Hi Clement, > On 31 May 2018, at 14:56, Clément Aubin <aubincl...@gmail.com> wrote:
[snip] >> >> All others: if you have any recommendation or counter argument, please post >> it quickly :) > > I don't have a good knowledge of the "old" message stream > implementation, however, I'm concerned about the ability of the > notification system to act as a messaging center. > > - When working on multiple documents, I might end up talking with > multiple people at the same time. What would happen to the notification > center ? Would I have a composite event with all of my conversations in > it, or one composite notification per user I'm talking with, which would > probably fill most of the notification center ? This is not supported ATM (see http://extensions.xwiki.org/xwiki/bin/view/Extension/Message%20Stream%20Application/) so no issue FTM. You can only send messages to a given user, to a group or to everyone. I don’t see a problem to add this feature later on if we need it. > - When being in a wiki, I might end up staying 1, 2 hours or more on the > same page, either to edit it or to read it and refer to it from time to > time. The messaging feature of the notification is interesting here if > and only if we have the ability to auto-refresh the notification center > every X seconds / minutes to check for new events or, in this case, new > messages. AFAIK, this feature isn't in place for now. I don’t see this related at all to messaging. You have the same problem for any kind of notifications. This is related to live notifications and something generic for the notifications feature. I don’t agree with "The messaging feature of the notification is interesting here if and only if we have the ability to auto-refresh the notification center every X seconds”. Again I don’t see why this is related only to messaging and also I find it more useful to have messaging than nothing (which is what you propose). I can imagine plenty of use cases where it’s still useful to have messaging without the auto refresh. I consider auto refresh to be a nice improvement to have not as a "must have or there’s no value”. > All in all, I'm very concerned that we try to ship a half finished > feature that cannot be used for real collaboration and that still takes > some place in the UI ; Maybe there’s a misunderstanding: We’re not trying to develop a messaging feature. This feature already exists and we’re not touching this. All that we’re doing is add the ability to display messages in the notification UI if some messages are present in the Event Stream table. This feature already exists today and not doing this would be to remove a feature that is now interesting to have thanks to the replacement of the AS by the Notifications UI. > only to have an equivalent of a feature that we > disabled some years ago. We disabled it for 2 reasons and not for the reasons you mentioned above: * Because the dashboard was on the home page and several users were not using the message sending UI and didn’t want it to take valuable space on the home page. This has been solved by moving the dashboard to a different space * Because you couldn’t know when someone was sending a message to you and it was hard to check your personal AS (you had to nav to your user profile). Those 2 problems are fixed so there’s no reason to not have it anymore. > If we want to promote messaging and more user > to user interaction in the wiki, maybe we should take more time to spec > it. If we had to vote for this, I'd say -0 for now. Yes and that’s why Caty is going to conduct a more general investigation on that. I understand your POV and I don’t disagree with it in general. However, I feel that: * I don’t like regressing/loosing features * This is an opportunity to have something now. I can tell you that if we do nothing now, nothing will happen for at least 1 year * We’re not redoing the messaging feature, just the display part in the notif UI * I don’t see how there would much changes even if we spend 2 years designing a new messaging center. * It’s not going to cost much (probably 1-2 days of work max). So at worse ,it’ll cost us 1-2 days if we had to redo everything. * Also note that we're not changing the store or anything so even if we redo it completely differently it doesn’t matter and we will not loose anything * This will allow us to gather feedback from the community about needs/improvements * We can never implement a feature in one go. We need to work iteratively. We just need to make sure the architecture is not going to break. Since we’re not touching this part, there’s no real risk So I really believe that it’s more positive than negative to spend this little time to display messages in the notif UI. WDYT? Thanks -Vincent > >> Thanks for you time and have a great day, >> >> Guillaume