On Mar 14, 2007, at 8:22 AM, Vincent Massol wrote: BTW thinking about this it's true that we take the result to jira or the wiki but there's one category that we don't summarize: things that haven't lead anywhere. For example, take this forum idea which we are debating. Imagine that we don't go forward with it. I think it would still be useful to keep a trace of that on xwiki.org. Something like "Things we've discussed but decided not to go forward with". So when someone asks or raises again the idea we could point them to that place. It would also serve as documentation so that any new comer will not be tempted to raise again some topic (unless he's read the old threads and decided he's something new to bring to the table). I believe this is what is advocated in http://video.google.nl/ videoplay?docid=-4216011961522818645 (How Open Source Projects Survive Poisonous People (And You Can Too). It's not easy to do and requires conscious thoughts but it's an idea to get the maximum value out of any discussion. We spend energy and time discussing something so it would be good that this energy be useful to a maximum extent. This could make it a little bit more useful. Thanks -Vincent

I know about Nabble, yet I don't think it is as powerful as a forum. One thing about mailing lists is that you need to subscribe to post a message and receive the answers. I know that you can unsubscribe at any moment, but you still get a lot of other messages along with what you're interested in. Another thing is that if we host a forum, we know how long it will be there. If we use mail, it will be in our account, and for the moment on some servers, like Nabble. But we don't know for sure how long it will be there, and how long it will display the xwiki mailing list. Again, personally I prefer mails, but we should listen to the users' voice. On 3/14/07, hritcu <catalin.hritcu(a)gmail.com> wrote: -- http://purl.org/net/sergiu

On Mar 14, 2007, at 9:09 AM, Sergiu Dumitriu wrote: As I said in my previous email you can do the same with Nabble but I don't like it (see http://www.xwiki.org/xwiki/bin/view/Community/ MailingLists at the bottom). I think it's a small price to pay for getting a good free software ;) But again it's possible and it's being used. I don't quite agree. What if ObjectWeb goes away? We won't have any SVN anymore? What if our hosting company goes away? We won't have any farm anymore or any jira, etc. The best we can do is duplication. Today our lists are backed up on at least 3 locations: - OW - Nabble - Gmane I think we're pretty safe :-) I haven't heard any complaint of users about this (especially about users knowing about Nabble). Can you point me to someone who's complained and to whom we've said they could use Nabble and then they said that it wasn't good enough? :) BTW one person isn't enough; we need several before deciding that it's something we need to address. BTW you may not have noticed but a lot of our users are already posting from Nabble (you can notice as Nabble inserts some text at the bottom of emails). Thanks -Vincent

Hi Vincent, On 3/14/07, Connected Performance <Uwe(a)connectedperformance.com> wrote: This is actually a good point we should probably consider. We hardly have any categorization with the mailing lists other then the split between users and developers. Both are high traffic, and while you, or Sergiu, are probably reading everything, you cannot expect your users to do the same. So don't you think that better categorization would help? Would it be possible to have it with the current infrastructure (e.g. we could for example tag posts, like we do for [Proposal] etc.) ? WDYT? Catalin

Vincent et.al., Here are my questions after having read the responses so far: - I didn't know about Nabble, I like it much better, it allows me to focus on the subjects I am interested in. But it only takes me from a flat list of EMAILS to a flat list of SUBJECTS. As soon as the community grows and the # of subjects with it, you have the same problem as before, but lets wait until more people suffer from information overload and hopefully we find out before they unsubscribe. - I realized that different people have different ways of communication and preferences in how to categorize topics (I prefer categories, you prefer a good searching machine). Let me state the difficulties I encountered, when I started using XWiki, maybe that sheds some light on what I am looking for: First of all I am using XWiki on your farm. I initially thought the xwiki-'users' in the mailing list are of the same kind until I realized that the overwhelming majority of posts where not about how to use xwiki, it was about how to make it work properly so others can use it. I wasn't looking for solutions of how to upgrade from beta2 to 3 without problems, I was looking for 'how to use Xwiki'. I found the FAQ, but some links in there didn't work and some entries seemed to be outdated, since XWiki didn't behave as described. So I did some trial and error stuff myself, some successfully some not. So I am looking for recommendation, so I can work effectively: - Become a SW expert in how the guts of the XWIki work? - Find a XWIKI expert and pay him/her? - Stay on the users list and ignore all the developer topics there and hope enough people have similar problems? - Wait some more months until more more users like me are part of the mailing list? thanks in advance UWe vmassol wrote: -- View this message in context: http://www.nabble.com/Do-we-need-a-forum--tf3399333.html#a9478008 Sent from the XWiki- Users mailing list archive at Nabble.com.

Vincent, Your points are understood. My intention though was to try XWIKI BEFORE I make a business decision. My assumption was also that the community might be interested in users like me, who help improve usability by raising and help solving issues. According to your statements this seems to be reserved for technology savvy developers. O.K. I live with what I've got. Thanks Uwe vmassol wrote: -- View this message in context: http://www.nabble.com/Do-we-need-a-forum--tf3399333.html#a9483352 Sent from the XWiki- Users mailing list archive at Nabble.com.

I haven't yet had a great reason to complain, but I can see the need emerging, as the community grows, for more narrowly targeted discussions. I have certainly bulk-marked threads on this list, and like Catalin I believed that forums offered the best opportunity for that (our XWiki host also hosts Roller and jForum), but given that tag-and-search offers potentially better targetting, it may be the better way. One caveat on tagging, though: the effectiveness of such a scheme depends on user behavior, much more so when the community is relatively small (which XWiki's certainly is, compared to, say, del.icio.us). The genius of it lies in the fact that each user benefits most individually by the same behavior that benefits the community as a whole, which is tagging items meaningfully. Nevertheless, such an ad-hoc mechanism can be unsatisfactory if too few people "get it". Like hashing, it walks a fine line between high collision rates (too many matches) and excessive sparseness (too few). A more robust way of combining tags ('and'/'or'/'not') would benefit tagging almost as much as with search terms, although because tags are cheap and can be liberally applied, the current "require all" works pretty well where there is a large population of liberal taggers. I do see, as the product and the community grow and mature, a growing distinction (witnessed recently) between "very advanced users/administrators/extenders" and the more casual user, and even finer distinctions within each of those groups. The difference is both qualitative and quantitative: obviously, for a successful project the growth of the population of casual users will quickly outpace and overwhelm that of advanced users, and self-help facilities will quickly be critical to success. The advanced users/admins/extenders will have more questions needing individual attention, and more likely to end up in JIRA. One last thought: generally speaking, user fora and other self-help mechanisms do tend to grow of their own accord, so it's not necessarily a matter of major concern - if there's a big enough perceived need, someone will host it, whether anyone else wants them to or not, and the quality of the idea and the implementation will tend to attract those whose benefit it best serves. Under that heading, I intend to investigate Nabble; it does sound interesting; sort of like subscribing to the mailing list with a shared GMail account, if I understand it correctly, no...? brain[sic]

Me too :> And there's the rub, of course. The variety of topics that arise is as varied as the human imagination; as such it defies categorization, especially a priori but even after long experience, because of the variety of personal perspectives and because as a community matures it changes. Any attempt to sort it all into neat categories fails for several reasons: overlap in categories; difference of perspective; ambiguity of labeling; probably many more that I can't think of. That's why, like you, I favor Google's philosophy that searching is better than filing - when the tools are available to do it effectively. Here you make a very good point. I'm not certain whether this is included in what you mean by 'cross pollination' but I can certainly see a problem with a highly-segregated issue taxonomy not making the necessary connections between topic posters and monitoring listeners. A hearty Amen to that. Of course, they're not mutually exclusive, and those who find it necessary to create other mechanisms are always free to do that, and if they do, and are successful, they are very likely to contribute to the effort.