#353926 lists.debian.org: Please make the debian-devel-spanish list open to subscribers-only #353926
- Package:
- lists.debian.org
- Source:
- lists.debian.org
- Submitter:
- Javier Fernández-Sanguino Peña
- Date:
- 2025-02-08 08:09:02 UTC
- Severity:
- wishlist
- Tags:
Please, make the debian-devel-spanish mailing list open only to subscribers. The load of spam there, vs. posting from regular users, is very high. For example, july last year had 4 valid e-mails out of 34 postings to the list (that is, 30 where spam). See: http://lists.debian.org/debian-devel-spanish/2005/07/threads.html In February this year (although it did not end yet) there have been 4 relevant e-mails out of 12: http://lists.debian.org/debian-devel-spanish/2006/02/index.html Even though some of the postings might come from unsubscribed users, I think many developers would gladly accept becoming moderators of the mailings of unsubscribed users. I rather reject posts to the mailing list myself than manually reviewing the web archives and reporting as spam tons of messages. Thanks Javier
Hallo! Du (Javier Fernández-Sanguino Peña) hast geschrieben: we generally don't have moderated lists (beside some very few exceptions, and most of the people are happy with that. please report spam that you get through the lists. see http://lists.debian.org/debian-devel/2006/02/msg00946.html Cord
well as throught mail. To no avail, all the spam I reported is still there. See, for example, for a S/N ratio for that list: http://lists.debian.org/debian-l10n-spanish/2005/05/threads.html I have reviewed the *full* 2005 year for that list and reported spam in both that year and 2006. Same for the debian-doc mailing list. Quite sincerely, when I see those web threads full of the spam I have reported time and time again I get quite frustated. In any case, the fact that you have not set moderation in the past does not mean that it should not be done in the future. There's people in the d-l10n-spanish (and other) mailing lists willing to moderate, the lists don't have that much cross-posting (not like -devel), and is most conversations are among subscribers only. I do believe it be really easy to handle the moderation load. I also can't understand why we have all those spam reporting functions when people using them (like me) don't get attention. I don't understand why requests from long-standing people don't get honored: I've been a member of that list for 6 years and I'm currently the main coordinator for that language. I thought that would mean something, but I guess I'm wrong. Regards Javier
I got confused here and mixed debian-l10n-spanish and debian-devel-spanish. I've opened up a (different) bug to ask for subscribers-only debian-l10n-spanish too. Sorry for the confusion, Javier
I think that's a misrepresentation of the truth. The truth is that *everybody* would like to see less spam. From them, a non negligible amount of people would even be willing to moderate, if becoming a moderator would be the requirement to have less spam. So just because you have not actually asked people about spam does not mean they are "happy" with the level of spam. I am in fact terribly frustrated by the fact that we are still not doing everything we could do to stop spam more effectively.
Hallo! Du (Santiago Vila) hast geschrieben: we got 34288 mails incoming yesterday of which 2292 passed our filters. so we have a signal/noise ratio of .0668. if say 10-100 spam-mails pass our filters you should keep in mind that that is only 0.02%-0.2% of the incoming mails. I really have the opinion that you have no idea what is happening on our listserver, nor what we can do and what we can't do. Switching Lists to moderated would mean that someone would have to review between 30 and 300 Mails a day to check if they are appropriate (numbers for debian-doc yesterday). I don't think that we find enough people, who want to do that reliably (beside the problem that we don't have an easy method to handle moderation). It would also require a major change in running our mailinglists, something i don't want to do without a real discussion and maybe a GR, only because of two people (out of 90000 unique subscriber, which doesn't complain). So now think what you personally want. * keep it like it is, with 0.02%-0.2% of incoming spam passed through * moderate with 30-300 mails to look at a day so the wontfix will stay, as long as we don't do a complete redesign of lists.d.o. i won't close the bugs, but we also won't moderate those lists. Cord PS: i hope this mail was polite enough, because i'm really pissed, that people cry out loud if our filters stop a few legit mails and we do our best to not let that happening again, and then the same people again cry out loud because the spam level is too high.
Just switching lists to moderated would be a nightmare. I agree that it would not be practical. However, a lot of time ago I proposed a semi-moderation scheme. To summarize, the procedure would be as follows: * Messages having a high spamasassin score are discarded, as now. * Messages having a very low spamassassin score are allowed, as now. * It would just be the messages in between the ones to be moderated. Moreover, the moderators would not have to be you, listmasters. The moderators would be volunteers from the same lists that switch to this scheme. This procedure would prevent a lot of spam from reaching the lists, and, at the same time, would not mean a lot of work for moderators. It would not require a major change in our mailinglist, only a few procmail recipes (well, maybe more than a few ones, but definitely it would not be a major change). Also, it would not have to be done for all lists at once. We could try it on a few lists, and if it works, extend the procedure slowly to other lists. I hope this mail was polite enough as well.
Cord, [ I think you were not a listmaster yet in early 2003 ]. Please take a look at Bug #175744, where I explain the semi-moderation scheme and provide figures showing why I think it would improve things considerably without a lot of work. Thanks.
While we are at it, please note that the real signal/noise ratio in the lists (as perceived by the subscribers) may not be measured by mechanical means[*]. For a subscriber, it's not the spam that the filters stop what counts, but the spam that they do not stop compared to the legitimate messages in the list. If a user complains that one out of two messages is spam in a given list, don't tell him "sure, but the amount of spam we stop is much higher". [*] The only way to measure this accurately is by counting false negatives. Only people who actually read the list can do this. No procmail logs in the list server will tell you how much spam that was allowed should not have been. This is in no way to tell you that your work fine-tuning filters is not appreciated. It is, but please don't measure their effectiveness just by looking at the logs, as they will never tell you about the false negatives. Thanks.
Hallo! Du (Santiago Vila) hast geschrieben:
[ I take your three latest mail into this response. ]
certainly.
i said that the amount of daily mails that would go into moderation is
too much to handle for everyone.
I try to get the data for yesterday. we had 30123 spamassassin
classifications.
we currently cut at score 4.0
the distribution of scores (i int() the scores)
2 -9
2 -8
18 -7
50 -6
83 -5
328 -4
228 -3
78 -2
154 -1
132 -0
233 0
221 1
25 2
18 3
100 4
80 5
assumed that scores between 2 and 4 go into moderation there is the
really small number of 43 mails for _all_ lists that day that went to
moderations.
so even if we would extend the range for checking for 0-5 there would
be 597 mails for 192 mailinglists. (i think it would be a good idea to
have different 'grey'-ranges for subscribers and non-subscribers.)
so. your idea doesn't produce the amount of mails that i was
expecting.
it is do-able.
now come the problems... we already have a lot of procmail-rules where
nobody knows what they are good for.
That whole system we currently use as major communication for the
project is messed up beyond repair. thats the main reason why i set it
as 'wontfix'.
We could set smartlist to have a moderated-list, but that will
moderate everything (that wasn't drop by our spam-filters), or we can
set it up as unmoderated and add those moderation-rules somehow, and
nobody knows if that will work.
right.
done.
correct. for that we currently esatblished the reporting-mailadress.
Cord
Well, I never reached Joey Schulze's level of expertise writing procmail recipes, but I maintain the Debian packages for both procmail and smartlist, so in theory I'm supposed to understand procmail recipes. Would it help if I try to document the current set of recipes at lists.debian.org? I would only need read-only access for that. If understanding the current code is the real problem, I'm definitely available for help. Once (or better: if) we understand the current code, we could try new things with a newly created test list. If those new things work, we could then try with a real list, like the ones proposed by Javier.
Hi,
Nearly 10 years later... I'm just sitting at DebConf15 and check
through open bugs for lists.debian.org
I wonder if we should just close these bugs or what to do with it.
Do you have an opinion?
Yours,
Cord, Debian Listmaster of the day
Hello, I am contacting you with regards to using your name for funds claim of long overdue dormant funds for Investment belonging to a late depositor. Let me know if you are interested for more details. Best Regards,
Hello, I am contacting you with regards to using your name for funds claim of long overdue dormant funds for Investment belonging to a late depositor. Let me know if you are interested for more details. Best Regards,
Final Notice. You are among the beneficiaries of 2024/2025 grant for all scam victims and relatives reconfirm your email if active for more details Thank You. Regards Mr. Rowland Cole ( Financial Crimes Enforcement Network)
Final Notice. You are among the beneficiaries of 2024/2025 grant for all scam victims and relatives reconfirm your email if active for more details Thank You. Regards Mr. Rowland Cole ( Financial Crimes Enforcement Network)
Final Notice. You are among the beneficiaries of 2024/2025 grant for all scam victims and relatives reconfirm your email if active for more details Thank You. Regards Mr. Rowland Cole ( Financial Crimes Enforcement Network)