email address as bugtracker link for non-small projects?

On 25/08/10 18:29, Andrew Flegg wrote:
> I am strongly against this, for the reasons outlined above. We
> shouldn't be putting up *unnecessary* barriers to someone getting
> their application into Extras, and this *is* one of those, IMHO.

Well reasoned

+1

David

--
"Don't worry, you'll be fine; I saw it work in a cartoon once..."

Am Mittwoch, den 25.08.2010, 13:00 +0200 schrieb Andre Klapper:
> http://wiki.maemo.org/Packaging#Bugtracker_location states
> "It is possible to [...] use [...] an email address."
> No mentioning of size of project.
>
> http://wiki.maemo.org/Extras-testing/QA_Checklist#Lack_of_bug_reporting_database states "An e-mail address for small projects (wallpapers, for example) can be used".
>
> Now which one is correct? (Plus the same topic/info with slight
> differences spread on different pages is not cool.)

There is also a third page in the wiki which should not be forgotten after the discussion is done:
http://wiki.maemo.org/Bugtracker

There also is written "just for small applications".

Best regards Uwe

> On 25/08/10 18:29, Andrew Flegg wrote:
>> I am strongly against this, for the reasons outlined above. We
>> shouldn't be putting up *unnecessary* barriers to someone getting
>> their application into Extras, and this *is* one of those, IMHO.
>
> Well reasoned
>
> +1
>
> David

Agreed (Saves me lot of typing and arguing :-)
E-Mail is enough. A channel there must be, but this suffices.

Best regards
-Tom (1.5% of all Maemo downloads, no real bugtracker :-)

Am Mittwoch, den 25.08.2010, 18:29 +0100 schrieb Andrew Flegg:
> On Wed, Aug 25, 2010 at 18:06, Uwe Kaminski <jukey@ju-key.de> wrote:
> > As written a few weeks ago at the testingsquad mailing list:
> > https://garage.maemo.org/pipermail/testingsquad-list/2010-August/000116.html
>
> However the Testing Squad is a group for applying the policy, not
> making it. The correct place for the discussion is here.

Thanks for the hint. I did not know that. A lot of discussions regarding
the policy are done at the testingsquad mailing list [1]. There are also
some of the rules discussed in the wiki [2]. In addition to that I was
unable to find a clear definition what the testingsquad list is for. I
suggest to define it better here:
http://wiki.maemo.org/Community_mailing_lists

> >> Beside all technical reasons which make a reporting solution failing...
> >> I never would force developers to do something or not.
>
> Except you are?

As explained I would do this differentiation for extras-(testing) only.

> >> Even if the choice of their preferred way to send bugs is something
> >> good for themselves (often) but bad for reporters (email).
>
> There's no evidence that email is bad for *reporters*. In fact, it's
> easier: there's no need to check for duplicates. You fire off an email
> which says what the steps to reproduce are, what you expected to
> happen, what happened and how reproducible it is; along with
> environmental information.

But maybe 10 or 100 people fire off an email. At some point more effort
is spent into frustrating reports by many people than in developing the
software. That's not bad for this application or package itself but
that's bad for the community which looses people willing to test and
willing to provide well made bug reports.

> However, users using the Extras repository (which is, as you say,
> enabled out-of-the-box) are exposed to it through HAM: this makes no
> exposure of the bug tracker field. Only if a user goes to
> maemo.org/downloads/ will they see the "Report a bug against this
> application".
>
> At that point, it reflects badly on the developer if they never
> respond to a bug report.

Maybe. But at least people can see if it's worth the effort to spend
time in reporting bugs or suggest enhancements. It may be easy to fire a
bug mail but it takes a lot of time to reproduce errors, to get log
files or debug stuff and to enter a report in a foreign language. If you
do this in more than one application it's easy to calculate that it
costs time. Time which could be saved if a tracker is available. Time to
get the testing queue "clean".

> The QA criteria are NOT designed to make every single package on
> maemo.org excellent.

That's reasonable. But the process itself contains a _qa process_ which
is one of the _requirements_ to bring a package into extras. If there
already is a qa-process the way to handle this process should be fair.
And this means better bug reporting possibilities as more complex an
application gets.

> > I'm able to understand that people don't have time but that's
> > why I also would like to see that developers understand that _I_
> > don't have unlimited time too.
>
> Then don't spend ages reporting bugs in applications which have only
> email-based mechanisms. Or send a quick mail first and ask if the
> developer needs more info.

That is what I will do. But why not prevent packages with email from
hitting the test queue? Other testers should spend ages?


> > As solution for this problem I would love to see a whitelist of projects
> > which allow email adresses:
>
> I am strongly against this, for the reasons outlined above. We
> shouldn't be putting up *unnecessary* barriers to someone getting
> their application into Extras, and this *is* one of those, IMHO.

As far as I see this discussion is a debate "bring fast and much
application packages into testing" vs. "encourage people to spend time
with testing software and give feedback before they hit extras".

Thanks, Andrew, for your opinion. It's sometimes hard to see problems
from an other perspective but even in this case there is no good or bad,
no email allowed or not allowed. IMHO it should be a compromise. Maybe
there are other possibilities to solve this problem?

More voices are welcome! :)

Best regards Uwe

[1] Examples for discussions in the testingsquad-list:
https://garage.maemo.org/pipermail/testingsquad-list/2010-July/000089.html
https://garage.maemo.org/pipermail/testingsquad-list/2010-August/000128.html
https://garage.maemo.org/pipermail/testingsquad-list/2010-April/000080.html
https://garage.maemo.org/pipermail/testingsquad-list/2010-March/000041.html

[2] http://wiki.maemo.org/Talk:Bugtracker

--
http://internettabletblog.de

Am Donnerstag, den 26.08.2010, 08:13 +0200 schrieb Thomas Waelti:
> > On 25/08/10 18:29, Andrew Flegg wrote:
> >> I am strongly against this, for the reasons outlined above. We
> >> shouldn't be putting up *unnecessary* barriers to someone getting
> >> their application into Extras, and this *is* one of those, IMHO.
>
> Agreed (Saves me lot of typing and arguing :-)
> E-Mail is enough. A channel there must be, but this suffices.

Thanks for your point of view. I see the points from the developers
perspective but in the role of somebody willing to spend time to help
bringing applications to extras I have an other view. :)

> Best regards
> -Tom (1.5% of all Maemo downloads, no real bugtracker :-)

Congratulations. :) And which place in packagetesting[1]?

Best regards Uwe

[1] http://maemo.org/profile/list/category/packagetesting/

--
http://internettabletblog.de

On Wed, Aug 25, 2010 at 8:29 PM, Andrew Flegg <andrew@bleb.org> wrote:

> I am strongly against this, for the reasons outlined above. We
> shouldn't be putting up *unnecessary* barriers to someone getting
> their application into Extras, and this *is* one of those, IMHO.
>

I already wrote about the topic on the testing list, stating my ambivalent
feelings
about the topic. After some pondering, though, I must say I don't see *ANY*
point in having an email address in the bugtracker field. Whatsoever. If we
say email is good enough, we already have a maintainer field which is an
email address, the bugtracker is IN ADDITION to this. So as far as I'm
concerned, emails in bugtracker fields should die. Thus for me the question
is
whether we should require a proper web-based bugtracker or not. If not, that
field should be optional. If yes, then everybody should have it.

On Thu, Aug 26, 2010 at 18:05, Attila Csipa <maemo@csipa.in.rs> wrote:
>
> After some pondering, though, I must say I don't see *ANY*
> point in having an email address in the bugtracker field. Whatsoever. If we
> say email is good enough, we already have a maintainer field which is an
> email address, the bugtracker is IN ADDITION to this. So as far as I'm
> concerned, emails in bugtracker fields should die.

Although one *could* have a different email address for bugs rather
than general support.</devil's-advocate>

> Thus for me the question is whether we should require a proper
> web-based bugtracker or not. If not, that field should be optional.
> If yes, then everybody should have it.

That field *used* to be optional, because if it wasn't present that
implied the developer wanted email; but ISTR a discussion which
concluded that that wasn't explicit enough, and so XSBC-Bugtracker was
made mandatory.

Personally, given the issues we've had with the QA process, I don't
think strengthening any rules which, at best, have tangential impact
on users is desirable. And I'd like to see concrete examples of where
testers have had real problems with "complex" apps like MyContacts
using email.

If there are some, and developers don't want to have the "hassle" of a
big web based application, a blank "Bugtracker" should equate to
http://maemo.org/packages/view/<package>(/<version>). That gets the
testers the visibility of previously mentioned issues, which seems to
be the main problem.

Cheers,

Andrew

--
Andrew Flegg -- mailto:andrew@bleb.org  |  http://www.bleb.org/
Maemo Community Council chair

On Thu, Aug 26, 2010 at 8:13 PM, Andrew Flegg <andrew@bleb.org> wrote:

> Although one *could* have a different email address for bugs rather
> than general support.</devil's-advocate>
>

Yes, well, nitpicks :) The only problem with the field is that some people
tend to forget to modify the maintainer field, leaving upstream addresses.
This is also an error so we might as well solve two problems in one go.


> If there are some, and developers don't want to have the "hassle" of a
> big web based application, a blank "Bugtracker" should equate to
> http://maemo.org/packages/view/<package>(/<version>). That gets the
> testers the visibility of previously mentioned issues, which seems to
> be the main problem.
>

I could go with that. That's no worse than a bugtracker put in just to
pass QA, so functionally, we're not losing anything.

Andrew wrote:
> If there are some, and developers don't want to have the "hassle" of a
> big web based application, a blank "Bugtracker" should equate to
> http://maemo.org/packages/view/<package>(/<version>). That gets the
> testers the visibility of previously mentioned issues, which seems to
> be the main problem.

I could perfectly live with this interface as bug tracker (it also would not be a good choice for complex projects but we would have visibility at least). Unfortunately there is a major bug with this interface: Commenters don't get an email if there is a new response:

https://bugs.maemo.org/show_bug.cgi?id=6914

Best regards Uwe

On Thu, Aug 26, 2010 at 9:12 PM, Uwe Kaminski <jukey@ju-key.de> wrote:

>
> I could perfectly live with this interface as bug tracker (it also would
> not be a good choice for complex projects but we would have visibility at
> least). Unfortunately there is a major bug with this interface: Commenters
> don't get an email if there is a new response:
>
> https://bugs.maemo.org/show_bug.cgi?id=6914
>
>
Is this still broken for you ? I have just seen Niels' comment, and I do
actually get mails on the packages I commented on (like pygtkeditor). The
mails are coming from garage at maemo.org, are we sure it's not some
validation/spam filter playing a prank on people ?