[kde] [Bug 333888] New: Quality Assurance should supersede short-term temporary and local advantages

[Uwe Dippel]

https://bugs.kde.org/show_bug.cgi?id=333888

            Bug ID: 333888
           Summary: Quality Assurance should supersede short-term
                    temporary and local advantages
    Classification: Unclassified
           Product: kde
           Version: unspecified
          Platform: unspecified
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: NOR
         Component: general
          Assignee: unassigned-bugs at kde.org
          Reporter: udippel at gmail.com

Running KDE, and being a great fan of this desktop, I perceive urgent needs for
some quality assurances not thrown overboard for 'quick fixes'. Software
quality does not improve on quick fixes!
It looks like there is too little done on a software quality assurance.
It is like in mathematics: To prove is difficult, usually; to prove the
contrary is easy: a single example destroys the validity of the hypothesis.

KDE has had a black spot for years: the desktop search. And various people and
at various locations, stated that 'with the next version of Nepomuk the problem
is solved'; and bug reports, including my own, were invalidated with 'not using
the latest version'. Where that 'latest version' did not work any better than
the one reported.

Worse is what has happened as consequence: a totally new effort was undertaken;
by a rewrite/rename. Though there was no quality control at all. Everyone just
believed that the new code by Vishesh Handa would be the silver bullet.

Serious mistakes:
1. No checks were done that it actually worked on a large scale
2. It was - despite of previous problems with Nepomuk over some years -
accepted despite its lack of a general deactivation
3. More programs were allowed to resort to this software as dependency (e.g.
Dolphin)
4. (Similar to Heartbleed) Nobody cared to look at what it does / does not: It
does not exclude network drives - a fact acknowledged by the author

Now the search engines are full of "... eats all my RAM" / "slows down my PC to
a halt" et cetera. 
In my case it is a quadcore i5 that takes almost 10 seconds to bring up an
xterm on an otherwise idle machine. Uninstall would remove Dolphin, Kmail and
bunch of other useful applications. 
Yes, I am and I have to be connected to some network drives. 

Software quality, and professional approaches have not been applied by the KDE
group, since the author acknowledges that network drives are not taken care of;
and that a bug report needs to be filed to this behalf. 
Alas, and sorry, but this is real and serious crap. How can a distinguished
project like KDE allow unchecked and obviously un-audited code (and concepts)
to simply replace other software; when even the author (who looks like a
nincompoop in this case) has simply not have the time to think of network
drives? And how did nobody else in the KDE project just overlook this problem?
And how did everyone else overlook that an untested, seemingly
written-from-scratch software enters an update with system-wide consequences
(mail and file utilities) obviously un-audited and unchecked?

Yes, bugs can happen; I myself have introduced hundreds or thousands in my own
code, for the last 30+ years. One thing has not happened, though: a blind,
auto-piloted flight with completely new software in production; hoping and
praying it would work,; and not even include a fall-back alternative in case
that it might fail.

In a nutshell: It is highest time that the KDE project develops and implements
clear software writing and commissioning guidelines, where a fresh program does
not immediately and without fall-back alternative take over relevant system
functions.
This is not a software problem, it is an organisational shortcoming of serious
relevance and implications, and I would hope and wish that the KDE-project
immediately starts looking into these  aspects; and creating unambiguous rules
on introduction of new software and new features to be based on a complete
feature control, roll-back ability and security audits.


Reproducible: Always

-- 
You are receiving this mail because:
You are the assignee for the bug.