On Bugzilla and spreading FOSS, was: Where Fedora Went Wrong...

Les hlhowell at pacbell.net
Wed May 16 19:05:40 UTC 2007


On Wed, 2007-05-16 at 09:28 +0200, M. Fioretti wrote:
> On Wed, May 16,  R. Sundaram (sundaram at fedoraproject.org) wrote:
> 
> > You can choose to not contribute anything back and only expect to
> > get everything for free but if everybody chooses to be selfish you
> > wouldn't get what you are getting. This is a ecosystem that relies
> > folks investing resources like their own time or money. If you have
> > the time to complain you have time to contribute.
> 
> I have already explained (1) why I believe the "This is a COMMUNITY,
> so you must contribute to it in some way" slogan, even when expressed
> politely and in full good faith as Rahul just did, is... let's say
> counterproductive, these days, so I won't repeat it here.
> 
> I just wanted to point out one specific issue with another statement,
> also because I don't think the solution can be implemented by single
> users or developers. Rahul also wrote:
> 
> > Bugzilla reports are very *useful* form of complaints (which are
> > more of constructive feedback to the right people) but that doesn't
> > even have a representation in threads like these usually. My problem
> > is not even that people are complaining. It is that people are
> > complaining in places where nobody who can fix the issues are
> > involved.
> 
> Complex, rigid, very formal bug tracking systems are indeed very
> useful to keep under control, with the smallest possible effort,
> development and maintenance of every complex project, not just
> software. I use such systems in my daily job and can confirm that OK,
> using them isn't the most pleasant activity in the world, but without
> them it could be much worst. In this sense, I agree 100% with Rahul's
> comments.
> 
> The only problem is that expecting that software *newbies* will ever
> understand by themselves how to correctly use anything like bugzilla
> is just foolish. This has nothing to do with FOSS or Fedora, it's just
> human nature. This chicken and egg problem happens all the time on the
> lists of OpenOffice, Mozilla, Fedora, OpenSuse...:
> 
>       Q: "I am unable to do X with this software"
> 
>       A: "OK, then just learn to use this other software, whose
>       interface is even less usable of the other one, to select which
>       of these words and sentences you had never heard before best
>       describes your problem"...
> 
> This is not an attack to Rahul. Probably no single developer or other
> kind of FOSS volunteer has the time, the need, the desire or, usually,
> the skills, to replace bugzilla and friends with something really
> usable by non programmers. Or the time to read all the generic
> complaints, gather all the information missing from them and type
> everything again in bugzilla.
> 
> So, if FOSS must stop being a niche on the desktop, if all the
> excellent people who have a lot of skills but not the one of thinking
> like a software engineer (~95% of humankind?) must contribute to
> software debugging, maybe the only way to make this happen is that all
> the big companies which want to make money on FOSS services (from IBM
> to Sun, Red Hat, Novell... ) fund *one* big serious usability study,
> with *lots* of tests on real people, and pay one team of full time
> developers to replace the current crop of FOSS bug tracking systems
> according to the results of that study.
> 
> I say "one study and one new system" because this is one case where
> choice is NOT good. If you put a *decent* bugzilla substitute on
> openoffice.org, a newbie who learns how to use it there will be much
> more likely to report a Fedora, Ubuntu or Suse installation problem
> (="to be involved, that is to become and *remain* one more happy FOSS
> user") if he or she finds the very same interface on those other
> websites.
> 
>    Marco
> 
> (1) it is the seventh of the "Things we're tired of hearing from
>     software hackers", http://digifreedom.net/node/56
> 
> -- 
> Help *everybody* to love Free Standards and Free Software:
> http://digifreedom.net
> 
Wonderful links, Marco.
    I am at heart an applications programmer, I took a BSIT to learn
more about Networking.  I have experience running an older (pre Solaris
8) Unix system on Sun systems, and I have programmed on 9 or 10
operating systems with a variety of languages.  Learning about
installing and running FC6 is just a continuing part of my education.
My real interests are in AI, Robotics, real time programming, VR, and
parallel computing.  Having to learn and master networking, system
admin, and the myriad tasks of system admin, simply interfere with my
true goals.  MS failed me because I have been innundated with viri,
costing me files, time, effort and being just too much of a nuisance to
bother with any longer.  

    What I want is a secure (read impenetrable to viri and bad guys)
system, stable for development work, with some tools.  I only need a
limited set of tools, because I can create my own as I go, which is
generally what I have had to do since I have been quite out on the
"edge" for some time.  What I don't want is to have to fiddle with a
bunch of books, on line tutorials, installation guides, and other stuff
to just get a fundamental system with adequate graphics and sufficient
horsepower up and running.  I need a good, simple, effective backup
system, one where I point it to a destination, and say "backup my
system" once a week or so, and "backup my users" once a day or so, and
"backup this folder" on demand.  The backup should be read only to any
system except that backup software, and the backup itself should be
"self extracting" to the fundamental system.  Thus if the system does
the unthinkable and goes to its knees, I would only need to get a boot
cd up, point to my "ultimate backup" and say something like "run
backup>new disk".  It should not be tied to a serial number, or have a
limited number times to run.  It should be capable of password
protection and encryption for secure users, and it should not take any
longer or more information to setup than I have mentioned here.  The
password could be installed in the base software at installation time
and could include a hint to recover it.  Super secure systems would not
have the hint capability.

    This backup should be part of the default system installation, but
could be turned off in favor of some other backup for those that are
"uncomfortable" with a single solution.

    A bug reporting system should have the builtin capability to get the
required file data, the log data, and the required application data
(name, revision and date) along with whatever description that the user
could provide about what action was happening on screen when the crash
occurred.  This should require almost no interaction.  The routing
should be part of the application name for example:
            Firefox 
            version 1.0.10.00
            date: 23 May 2007
            contact: bugreports at .firefox.org
    or some such.  This would provide the crash software with a location
to deposit the report.
    The report could include such data as PS, log, os version date or
whatever the community (oooh that dreaded word) decides is required.
    The originator would have the choice to send all the data or only
some of it, in case it would be a problem for their security.  This
could even be setup by an 
    admin via a ".conf" file.  If the originator wanted an automated
reply would send him/her an email with the bug report number.

    Most users would not object to this, and with the prompt to allow
selectively sending or not, their wishes could be attended.  If the
system failed so badly that the report could not be sent, the report
could be placed on the desktop with a title "send report"  When the
system was restored, they could simply click the desktop icon, and it
would ask "send or delete" and away you go.

    This is only my personal view of the process, and it is by no means
comprehensive or implementable, and may or may not be useful. YMMV.

Regards,
Les H




More information about the fedora-list mailing list