David Cary Hart wrote:
We are listing a large number of ranges that SORBS does not. There
are numerous operational differences that make both lists useful.
It's ok if the lists contain different entries. my concern is if a list
includes another one. you can setup aggregate lists if you want, but
this is not what I am talking about.
The issue is that if a list includes (most or all of) another list, then
we need to add _OVERLAP_ style rules to avoid a single "reason" getting
a too high score. if on the other hands lists are kept independent, then
finding an IP on two (or more) is meaningful.
The other issue is the one I cited: the unoisting process gets more
complicated. which isn't good.
ALL DHCP lists include some static, generic hosts. Most of the white
listing that we handle is also listed by SORBS, some by NJABL. The
fact that we choose to discuss the matter is part of the quality
management process.
I know I'm nitpicking, but I find the name misleading. why not choose a
different name? (I don't like dul nor duhl either).
* Most of these are in mixed dynamic and static ranges. We are
white listing these immediately upon request and verification.
Can you detail this process please? Why isn't it automatic?
It is fully documented on our site. We prefer to provide EXTREMELY
expeditious handling (usually within a hour). Furthermore, we choose
to have a meaningful dialog to help them get unlisted.
What is automated is that every removal request generates the
necessary data (from dig to whois) to evaluate the issue.
Frankly, we have made enormous progress is a relatively short period
of time.
as long as you keep doing so, I can only congratulate you. but if one
day you don't have the time to manage this this way, then please do
something to handle that.
It's a balancing act.
I'd say random art... Thank you for your participation to the
"Balkanize The Internet" project.
Nonsense.
Don't take it bad. It was voluntarily provocative.
The endeavor is well documented and proactively managed. Do
you have anything constructive to add?
My main concern is list dependencies. a well-known example is the
various uribl overlap. sometimes, these are enough to give a huge score,
so one needs a lot of tuning and tweaking to get a more realistic score.
This is one reason I don't like list inclusions, except for things like
what xbl does: different results depending on the source.
regards,
mouss
