[reportlab-users] encoding errors

[François Pinard]

[Dinu Gherman]

>Andy Robinson:



>> Yes, if people are disciplined.  But most real-world project trackers

>> grow in length forever, and create differences of opinion.



>I tried to think more deeply about this.


Beforehand, I would like to say that I find Andy's attitude and opinion 
pretty refreshing.


>There are very successful real-world projects like GNU, Linux,

>Mozilla, Apache, Samba, Python, Perl, Ruby, Zope, Plone, and

>Django, to name just a few.  I am sure they have very long bug

>lists in their trackers and generate lots of different opinions.


I know most of these projects much more as a user than a developer, and 
undoubtedly, they are successful.  However, I do not think their success 
is a consequence of using a tracker.  One could also enumerate 
successful projects which do _not_ use trackers.

In the time when almost each and every maintainer was reachable and 
speakable, I used to report tons on bugs on a flurry of projects.  Now 
that I have to learn the toys of every maintainer, and each its own, 
often supplemented by pre-requisite study of the conventions set for 
correct filling forms, each project its conventions; now that I face the 
poor UIs and horrid editors that most trackers force us into; now that 
nice humans deprive me of the pleasure of interacting with them without 
interposed robots, this is not much enjoyable anymore to report bugs, so 
I almost stopped doing it.

The GNU Autoconf maintainer once decided for a tracker, when I routinely 
submitted a documentation corrected about a sentence which was not fully 
true.  The maintainer decided to push my letter himself into his new toy 
(this was at a time maintainers were less aggressive at forcing users 
into directly using their toys).  It took half a year, and a dozen of 
automated messages informing me of the "progress" of my report, before 
the correction made its way.  Priorly, for such things, I usually 
emailed the maintainer in the morning and received "Thanks!" at noon.

I could tell similar tales or horror stories for pages.  Let me be 
brief.  Some might even argue that my experience is atypical, or that 
I lie.  I do not want to be pulled in such arguments.  The fact is that 
after a few dozens of similar experiences with various projects, 
I really gave up on trackers.  Life being what it became for me, I do 
not have so much free time for participating into projects.  So if 
a project uses a tracker, I prefer to be a mere silent user for that 
project, and wander elsewhere.

The pattern I observed for many years is that many maintainers now put 
a high bar for bug reporting.  For users who persevere enough, a bit 
like in Pavlov experiments, they learn to love and enjoy their misery, 
and then become captive for a few projects only, instead of learning to 
be good citizens in the whole realm of free software.

The problem of sharing support is finding good collaborators.  Bug 
trackers may be a communication device between these people, but so is 
email.  Bug trackers do not create collaborators, and a project does not 
really acquire speed merely by using a tracker (unless the maintainers 
are pretty disorganised to start with, in which case the real problem is 
elsewhere).  It also happens that some maintainers use bug trackers 
because they feel unable to filter spam out of their incoming email.  
Once again here, trackers do not address the real problem.

Does GNU use a project-wide tracker now?  It surely was more successful 
when not, then :-).  Just teasing of course (yet, the truth still is 
that the GNU project once had a crisper visibility and authority).

For Python, they are unsatisfied with the tracker they used for a lot of 
years, but feel pretty stuck with it, unable to change or improve on 
that aspect.  Some time ago, they decided to do something more formal 
about it, set up a committee, called for votes, etc.  The big thing as 
usual.  I'm not closely following the Python project anymore (their 
administrative overhead is just incompatible with my desire for life), 
but I guess improving their tracker will have taken a lot of time.

For Perl and Samba, Larry and Andrew surely were very speakable.  Andrew 
never forced me into the bug tracking system he created.  While Larry is 
a true gentleman, his spirit does not spread anymore beyond the riot of 
haughty collaborators surrounding him, he is buried alive in the crowd; 
and despite I've been using Perl since version 1, this was enough for me 
to decide for a better world, and this is how I happily found Python.

I think that ReportLab would be a lesser project with a bug tracker, 
than without.  Currently, support is great, quality is there, and if 
there are administration duties, at least these are not thrown at users.  
Let's keep it that way.

Now, if internally, Andy wants to use a tracker for himself, and if this 
is to remain his own problem, he surely is grown up enough to do so! :-)
If he ever wants to publicize a tracker for ReportLab users, I really 
wish he will choose something nearly transparent, and email-compatible.

-- 
François Pinard   http://pinard.progiciels-bpi.ca