[yadifa-users] My reasons for giving up on YADIFA
anandb at ripe.net
Thu Apr 13 12:22:50 CEST 2017
I've just seen the previous thread from Nis Wechselberg, wherein he says
he's considering moving to yadifa, and why things are so quiet.
I used to be on this mailing list, but left some months ago when I
decided to give up on yadifa. I've just rejoined it, because I'd like to
answer Nis's question, and also to respond to the yadifa R&D team.
First of all, let me start by saying that remaining anonymous doesn't
help at all. People tend to connect better when they know who's on the
other end of a conversation. I have seen several past messages on this
mailing list that are sent from the "R&D team". I have not seen any
other software development project trying to maintain this kind of
anonymity. I don't know why the yadifa team is doing this, but it
doesn't feel right. Please start by posting your messages from your own
accounts, and sign them with your name. Let folk see and feel that there
are people behind this project.
Next up, communication about new releases has been rather poor. I have
seen the odd tweet here and there, but the mailing list has been
completely quiet. I know people are fond of twitter these days, and like
to tweet everything, but this is, IMHO, not a substitute for making
announcements to the mailing list. Tweets can be easily missed. It's
like any other social media. You see the newest stuff when you check it,
but older stuff is too far back in the history. No-one has time to sit
and scroll through old tweets. Tweets are by nature ephemeral. I prefer
email, which I can look at my leisure, and filter and organise.
Start a publicly available bug tracking system, where users can report
issues that everyone can see and join in the discussion and provide
feedback. In the absence of that, we're reduced to posting bug reports
by email on the mailing list, or to a private address ([email protected] or
something like that). This is not transparent at all. It's almost as if
the yadifa team doesn't want yadifa's bugs exposed in any way. All
software has bugs, and there's no shame in having them exposed.
And when I report a bug by email, I have no idea whether it will be
fixed or not, or in which release. There's very little engagement about
it with the developers. There's no opportunity for other users to
comment on it, and provide feedback or ideas. It's all done behind
closed doors, and then a fix or feature is released as a fait accompli.
And I've found that in some cases, I didn't like the fix or feature, but
it was already done, so it wouldn't be adjusted or retracted.
In terms of software quality, I also found yadifa to be lacking. During
my last round of tests, it would still crash on me, because it didn't
like something or another in the contents of an AXFR or IXFR. This
scares me. It means that yadifa is still unable to handle failures
gracefully. Instead it crashes, and leaves undecipherable log entries
that I have to send to the developers. And in some cases, when they've
fixed it, all they've said is "yeah we found and fixed the bug" with no
proper explanation as to what happened.
The yadifa team should look at the NSD, Knot, PowerDNS and BIND
communities. Even though these projects have some differences (for
example, BIND tends to keep bug reports private until a fix is found,
whereas the other have completely open bug trackers), in general, they
are all open, communicative and very responsive to their users in
transparent ways, with bug trackers, active engagement with users
publicly on mailing lists rather than in private email conversations,
published roadmaps showing what features users want and are planned. It
gives me great joy to interact with the developers of each of these
software projects. I know that if I have an issue, it will be addressed
quickly and transparently, and I will have the opportunity to discuss it
with the developers, and ask other users for feedback by pointing to the
issue in the bug tracker. Or ask via the mailing list.
I have nothing personal against anyone on the yadifa team, and this is
not meant to be any kind of lash out against any individual. However, as
a project, yadifa has failed to keep my attention, despite my active
involvement in it. I've spent enough time on it, and now I have other
things to do. I will not be going back to it. If some of these aspects
of yadifa get better, then it will of course be a good thing for those
who are using it, but I have no intention of reviewing, testing or using
yadifa for now.
More information about the yadifa-users