On 5/28/05, sascha brossmann <[EMAIL PROTECTED]> wrote: > 1) spam on a mailing list according to my experience strongly harms > ~ its perceived quality.
As has been suggested already (implied by my message to which you responded and then explicitly in another message), sending mail to a list could be restricted to subscribers only, as long as there's another mechanism, such as a web form, for non-subscribers to post. This would probably greatly reduce the level of spam. > 2) the newbie argument is IMHO far overrated: this problem has been > ~ vastly discussed in the literature for interface & interaction > ~ design (see e. g. cooper/reiman, about face 2.0 - the essentials of > ~ interaction design, indanapolis: wiley, 2003 - very recommendable!) > ~ the bottom line goes like: think of your main and most important > ~ users as intermediate (neither rookies nor wizards) and > ~ intelligent, but with limited time. this can be applied to nearly > ~ every context. > ~ => e.g. a decent and beginner-friendly _how-to-subscribe_ is more > ~ helpful in the long run as the main users won't be impacted by > ~ negative side-effects. I'm not familiar with the research, though I'd point out that if the majority of your users are in the "intermediate" category, then statistically they'll dominate any study that doesn't attempt to normalize for each population. Newbies are important because they haven't necessarily committed to the system. The harder it is for them to set up and use their computers, the less likely they are to stick with Debian, Linux, or whatever. > 3) especially beginners normally don't get on mailing lists by chance > ~ but through some web site. => if they have questions which can be > ~ answered directly on this web site there is no need to post them to > ~ a list (which in turn requires also more of their time and > ~ attention to get the answers). => the better help there is at first > ~ hand, the better their needs are taken care of. [this especially > ~ applies to the type of questions that can be easily solved by > ~ search/rtfm] Concise and well-written how-tos are indeed invaluable. Unfortunately, if they're not available then newbies have to go *somewhere*. We can try to produce better documentation, but a lot of problems arise before they can be documented or are highly transient. For instance, I've been looking for a wireless card that'll work without having to use ndiswrapper. Everything that I've found online recommends cards that are no longer on the market. For something as fast-moving as computer hardware, anything other than mailing lists (or forums) are almost doomed to failure. > 4) mailing list participants tend to be annoyed by spam and rtfm-type > ~ noise which decreases their ability and motivation to read and > ~ answer questions Everyone is annoyed by spam, except perhaps for spammers. There are a number of Debian mailing lists, and they target different audiences. On debian-laptop I expect to see very basic questions from people who are just trying to get their laptops working and perhaps find searching for the answers too frustrating. On debian-user I expect to see just about everything. If a subject header doesn't interest me, I just move on. > to draft my position on this, i favor: > > - - vastly improving the help resources on the web at debian.org (1st > ~ stage) [maybe a simple expert system in addition to plain search, > ~ definitely better navigation, ... ] Absolutely. In particular, there's almost too much documentation, at least for some things. Duplication, especially when inconsistent, helps no one. > - - adding a realtime intermediate communication & participation stage > ~ there. think (maybe moderated) irc here w/out irc client software. A big problem with IRC is that the questions and responses are generally not archived. The more times a question is asked and answered in an archival format, especially as the phrasing changes, the more likely a subsequent search for that question will find at least one of the relevant answers. Not all questions can be answered quickly, either. > - - mailing lists (3rd stage) should then primarily host in-depth > ~ discussion of ideas and issues which cannot be dealt with on the > ~ first two stages. which means: less but more substantial traffic. > ~ and *no* spam at all. and, yes, subscriber-only, as well. Again, I think an appropriate division of mailing lists works best. I happily sift through the discussions on debian-user looking for anything that I can answer. I figure that if I can save a real Debian guru the time and effort to answer a question, that's more time for that guru to tackle a more difficult problem. For more in-depth discussion, I'd look at debian-devel, debian-legal, or some other more topically specific list. I'm not going to hang out on an IRC channel and try to pick through the noise. I'm not going to check a web forum. I'm not going to spend days writing up a how-to and polishing it so that it's easily readable (at least, not often). I *will*, however, check my email. Besides, with all of the documentation out there, whether on debian.org or elsewhere, users *still* turn to mailing lists first. There's something about the email interface to getting help that people seem to *like*. I'd guess that it's: 1) asynchronous -- they don't have to log into something and sit there until their problems are solved (in contrast with IRC); 2) personal -- you can't ask a how-to for clarification, nor how to interpret some bizarre error; 3) convenient -- the response comes right to your mailbox (in contrast with forums). -- Michael A. Marsh http://www.umiacs.umd.edu/~mmarsh http://mamarsh.blogspot.com
