I am sure whoever suggested doing an upgrade install don't have a
terrible experience unless...
The point is not about whose experience and technological skills and
knowledge; the point is to maintain a consistent quality output (even
though the initial *production* time is longer) to eliminate preventable
unnecessary future rework cost - which cannot be estimated because no
one could predict until it *happened*.
The concept is very simple - do thing right the first time, and this is
how we could improve productivity, work quality, and to reduce cost,
which then will become profitability which then will become pay raises
and bonus checks.
Doing an upgrade check is the minimum, and veterans like you should know
that compatibility list does not reliable because everyone has a
different interpretation about *compatible*. Some consider being able
to launch a program is compatible while others consider compatible means
retaining all functionalists.
And it's not uncommon for some sloppy programs couldn't even perform
well during its own upgrade/updates not to mention to do a system
upgrade. Finding a problematic driver and/or program alone after
installation could be a costly job.
So, the time spent on post-installation checking and troubleshooting
could be much higher than the installation time for an upgrade install.
So one should always include the *total time* spent not just initial
time without consideration of possible re-work time.
So it's not about whose experience is better or not, and it's about work
attitude - doing things right the first time.