when and probably none knew/know when a ‘reconsideration’ might be made, if that has not already been done.

I suspect the way things go is a ‘reconsideration’ would happen sooner if there were problems… apparent problems… with pain on the Linden side increasing the priority. Also, stats would tell the Lab’s management whether things were working. The chart above looks pretty good.

We can see most of the 305 items filed have been handled. All but 71 items have been handled as I write this. Also, some number of these items are duplicates. While many have been concerned about duplicates over loading the system I think it obvious that is not going to happen. Dealing with the 10 or so items filed per day is no big deal. Even if we take out weekends, that is still just 15 or so per day. Not exactly an overwhelming flood.

Plus duplicates can be marked handled pretty quickly. Say one does get 30 about 1 problem and 40 about another problem in a week. Those 2 problems would be entered and 68 items would be marked handled and closed. Hey, that is a support team that is really cranking through them…

It isn’t that bad, statistically at least. I am sure management knows that numbers game. Plus these public bug reports are taken into a private section of the JIRA known as MAINT. Those 70 reports in the BUG project would condense into 2 reports in the MAINT project. Performance would likely be measured on MAINT by management.

If you are following STORM-1716 (the Mesh Deformer item), you can see all the problems happening that the Lindens hoped to reduce. That JIRA thread has turned into a huge JIRA item that no one reads. Now complaints and ideas are repeating as new people fail to read and inform their selves before commenting or complaining.

Even those of us that follow the thread are burning out on it. So, it is obvious 1716 is the infamous remaining poster child for why the JIRA changed. I can only imagine how depressing it would be to work with such a mess every day. Trying to get real information out of such items would be a huge time waster.

Pain on the user side is pretty much immeasurable. Nor can we measure the number of reports that are not filed because people are unhappy with the change. Nor could we measure how many were not filed for whatever reason before the change. So, there is no objective comparison point.

I doubt there is any way to decide how many reports were just noise and whether that has changed. But, I do think one can measure in a general way the number of on and off topic or helpful comments added to JIRA’s. Since only the original reporter of a problem can comment on it now, I suspect comments are much more on topic. There are also far fewer comments, meaning less to read and time spent reading JIRA reports is likely way down.

Looking at things from the Linden side of the JIRA, things probably look pretty good. Form our side it is a problem only now and then. I see the degradation in the service the JIRA can provide. But, there is no way to quantify that such that Linden management can evaluate it. It is just invisible to them.

So, I doubt we will see any change in the JIRA system.

There are Problems

From outside the Lab we can see problems but we cannot quantify them. I think these problems are contributing to a loss of users and certainly a loss of people willing to own regions.

Events have been having problems for some time now. I’ve covered Toysoldier’s problems and complaints regarding the event crashes and massive lag since September. See: New Lag Problem.

The Lindens are busy with their assigned problems and work. Getting their attention is difficult. All the indicators of problems have changed and become rather restricted. We have no way to signal how extensive the pain is on our side of the system. The Sudden Massive Lag is an excellent example.

If a problem does get the Lindens attention, we have no way to know that. So, if they think they have it fixed, there is no good feedback system between the two sides of the problem. So, on one side the Lindens are all Yay! I fixed it! And the user is all OMG! I’m dying over here! Help.

Yeah, a problem with a specific symptom was fixed. But, the user is still experiencing the same symptom. The time to communicate the problem and get the Lindens re-started on a search for the cause is time consuming on our part as we have to start over. Plus we have to get it by the triage people that may often think it is a duplicate of a resolved problem. And… it may be solved and the solution is winding its way through QA. But, none of that information is communicated.

From the user side it is horribly frustrating.

So, we have a problem, that for all practical purposes, is invisible to the Lindens. I have no doubt the JIRA Change has improved efficiency and working conditions for the Lab.

From my viewpoint JIRA Change has reduced user feedback and communication. It is another hidden-from-the-Lindens contributor the degradation of the user experience and a drag on player retention.

The sad part is there is no way to quantify the problem in an objective way. That means there is no way to resolve the problem or even convince the Linden management of the existence of the problem. So, I say that unless we have a perceptive person in management at the Lab, the JIRA Change will become permanent.

Even more likely to make this change permanent is there is no solution offered to solve the problems that originally precipitated the JIRA Change.

10 thoughts on “JIRA Change Follow Up

  1. One possible solution would be to copy all bug reports to a public section of the JIRA or some other format. There the number of comments, number of unique commentors, votes, etc. could be used to indicate resident pain / interest levels. The Lindens would only be expected to look at these pages in aggregate not read every post.

    You are very correct, our feedback to LL has nearly vanished, they are working in a nearly opaque bubble!

  2. ERROR: That response to the Solve Media puzzle was incorrect.
    I should enter a Jira for that, but I won’t since I have stopped using the “New And Improved Jira”.

  3. I would like to make two points about your analysis.

    First, issues are “resolved” in the BUG project when they are “released” – that is cloned into an internal jira. Thta has nothing to do with them being resolved in the usual sense of the word, either by fix or decision to ignore. So that graph is merely an indicator of the triage process and doesn’t tell us anything at all about what’s actually happening to bug reports. I know you are aware of that, but it needs to be emphasized.

    Secondly, I think you told us before that you are one of those with privileged access, and therefore you may not be aware that when BUGs are “released”, the originator cannot add further comments. I have collected a large amount of clarifying data about large inconsistencies in physics weights which I cannot introduce into the relevant jiras for that reason.

    Similarly, I cannot offer counter arguments to what seem to be Lab comments based on misunderstanding of the reports. No doubt these are my fault for not anticipating and being sufficiently painstaking in the reports, but now I cannot correct that. Taking these effects together, I am increasingly feeling that it is no longer worth the effort of trying to help the developers this way. If the intention of the changes was to decrease user effort in bug reporting to ease their burden, then it has indeed been successful in my case.

    • I agree the chart is only representative of triage, as you say.

      While triage locks out even the original filer, an email to Alexa can get the item reopened.

      Misunderstanding is a problem. Knowing what to tell the techies is a bit of a mystical art. Without feedback from the techie and some conversation, neither side understands for sure what the other is talking about.

  4. Maybe I’m thick, Nalates, but it seems to me that making JIRA read-only would’ve accomplished the same thing from the Lindens’ standpoint without leaving residents in the dark.

    I’d much rather see the banhammer liberally wielded than residents silenced altogether, but apparently nobody has the spine.

    • I agree and think that would have been better.

      But, thinking the choice was made from a lack of spine is rather harsh. How the Lindens were treated may have had much to do with it too.

  5. I must recognize that the deformer mesh is a huge mess. But is also true that everytime a Linden or Qarl postead an update about a feature or any other change, they could also update the main jira so people dont need to scroll down and read the whole bunch of comments. The jira is outdated in comparison with the information given later on the comments. I know that Oz Linden reads the comments because I can see his answers from time to time. That means that everytime that info were given, a little update to the main post could be very helpful.

    And for the rest of Jira, I am not really sure what they prettend in first instance with this whole change. If they just want to evade the flood of comments, just make the Jira readable only except for the person who made them. Anyway, still havent too much sense. The flood of comments can be now replaced by the flood of bugs reported. But well, you and me know that most people dont pot bugs at all and even less now with tis situation. But of course, less bugs reports doesnt mean a bug free product. This are just making everything slower but maybe is what they want so they have more time fixing bugs from 2006.

    • Updating the status in the first post requires time and effort on someone’s part. If Oz and Karl are updating then they are not getting other tasks completed. If people want to be informed they need to put in the time. Shifting that responsibility is not a solution. But, people are people. While I’m not happy with the present system I am convinced it works better for the Lindens daily work flow. But, as I said, I’m not sure this solution is helping Second Life. But, that is opinion and hard to substantiate, as I pointed out.

      Several have suggested they could have made the JIRA read only. But, each person that has done that has also agreed there are problems from people not reading. Also, read only would not stop the number of duplicate/repeated reports as people try to add to a report. So, while I think that would have been better, I don’t know it would help that much, even if I would prefer it.

      It is obvious from the stats we are not getting a flood of bug reports replacing the comments.

      I don’t think we have any hard data to say we get fewer reports now.

      In 2011 Lindens were going through all the old bugs and clearing them out of the system. While we may have problems with similar symptoms, I doubt they are from the same software bugs as in 2006.

Leave a Reply

Your email address will not be published. Required fields are marked *