Remember the purpose of the SVN zips? Authors need feedback from people willing to test the addons, and submit information about bugs they find.
What about automating this. This is something that we could do as a community, and all benifit from.
The concept would work like this
The curse client, or some other tool, could collect certain logs and place them in a large pool. From this we can find ways to query the data and correlate certain errors with solutions - basically what we do now when someone posts an error and we say "Update to revision N". Automation of this process would also ensure that developers get all the information they need, instead of the cat and mouse game "did you delete your saved variables"?
The solutions can be propagated back to the user, as well, they can be actual solutions or requests for more data. There are a large number of privacy issues we would need to address, but the system would be opt-in, so it would be a trust relationship between the developers and the users.
WAU helped thread the community together, and it made our lives easier in one respect, but there are other processes that can be automated to make them both easier and better.
maybe something similar to buggrabber/bugsack that when an error occurs it saves it and maybe also save the SV file for the offending addon and upload the error and the sv file as an error report to wowace/curseforge
you could tie this to the users curse account and then they could tell others about the bug or us authors could look at the bugs listed by addon
The curse client, or some other tool, could collect certain logs and place them in a large pool. From this we can find ways to query the data and correlate certain errors with solutions - basically what we do now when someone posts an error and we say "Update to revision N". Automation of this process would also ensure that developers get all the information they need, instead of the cat and mouse game "did you delete your saved variables"?
The solutions can be propagated back to the user, as well, they can be actual solutions or requests for more data. There are a large number of privacy issues we would need to address, but the system would be opt-in, so it would be a trust relationship between the developers and the users.
WAU helped thread the community together, and it made our lives easier in one respect, but there are other processes that can be automated to make them both easier and better.
What do you think?
Looks good on paper. :D
Seriously, though, a type of "living" knowledge base would be excellent if someone (or a group of someones) could design and implement it well.
If I could drop a SV file from an error reporting addon manually and have it distributed to the respective authors, that would be neat. While a client is nice, I'd like the option to do it manually as well. This would alleviate the "ur loggin' mah keys!" complaints.
You've already acknolwedged the incoming "but you'll see my porn" complaints, but I don't see any issues other than that.
I think its safe to say that I and Curse would be happy to help work something up for this. It'd require a lot of design considerations to try to come up with a good round about feature set.
Honestly I don't think it'd pan out well because authors would get swamped with error reports that lack any contextual information from the users regarding how they were triggered. The only way to avoid it would be to force users to write a description of what they think triggered the error before it could be submitted.
Edit: It would also force authors to field reports regarding issues that the community can handle, like improper installation or missing libraries.
Doesn't force the author to field the report, they can just close it or whatever - "user error". Maybe we can even make it intelligent enough to pick up some things, like the common missing library error.
I agree that it'll need context though. I think the big benefit for someone like me is I don't need to find the project, log in, open a ticket, fill out the forms, etc. If it can handle that, I don't mind describing the problem in detail. I don't think it's going to make a tester out of a disinterested user, nothing can do that.
Doesn't force the author to field the report, they can just close it or whatever - "user error". Maybe we can even make it intelligent enough to pick up some things, like the common missing library error.
I agree that it'll need context though. I think the big benefit for someone like me is I don't need to find the project, log in, open a ticket, fill out the forms, etc. If it can handle that, I don't mind describing the problem in detail. I don't think it's going to make a tester out of a disinterested user, nothing can do that.
maybe do it like a firefox error report where when you get the error it will ask you what you where doing etc
Many errors that happen do not need a report, because it is a coding error, and a simple stack trace is usually enough.
However, an ingame bug reporter would be useful, something similar to the Blizzard Feedback UI for reporting bugs. The user can report what he thinks are bugs, saves it in the SV and the curse client can upload these and automatically clear the SV.
Sorting/categorizing the bugs will be painful though.
I think my intention was not to spam the authors with reports - rathter to accumulate the errors and contextual information in a central database. This could be searched by authors who either want to know if there are errors from their addons, or are looking for data on an error that they have already been told about. The submissions could contain information such as a list of addons the user has installed, and what versions they are, etc, even screen shots.
It would be opt-in by both sides. Noone would be forced into it.
Datamining for someone like curse or wowhead, is a similar concept. You send in data, and a database is produced which is quite useful.
Thats why I mention privacy issues, etc. And yes submission via web form would definately be an option.
Also, it could operate in a passive or active manner. ie. not ask any questions, just gather info. Or it could ask when errors occured. If the user has it in active mode - it could make a ticket. In passive mode it would not. It would just store it.
To the privacy issue: Maybe it could be a opt-in for specific addons so only addon authors of that specific addon can see the report. Of course this automated "issue reports" should be able to recognize duplicates.
This would most likely help a lot with missing libs and old versions, since that could possibly be caught by the client.
I wouldn't make these reports real tickets, rather a display for the author so he can resolve it if possible, or even better directly ask for more information.
I know I know, I dream a bit here, but it would be handy! :p
It would be opt-in by both sides. Noone would be forced into it.
Would have to make sure that there isn't a disconnect such that users are allowed to use the tool to file error reports if the author has no intention of reading them. Otherwise the author may be expecting to see error reports only on a forum or something and never know something is wrong because it's only being reported somewhere else (by tons of frustrated people who don't know why such an obvious bug isn't being fixed).
Would have to make sure that there isn't a disconnect such that users are allowed to use the tool to file error reports if the author has no intention of reading them. Otherwise the author may be expecting to see error reports only on a forum or something and never know something is wrong because it's only being reported somewhere else (by tons of frustrated people who don't know why such an obvious bug isn't being fixed).
Well it could also be the other community members might be willing and or able to field some of the common non code related issues.
Well, I do find it an interesting idea, and I do believe it would help a lot with tracking down reported bugs. How many times have we seen a bug reported but no error logs? Or concise way of reporting it? This way, everything would be in one place, would follow a set format, over time this could be a great method of error reporting.
Con: Doing this through the curse client would still require the CC to run in the background all the time, which many have disabled. So perhaps make it so that if they client does NOT auto-run, that upon startup those errors get sent to wherever they should go.
As a user I'd be all for sending in more stuff, if I had a way to be sure it woudn't just be me being stupid >_< (had a lot of that recently)
If as Syl said it went to central DB, maybe give an option that if you submit your error lua, then it can check against existing issues / answers and give you any solutions that it find for your exact error?
ie: Error in PRat 3.0, that just needs an update to an alpha/bets, and someone already hit it, sylv posted solution, then you get an auto-response, to update and try again.
If that makes sense. not sure what the DB load would look like though for that.
To make it easier, have it pick a mod, or mods (swatter / bugrabber) and you manually upload the lua file? Kinda like wowroster does (for those familiar) where you can use a client to upload OR you can go to ___ website and do it manually for those who don't like auto-client stuff.
To make it easier, have it pick a mod, or mods (swatter / bugrabber) and you manually upload the lua file? Kinda like wowroster does (for those familiar) where you can use a client to upload OR you can go to ___ website and do it manually for those who don't like auto-client stuff.
That beats my idea.. I love it. You have my vote there!
To make it easier, have it pick a mod, or mods (swatter / bugrabber) and you manually upload the lua file? Kinda like wowroster does (for those familiar) where you can use a client to upload OR you can go to ___ website and do it manually for those who don't like auto-client stuff.
Yes, yes, definately. I said that in my first post :P. Must have option, I don't like running clients that send out information, consider me paranoid ;).
Just because you're paranoid doesn't mean they're not watching you. ;)
Spennig's idea would also cover the power-user types who occasionally feel the need to hack an established mod 'just to see what happens'.
However, if Spennig's idea didn't make the final cut, that is something that would need to be considered. If the errors were automatically reported, how is the author to know (unless it's a very obvious change on the user end) that the user is mucking around with stuff, and this isn't a problem with the code as it is on the svn?
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumWhat about automating this. This is something that we could do as a community, and all benifit from.
The concept would work like this
The curse client, or some other tool, could collect certain logs and place them in a large pool. From this we can find ways to query the data and correlate certain errors with solutions - basically what we do now when someone posts an error and we say "Update to revision N". Automation of this process would also ensure that developers get all the information they need, instead of the cat and mouse game "did you delete your saved variables"?
The solutions can be propagated back to the user, as well, they can be actual solutions or requests for more data. There are a large number of privacy issues we would need to address, but the system would be opt-in, so it would be a trust relationship between the developers and the users.
WAU helped thread the community together, and it made our lives easier in one respect, but there are other processes that can be automated to make them both easier and better.
What do you think?
you could tie this to the users curse account and then they could tell others about the bug or us authors could look at the bugs listed by addon
-
View User Profile
-
View Posts
-
Send Message
Project AuthorLooks good on paper. :D
Seriously, though, a type of "living" knowledge base would be excellent if someone (or a group of someones) could design and implement it well.
You've already acknolwedged the incoming "but you'll see my porn" complaints, but I don't see any issues other than that.
Edit: It would also force authors to field reports regarding issues that the community can handle, like improper installation or missing libraries.
I agree that it'll need context though. I think the big benefit for someone like me is I don't need to find the project, log in, open a ticket, fill out the forms, etc. If it can handle that, I don't mind describing the problem in detail. I don't think it's going to make a tester out of a disinterested user, nothing can do that.
maybe do it like a firefox error report where when you get the error it will ask you what you where doing etc
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumHowever, an ingame bug reporter would be useful, something similar to the Blizzard Feedback UI for reporting bugs. The user can report what he thinks are bugs, saves it in the SV and the curse client can upload these and automatically clear the SV.
Sorting/categorizing the bugs will be painful though.
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumIt would be opt-in by both sides. Noone would be forced into it.
Datamining for someone like curse or wowhead, is a similar concept. You send in data, and a database is produced which is quite useful.
Thats why I mention privacy issues, etc. And yes submission via web form would definately be an option.
Also, it could operate in a passive or active manner. ie. not ask any questions, just gather info. Or it could ask when errors occured. If the user has it in active mode - it could make a ticket. In passive mode it would not. It would just store it.
To the privacy issue: Maybe it could be a opt-in for specific addons so only addon authors of that specific addon can see the report. Of course this automated "issue reports" should be able to recognize duplicates.
This would most likely help a lot with missing libs and old versions, since that could possibly be caught by the client.
I wouldn't make these reports real tickets, rather a display for the author so he can resolve it if possible, or even better directly ask for more information.
I know I know, I dream a bit here, but it would be handy! :p
Would have to make sure that there isn't a disconnect such that users are allowed to use the tool to file error reports if the author has no intention of reading them. Otherwise the author may be expecting to see error reports only on a forum or something and never know something is wrong because it's only being reported somewhere else (by tons of frustrated people who don't know why such an obvious bug isn't being fixed).
Well it could also be the other community members might be willing and or able to field some of the common non code related issues.
Con: Doing this through the curse client would still require the CC to run in the background all the time, which many have disabled. So perhaps make it so that if they client does NOT auto-run, that upon startup those errors get sent to wherever they should go.
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumIf as Syl said it went to central DB, maybe give an option that if you submit your error lua, then it can check against existing issues / answers and give you any solutions that it find for your exact error?
ie: Error in PRat 3.0, that just needs an update to an alpha/bets, and someone already hit it, sylv posted solution, then you get an auto-response, to update and try again.
If that makes sense. not sure what the DB load would look like though for that.
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumThat beats my idea.. I love it. You have my vote there!
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumSpennig's idea would also cover the power-user types who occasionally feel the need to hack an established mod 'just to see what happens'.
However, if Spennig's idea didn't make the final cut, that is something that would need to be considered. If the errors were automatically reported, how is the author to know (unless it's a very obvious change on the user end) that the user is mucking around with stuff, and this isn't a problem with the code as it is on the svn?