Subject: To hit some of the points I have an opinion on:
Author:
Posted on: 2013-04-13 18:43:00 UTC
First off - the important thing to remember here is that this is a prototype. We can change literally everything about it, if we want.
Probably the biggest feature we've added is editing - this is meant to be how one fixes a typo, broken link, etc. In fact, we'd considered making posts lock after a given time after posting. Edits should be in-place, without affecting ordering, newness, etc. The intent here is that if you have new ideas to bring up, they should be in a new post. The "edited" tag is there so people can't edit away inflamatory posts and then cry "OMG you're so mean!" (Hopefully, this won't be a concern - but we decided that safety was better than sorriness on this one.) Personally, I don't think that the edited flag should appear at all on the main views - it should only show up with the posts themselves. I'd be interested in hearing what you think?
One of the big themes that seems to come up in your post is how we define "newest" - which, fortunately, is completely up to us! You do raise a good point, that we've seen that bumping on the Classic Board causes trouble, while not bumping on a Modern Board is a pretty strong contradiction of what people expect. And you're very right that we want people to be interacting with essentially the same set of threads, regardless of view.
There may be a reachable middle ground here... We could have threads float - rather than a reply automatically bumping a thread to the top, it adds to the thread's buoyancy. As threads age, they become sinky, with their bouyancy counting for less and less. (This could be a very aggressive sinking policy - after, say, a week, they could be back in their time-ordered position.) It would be a change for both views, but it keeps interesting threads from falling off the front page for a while, while also enforcing a death time.
And on individual comments:
On flagging new posts: Unfortunately, if someone visits us from an IP that they've never visited from before, on a PC they've never used before, and doesn't log in, we have no way of knowing that they're them. This is simply one of the fundamental limitations of the internet. What we could do is, once people have logged in, park a cookie on their box that says "I believe that this person is the Sunflower Official", and then reference that to their last known visit time - this cookies wouldn't give them any permissions, of course, but it would give us cross-computer last-visited time once we know who that computer is. We could also set it up to work in multiple ways - if you don't have any of our cookies, it highlights the last 24 hours, and if you do have a cookie, it highlights everything since your last visit instead.
I don't recall what the status of the search bar is in the prototype - when it's done, of course, it will be a full text search, and do all the expected shiny things.
The dual setup does risk fun with offtopicness, but there are a few mitigation strategies: Firstly, tags on the Classic Board are forums on the Modern Board, so as long as people tag things appropriately, they should show up correctly. Secondly, and this one is a bit more hand-wavey at the moment, replies can be tagged as well. So, if a newbie thread turns into a badfic discussion, a badfic discussion post can be tagged with "badfic discussion", and the entire thread will also appear in the Badfic Discussion forum. (With appropriate tags to make it obvious that it exists in both places, user configurable, subject to design review and usability study, etc, etc.)
Note visibility is trivial to add - every main Board page should have the same header with constitution, wiki, etc, IMO. It's all of a single line of code to add the header, so we can change our minds on it for free.
I'm not sure that slow subfora is a huge problem - as long as they're all visible from the main Modern Board page, we have a single place that shows where all the new posts are, so new posts aren't going to get lost in the shuffle.
One of our thoughts with logging-in is that a lot of people (myself included!) don't log into the Board to post. We would like to keep this ability - set it up so that users can decide if they want non-logs to be able to post as them or not.
Regarding spambots, it'll be a lot easier to handle the content of spam than the sources, I suspect. Given that we have members with multiple computers, using proxies, not logging in, etc, detecting content is a lot easier than detecting a deliberately stealthy user. We definitely plan to create repeat-thread detection, which will happily snark at anyone trying to ypur the board.
*Dons the Security Fedora to address the point on HTML*
Allowed HTML tags will be whitelisted, not blacklisted - we will allow only tags that we know to be inoffensive, and disallow (probably by stripping them out) the rest. This is the only way to really protect against cross-site scripting: there are thousands of tricky ways to get javascript to run on a page, it's impossible to blacklist them all.
This is going to be a large set of tags - text formatting, links (minus certain vulnerable attributes like onClick), images, tables, etc, are all totally legit.
I suspect that we would want to switch over to BBCode-esque tags, rather than raw HTML - it's much harder to escape the post section with bbcode that we translate, on the backend, to HTML.
As far as mod powers are concerned, that's not really a question of what the prototype supports as much as it's a question of how much power we, the community, want to give our mods. Everything you've listed is totally doable.
Hosting is a very good question. You're right that we don't want it predicated on a single member, but at the same time, we will want someone relatively technically savvy in a Nameless Admin position paying the bills, keeping an eye on the database, etc. Sadly, this will almost definitely not be free. So, it's a big question. I am totally against a mandatory community fee, and really have no good ideas beyond that. Hosting is probably not going to be expensive - I'd be surprised if it was more than $20 a month.
Tags will have to be chose-from-a-set to keep the forums under control - we could possibly have two sets of tags, a canonical set that map to subfora and an enter-your-own set that don't? Even that seems complicated. I dunno.