Active Topics
-
Porting apps to Leste (29)
to Maemo 7 / Leste by Arno_11 - 3 days, 9 hrs ago -
Install Fonts in N900 (68)
to Maemo 5 / Fremantle by teroyk - 4 days, 12 hrs ago - more...
|
2010-01-18
, 19:26
|
|
Posts: 1,418 |
Thanked: 1,541 times |
Joined on Feb 2008
|
#22
|
Originally Posted by VDVsx
Could you give a short summary of the plans Flandry is talking about? What are they, who made them, and how many people agreed to them?
I'm responsible for the coordination of these improvements, but unfortunately Neils(X-Fade) that will implement the improvements is busy with things that have more priority at the moment. I think he will accept help if someone want to jump in and contribute some code.
|
|
2010-01-18
, 19:29
|
|
Posts: 329 |
Thanked: 142 times |
Joined on Oct 2009
|
#23
|
Originally Posted by Flandry
Actually, i think that's the problem: there are no individuals relied upon. Users are in fact discouraged from testing by 'there-be-dragons' warnings. No wonder when i peek around the package list and look at comments, i only see a few familiar names/faces
The relying on individuals doesn't seem to be getting us far, so maybe it's not a good strategy?
I have a proposal: make -devel the snake lair. Regular users will be discouraged from using it, but those with the know will be urged to try it more, to compensate. The main objective would be to make sure the app doesn't kill the device or make it go crazy.
Once the app is declared sane, it can go to -testing. Since there are (thanks to the above paragraph) now a lot more users, things get tested quicker and more thoroughly (law of large numbers says faults come out better on a larger testing base).
And the normal extras gets the high-quality apps that should be showcased. That's not to say they can't have any bugs, but those should be minor and more like annoyances
Also, +1 for official testers
__________________
I don't mean to crush your hopes, I just can't help it
I don't mean to crush your hopes, I just can't help it
 |
|
2010-01-18
, 19:31
|
|
Posts: 1,559 |
Thanked: 1,786 times |
Joined on Oct 2009
@ Boston
|
#24
|
Ok would you please point out where this is at? Gitorious?
__________________
Unofficial PR1.3/Meego 1.1 FAQ
Unofficial PR1.3/Meego 1.1 FAQ
***
Classic example of arbitrary Nokia decision making. Couldn't just fallback to the no brainer of tagging with lat/lon if network isn't accessible, could you Nokia?
 |
|
2010-01-18
, 19:37
|
|
Posts: 903 |
Thanked: 632 times |
Joined on Apr 2008
|
#25
|
Re: The quarantine issue:
I think that the 10 day quarantine just slows down the whole system. It certaintly slows down me.
I promoted a package to testing that had a rather serious bug in it: one of the dialogs in the settings interface would not open. It was a rather small issue, because I'm not sure that hat many people would have a need to open that dialog, but for those that did try, it would be very confusing.
After 3 or 4 days into the testing process someone raised it to my attention and it was a simple bug, one line needed changing.
I did this and uploaded a fixed version to -devel. Problem was, this was during a builder outage and the builder would not build it. Finally three or four days later I got the package built and went to promote it to -testing. I saw the version with the bug in it that had been sitting in -testing had 11 karma. The 10 day quarantine was going to be up soon, but should I promote it to Extras and then queue the new version into -testing? With the bug?
I decided against it and threw away 9 days of quarantine and 11 positive votes for a one line change. This fixed version got the required karma in a short amount of time, but am still waiting for the quarantine to be up... (which I think is completely useless).
It's bad enough you have to get 10 positive votes even for a 1 line bug fix, why delay it any further?
Additionally, it says on the wiki:
http://wiki.maemo.org/Help_testing_software
So, how are any bugs going to be found in +10 karma apps if people aren't even testing them?
I think that the 10 day quarantine just slows down the whole system. It certaintly slows down me.
I promoted a package to testing that had a rather serious bug in it: one of the dialogs in the settings interface would not open. It was a rather small issue, because I'm not sure that hat many people would have a need to open that dialog, but for those that did try, it would be very confusing.
After 3 or 4 days into the testing process someone raised it to my attention and it was a simple bug, one line needed changing.
I did this and uploaded a fixed version to -devel. Problem was, this was during a builder outage and the builder would not build it. Finally three or four days later I got the package built and went to promote it to -testing. I saw the version with the bug in it that had been sitting in -testing had 11 karma. The 10 day quarantine was going to be up soon, but should I promote it to Extras and then queue the new version into -testing? With the bug?
I decided against it and threw away 9 days of quarantine and 11 positive votes for a one line change. This fixed version got the required karma in a short amount of time, but am still waiting for the quarantine to be up... (which I think is completely useless).
It's bad enough you have to get 10 positive votes even for a 1 line bug fix, why delay it any further?
Additionally, it says on the wiki:
Finding an App to Test
The applications waiting for testing are listed here by age (the oldest is first). When picking an app to test, note that those that already have 10 or more karma are "done". Don't waste your time on those. Ditto if the package has < -5 karma: such an app obviously needs to be fixed and re-released.
So, how are any bugs going to be found in +10 karma apps if people aren't even testing them?
__________________
-Brent
Author of TouchSearch -- web searching software for Maemo 5.
Mobile Device lineage: Palm Z22 -> Palm TX -> Nokia N800 -> Nokia N900
-Brent
Author of TouchSearch -- web searching software for Maemo 5.
Mobile Device lineage: Palm Z22 -> Palm TX -> Nokia N800 -> Nokia N900
|
|
2010-01-18
, 19:43
|
|
Posts: 1,559 |
Thanked: 1,786 times |
Joined on Oct 2009
@ Boston
|
#26
|
Originally Posted by BrentDC
Well, that bit of wiki was my own edit from a while back when trying to streamline the process for new testers. The quarantine does still allow time for serious bugs to come out, but i think this is a good case for official testers covering the full spectrum of tests and user testers doing casual testing via usage during the rest of the process.
So, how are any bugs going to be found in +10 karma apps if people aren't even testing them?
__________________
Unofficial PR1.3/Meego 1.1 FAQ
Unofficial PR1.3/Meego 1.1 FAQ
***
Classic example of arbitrary Nokia decision making. Couldn't just fallback to the no brainer of tagging with lat/lon if network isn't accessible, could you Nokia?
|
|
2010-01-18
, 19:45
|
|
Posts: 5,795 |
Thanked: 3,151 times |
Joined on Feb 2007
@ Agoura Hills Calif
|
#27
|
Originally Posted by fms
Amen. If those are the worst problems experienced lately after the dire warnings made over and over, I would say the current process is basically trouble-free. Except that programs that should be released to users haven't been. Since there is overprotection (a useful chess term!), some people have basically ignored the warnings. This is the same thing that happens when cigarette taxes get too high -- smuggling starts.
Well, honestly, I would forgive the Load applet, no screencasts is no big deal. The Nintendo wallpaper thing has been obvious from the first day, I have no idea why nobody at Nokia noticed it. The wallpapers were technically ok though, so they passed through the approval process. This leaves Radio FM messup. Not much in terms of problems.
I do not think there is anything you can do to increase the number of people testing stuff. "Power users" generally do not test and they do not vote. They just use, even if it means enabling potentially unsafe repositories. Making every voter go through a checklist just reduces the number of people who vote. People who do go through checklists appear to be a very special kind of people, more like career Wikipedia editors. These are rare and they are weird.
So, as you might have guessed, I am very skeptical on any "improvements" anyone is going to bring to the Extras vetting process. If anything, we need less enforcement, not more.
Whatever enhances the collaborative process between users and developers should be encouraged. Whatever discourages the collaborative process should be abandoned. Regulations should be minimized, and carefully chosen.The ultimate aim is NOT protection. The ultimate aim is fun or useful programs.
|
|
2010-01-18
, 20:08
|
|
Posts: 1,559 |
Thanked: 1,786 times |
Joined on Oct 2009
@ Boston
|
#28
|
Originally Posted by fms
I think you missed the first page of the thread. You came in about the time i started throwing out pie-in-the-sky hypothetical suggestionsl because that's what was coming back at me. This is a fairly basic problem with fairly basic needs, as you and geneven and others have pointed out in response to my bait: we need to improve th testing process so it's not a barrier.
Sorry, but what plans are you talking about? I have not been agreeing to any plans so far.
Call me stupid, but I do not quite understand how Linux and open source are related to your plans for Extras control mechanisms. Would you be so kind to elaborate?
I started the thread to request that portions of the process that were discussed to be part of it actually be added. We don't really need ten testers; we need a thorough test of the app and 10 seemed like an arbitrarily good way to ensure that. We do't need 10 days of quarantine... " " "
So, let's deputize the really active testers as "official", give them pet criteria to cover if they want, make the new requirement 3 or 2 or 1 official tester who will verify all points as assigned, and leave the other testing to power users during the quarantine period.
And let's not make new versions go through the same process as first-time apps.
Edit: as for the individual vs community example, it's the difference between relying on one person (xfade in this case) vs a team or community to get things done. He's "busy with other, more important things" which says to me right there that there is too much hanging over that one person's head.
__________________
Unofficial PR1.3/Meego 1.1 FAQ
Unofficial PR1.3/Meego 1.1 FAQ
***
Classic example of arbitrary Nokia decision making. Couldn't just fallback to the no brainer of tagging with lat/lon if network isn't accessible, could you Nokia?
Last edited by Flandry; 2010-01-18 at 20:11.
|
|
2010-01-18
, 20:22
|
|
Posts: 1,418 |
Thanked: 1,541 times |
Joined on Feb 2008
|
#29
|
Originally Posted by Flandry
And that was the very first post at the very first page of the thread.
I think you missed the first page of the thread. You came in about the time i started throwing out pie-in-the-sky hypothetical suggestionsl because that's what was coming back at me.
we need to improve th testing process so it's not a barrier.
We don't really need ten testers; we need a thorough test of the app and 10 seemed like an arbitrarily good way to ensure that. We do't need 10 days of quarantine...
So, let's deputize the really active testers as "official", give them pet criteria to cover if they want, make the new requirement 3 or 2 or 1 official tester who will verify all points as assigned, and leave the other testing to power users during the quarantine period.
And let's not make new versions go through the same process as first-time apps.
Edit: as for the individual vs community example, it's the difference between relying on one person (xfade in this case) vs a team or community to get things done. He's "busy with other, more important things" which says to me right there that there is too much hanging over that one person's head.
|
|
2010-01-18
, 20:38
|
|
Posts: 1,559 |
Thanked: 1,786 times |
Joined on Oct 2009
@ Boston
|
#30
|
Originally Posted by fms
The weird people who test or edit wikipedia (much like the weird people who spend time arguing about computer platforms on an Internet forum) are exactly what we need in that role, and empowering them to follow the letter of a QA checklist is a great way to assure QA. That you don't agree with one or more items on the QA checklist is a different issue that ought to be pursued in a discussion on the topic of what constitutes accepable quality in an Extras app. The issue at hand is getting the established guidelines for Extras QA verified in a more expeditious manner. That was the thesis of my first post (but not the one you replied to, which you still seem to think was the OP) and is still my goal with this thread: rattle cages or squeak wheels until the means of getting the actual process in agreement with the agreed process is clear.
Well it has long been agreed that 5 votes is usually sufficient. In fact, it was agreed in an IRC meeting months ago.
...and the only people willing to be "official testers" are going to be above-mentioned career wikipedia editors, then ones who vote package down "because it has no bugtracker URL" (kinda nonsensical, as each package has maintainer's name and email address).
This, again, has been agreed months ago.
__________________
Unofficial PR1.3/Meego 1.1 FAQ
Unofficial PR1.3/Meego 1.1 FAQ
***
Classic example of arbitrary Nokia decision making. Couldn't just fallback to the no brainer of tagging with lat/lon if network isn't accessible, could you Nokia?
 |
| Tags |
| extras-tesing, finishing the job, quality assurance, quarantine, software quality, user testing |
«
Previous Thread
|
Next Thread
»
|
All times are GMT. The time now is 05:47.
Computer and Console Emulators for Maemo, Symbian, and more.
Pack Rat: The Maemo Package Aggregator.