On 4/7/2014 5:23 PM, Jeff McWhirter wrote:
Hi all,
I've been staying out of this discussion since I don't use the IDV or
the TDS anymore but John's note raises some red flags that I think
Unidata and its governance committees should address.
The rollout on friday hit a few snags, compounded by me being sick
so I couldnt respond quickly.
Considering what Unidata has gone through in the past with losing key
personnel and the impact that that has had on its products it is
disconcerting that one of its core products is still tied so closely
with an individual developer. I don't fault the developer - John is
incredibly dedicated and works very hard - rather I fault management
and its continuing inability to address these (albeit difficult) issues.
You might think "Why didnt we find these problems before
releasing?" Because we dont have a separate test team, like
commercial software release cycle does. We constantly improve our
testing, but plenty gets past us. We rely on the users to find the
last set of bugs. And its all but impossible to get users to try
beta software. Fact of life.
This is a cop-out. There are 5 engineers who work on THREDDS
(http://www.unidata.ucar.edu/committees/usercom/2014Apr/agenda.html)
and 2 engineers that work on the IDV. The TDS has been in production
for over a decade. These were *not* edge case errors. They were
errors in core functionality and should have never gotten into production.
-Jeff
Hi Jeff:
Thanks for your thoughts, and your perspective.
There were 2 defects that would normally have been caught and fixed
within a few hours, in normal circumstances. The inability of some to
tolerate a few days of intermitant downtime highlights the tension
between thredds.ucar.edu as an operational server and as a vehicle for
continuous innovation.
The TDS is complex and possibly over ambitious software for the size of
Unidata, which has about 13 software enginers and about 15 software
systems and projects. The TDS is providing access to large collections
of GRIB files in a way that no other software in the world can do. This
has occupied more of my time than I wanted, to the neglect of other
issues; in retrospect Im not sure if GRIB is worth it. But I think we
have the bulk of it done, so we'll see how it rolls out.
Meanwhile, we feel like we have some solid ideas on how to improve
things going forward. Thanks everyone for your continued support and
feedback.
John
