[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: spam (Re: Bug#202373: traitor offstage)

* Don Armstrong

> On Fri, 20 Jul 2007, Oleg Verych wrote:
>> When i looking trough say 1024 messages in bug-dist, all i have are
>> lines with author, subject, data. The Subject have bug number and
>> what was set in it by author. Sometimes wordings there are not
>> attractive without context. And the big one will be the package name
>> there. E.g. i don't care for "Subject: Bug#OMEGA patch" for ncurses,
>> but care for SLang, etc.
> If you don't care about specific packages,

Well, i'm not a machine. In some cases: interesting subject with
(sometimes) interesting/useful package can bring me to the message.
Otherwise i will concentrate only on stuff, that matter to me. But
again it depends.

> then you don't want to read
> -dist;

Why not if currently i can?

> you want to subscribe to the PTS for those specific packages

What i'm talking here and in "headers" thread is: i don't want to
subscribe to many, many things i easily can lost in.

> (or otherwise filter -dist appropriately.)

Because i don't want to bounce messages in case of any problems. Also
i don't want to generate useless traffic. I've found Gmane and it
satisfies me:

* i don't have huge incoming traffic at all;
* i download and read message bodies *only* if i find subject interesting
* all i use is new reader (not complex setups with maildirs or MTA
  transports or whatever)
* i handle *huge* amount of information now, including study archives,
  without distraction from up to date state.

>> Again, this is optimization thing, not necessarily suitable for
>> skilled developers with bts tool and by-package MLs with less
>> traffic. But i've found more efficient to check one place for bugs
>> reports. Other on-topic MLs are for discussions like this one.
> I'm still not certain what your use case is;

Thing is: *following* ever changing information. This is fuzzy thing
for me ATM.

> if you're looking for information on a specific package,
> bugs.debian.org is the place to look.

If i can easily follow current state of affairs for some time, then
sure i can look for all bugs and other things. But prior to that i
have nothing to do there.

> If you want to know about a certain error string,
> bugs.debian.org/cgi-bin/search.cgi is far more accurate than google or
> gmane.
> If you just want to see the bugs scroll by, then that's about what
> -dist is useful for... but then you have to filter it out yourself if
> you want it filtered.

All i'm asking pkg name in Subject next to bug number and one or two
message-ids in References. Thus everybody in that _massive_ lists can
follow information easily *without* big overhead. Because

>> Also reducing size of header "X-Debian...: pkgname" to just pkgname
>> in Subject is a little size optimization.
> Size optimization isn't the point; the point is utility. It's much
> easier to filter based on discrete headers than Subject:, especially
> when Subject: foo pkg breaks bar pkg makes things even more useless.

without downloading body (this is feature of mail<->news) i can filter
out things:

Subject: Bug#OMEGA, ncurses: patch......"

i don't care

Subject: Bug#KAPPA, libslang2: patch......"

i do care.

>> One more thing about size, if you permit. Why empty headers are
>> added? It's a waste IMHO.
> What empty headers?

|X-Debian-PR-Message: transcript

Like the first.

Reply to: