Hey Bill,

> Do you know why his students have at least twice in the past engaged in 
> deceptive spamming to gather data?

Yes, I do, see Footnote ** of my previous mail; But to recap: It is a complex 
problem between how academia is setup, what it incentivizes, what it requires, 
what it rewards, and who does network measurement research (usually people 
without operational experience working on things technically too big for a 
single PhD). 
You can, of course, now simply tie that to Tijay, who at least starts to engage 
with the issue now, and tries to be part of a solution instead of the problem;
Then again, as I said, it is a wider problem, and there surely is a ton more 
science-y-groups that are doing similar things, but simply did not pop up on 
your radar yet.

I for one personally prefer working on fixing the underlying problem instead of 
keeping beating on a person that is actually developing awareness of the issue. 
Because academia will not go away. PhDs doing research won't. So,... well. 
Let's fix things instead.

> Frankly, there's no chance that I'll ever cooperate with that person's 
> research.

This is fine. Your choice, and also the reason I am trying to build one stable 
AS for that kind of research, so people can easily enforce never being involved 
on multiple network layers, without running the risk of, e.g., blocking IPs 
that are later used by a different entity (when using stuff like EC2, for 
example). Similarly, that's the reason why such an entity should have a board 
checking planned research that actually involves non-academics. In the end, 
well done research can be really beneficial (e.g., by figuring out changes to 
RFCs); It just MUST not cause harm, and that harm is often not apparent, 
certainly not to people usually found in IRBs, and often not even for people 
that are not actively involved in that _specific_ type of system (Say, DNSOP 
assessing Mail measurements, or MAILOP commenting on some BGP experiments, even 
though, of course, there are community overlaps).

> His tenure should be pulled. He has a pattern demonstrating ignorance of his 
> field of study.
This is not fine; At least in my personal opinion. It is essentially what I 
talked about in my previous mail when referring to rather frustrated comments 
that do not necessarily take into account that the person on the other end of 
the communication is also a, well, person.
Furthermore, it--again--does not really help in mitigating the underlying 
problem, which is a bit bigger than just one person doing research, but instead 
may even make it worse.
I would usually have this discussion off-list, but given that you posted this 
to the list as well, it might be worthwhile to talk about it somewhat openly, 
as it relates to how we want to interact with each other here.

The reason I believe that comment is not fine is that it rather directly 
attacks Tijay as a person, without considering the context of events. Your 
earlier statements in that context were even a bit stronger. In any case, this 
is certainly not in line with how I would expect a rather senior engineer to 
communicate, especially given that this message may also be read by more junior 
people directly. In the end, we are all shaping the tone in the community.

Specifically, your call for revoking tenure is relatively close to "make that 
person lose their job and prevent them from ever working in that field again". 
This is usually done when people demonstrate significant neglect of ethical 
standards and practices in their _academic field_.* The problem here, though, 
as I mentioned above are those standards and practices (or rather: Especially 
the incentive structures and requirements) in the field, which are just 
incompatible with how the Internet works (and this issue is certainly not 
restricted to mail). So, you are basically putting a lot of (kind of justified) 
frustration at academic research at large into comments towards a single person.

Furthermore, when I think about the impact of one of the measurement studies 
(crashing MXes of smaller ops running off-the-shelf frameworks), I cannot shake 
off the feeling that a reply to one of those operators posting to mailop asking 
for help with the issue (without the measurements having been tied to Tijay) 
could very well have been: "Ah, yeah. Unethical scans. But you should have 
figured that out by yourself. Maybe, if you can't debug something that simple, 
and figure out how to block it, you shouldn't be running mail-servers on the 
Internet. *shrug*" 

Which, I think, illustrates the point about tone I hinted at earlier a bit more.

Instead, I would expect clear, yet respectful, communication that takes the 
underlying problem into account and works towards a solution, making sure that 
the problem does not occur again in the future, independent from the people 
involved, while the people involved are allowed to grow. 

But that may just be me.

With best regards,
Tobias

* Note: I am not saying that all academics are always doing crap work/are 
wrong, nor that all academics always do good work/peer reviewed and published 
papers are always right; Just that the way this all is set up really encourages 
fuckups en grand going unnoticed, i.e., I claim that fuckups are farfarfarfar 
more likely when people start to do measurements involving even mildly complex 
protocols which just accumulate corner cases you usually only know about when 
having run systems (usually even at a certain scale) for some time; There also 
is a lot of good work out there, but it's not like any small mistake couldn't 
have happened there with disastrous consequences; It just didn't. Beating on 
single people then creates an environment that is not really good at 
learning/improving/mitigating issues. Also note, that I might be considered an 
academic myself (and certainly did stuff in the past I am sure as hell not 
proud of. Usually ran into kind people who explained [repeatedly] and helped me 
grow; Kind of the people I want to be like for others, regardless of whether 
they are students I work with, or colleagues in an industry setting.). Bottom 
line: Look at the system, not the people.



_______________________________________________
mailop mailing list
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to