Meta Discussion / SL II

For lack of a better title, I called the page "SL II". I hope it does not raise too high expectations. Most of this has been talked about several times already. So not much news here.

• New discussion pages: real forum like. That means
  • One or more threads per page possible
  • Posts are signed automatically with author name and timestamp
  • No one can edit foreign posts, but only their own posts (except for librarians and admins)
  • Posts in a thread are listed in chronological order
  • No archiving (if you edit your own post, only the newest version is kept)
  • No minoredit (if your comment is not important enough to warrant a major edit, don't make the comment in the first place)
  • Instead posts can be filtered out
  • Advanced features as time permits: RSS feeds per page-forum, per thread, mail interface, ...
  • If you have no idea how this can look like, look at examples at http://www.phpbb.com/ or http://www.phorum.org/ (although it will be integrated in the typical SL layout)

• Different domains (or sub-wikis or areas)
  • Domains:
    • Reference: strict editing rules, wikipedia like, content: e.g. terms, people, history, etc. (read: factual information)
    • Teaching: strict editing rules, teaching material only. Think of the wiki version of Ishi's elementary go series (or other books) to get the picture
    • Playing: strict rules for content, but editing less strict: problems, BQM, ongoing game, ...
    • Homepages: just like now (maybe pages can be protected from foreign edits)
    • Social (for lack of a better name): all the rest, no strict rules, just WikiEtiquette
  • Tools / Navigation
    • Searches, RecentChanges, FrontPage etc. separated by domain (of course there will always be a search-all / RC-all / ...)
    • Special navigation for some pagetypes: e.g. people will always show search field and "A-Z" bar to look for other people entries
    • SGF files (already in beta, but not ready for prime time)
    • Image upload (strict rules to avoid copyright conflicts)
    • As time permits: templating system that can be used to store different fields in the database (again people example: name, country, date of birth, rank, etc. which then in turn can be searched / sorted / etc (e.g. auto-generate time-line or sensei-pupil relationships))

• Editing rules, considering as example the teaching section:
  • no discussion on page
  • no definition of terms, no problems, etc (are in other sections)
  • content may be signed (if someone writes e.g. a treatise on the monkey jump I think it is ok, in contrast to e.g. the reference section, where there is no signing)
  • has to be linked from meaningful index pages (or be part of a well linked path)
  • has to have meaningful keywords and intended playing strength
  • depending on subject: meaningful links to problems, BQMs, etc.
  • too harsh? : new pages start out in social domain and have to be approved to get into teaching domain

• Migration
  • All pages start out in social domain
  • People pages and problem pages (e.g. problem series) will most likely be transferred en bulk
  • Other pages have to be transferred/approved by admins/librarians (depending on workload additional librarians will be assigned or special migration privileges granted to regular deshis))

Note: basically all this is up for discussion. Some points may be realized, others may not. I just jotted down what occurred to me (after going through previous discussions and emails). I don't give a timeframe for when features will be completed. My life is too unpredictable right now.

--Arno

Discussion

nachtrabe: Mostly looks good, though I am concerned with the transition plans. It may be best to emphasize a "fresh start" and require pages to be manually added rather than moving them over wholesale. This will cause a bit of a slow start and require SLII to be run simultaneously with SL for a while, but it will raise the overall page quality when SLII becomes primary (if nothing else a lot of the "people pages" could stand to be edited to fit to some form of standard).

DrStraw: I agree. When I proposed SLII I envisaged individual transfer of pages over time as people deemed them appropriate.

Hu: I am thrilled by the proposed changes. I would hope that if possible the changes could be added to SL, or folded in, so that there continues to be one site, with one point of reference, and most pre-existing outside links into SL will continue to work, perhaps with redirects or aliasing in some cases. Perhaps the folding in could be accomplished by having the current library as a domain in level between teaching and playing, with some pages chosen by reasonable consensus to seed the reference section.

I think that the forum aspect is an excellent direction to take to handle the more dynamic aspects of SL. A lively forum can make a site very popular and it doesn't mean that it need be dragged down by lowest common denominator, if it is well moderated. r.g.g. is unmoderated and frequently filled with nonsense, but even a well moderated forum can benefit from a little nonsense now and then. The name calling and feuding can be moderated out.

I can recommend a forum as an example of good social software. It is political so it does have a share of name calling and extreme disagreement, but it also attracts a large number of well written thoughtful posts. I find it to be one of the most intelligent political forum on the web and the most lively, though a few others come close. Whether you agree with the politics or not, examine it as a piece of social software that does not fall into the phbb cookie cutter mold and is better as a result. It is the Democratic Underground.

For the sake of Arno and Morten, I hope that the proposed changes can have the effect of in some ways simplifying Sensei's Library. Generally simpler is better because it is easier to navigate, easier to explain to users, easier to administrate, and easier to maintain. When in doubt, simplify. -- Hu.

Arno: maybe I should have been more clear about this: SL2 is just a code name. The changes are done in place, just like e.g. introducing subpages was done in place. So URLs remain the same and links will still work. I thought if I should go as far and have separate link domains as well, e.g. to reference keima you would have to write [ref:keima] and for your homepage you would have to write [user:Arno], but decided against it. So domains are mostly separated by rules and attitude and to a lesser degree by technology (e.g. separate RecentChanges available, small differences in layout, etc.). If you think that we should have different linking domains as well, then speak out :o)

axd: Pity, otherwise you could have donated the old SL with hair and all to the KGS community :-)

nachtrabe: I'm a "burn the bridges and start over" kind of guy--I believe that doing this in-place will be a mistake and make it difficult to enforce a higher quality article overall.

As to the spaces: I'd separate out User and Discussion in their linking domains. These are going to remain relatively static in terms of their links and aren't going to move between spaces very frequently, but wouldn't do it for the others, simply because if something gets moved it will violate the link-space (also because it would be a PITA to figure out what every page I wanted to link to was part of). The consequence of this is that it would be [user:nachtrabe] to access my homepage, but just [keima] to access the keima page.

DrStraw: I agree with nachtrabe. While it would be nice to do it in place (and the truth is that there may be no choice) I think a better product will be available by starting afresh. I have seen too many example from the business world where a bandaid fix was attempted and problems persisted.

Arno: Hm, I guess I don't share your concerns. By creating domains we are in fact starting out afresh. That URLs are the same cannot be such a deterrent, or is it? As everything starts out in "social" (maybe we should call it "lobby" or "misc") the other domains will be empty. And as guidelines have to be agreed upon beforehand, and those guidelines will get enforced, I don't see how we need a new wiki to adjust ourselves to this new mindset. Those who don't want to adjust remain in the "lobby".

An idea for separating link domains: I don't share nachtrabe's concern that it's impossible to know whether to link to [ref:keima] or [teach:keima]. Furthermore, there could be an automated branching mechanism for [keima] that either redirects to the domain (if keima is only defined in one domain) or shows an intermediate page that allows to choose among the different pages in the different domains. But as I said, I'm not really in favour of this change. Separating out user pages could make sense, discussions will be in their own domain, as they don't have names, but only numbers (you will most likely link to them like [topic:1234] or [posting:3456] -- I'm coding that already).

DrStraw: Ok. Maybe I misunderstand. As you are the author I will bow to your judgment. I'm sure you have given it more thought than I.

Morten: Some other points:

• The layout on different domains will be different so that one knows at a glance whether a page is in the Teaching domain or the Reference Domain. (As is the case with discussion pages now)
• The links will also be slightly different to indicate the difference between [teach: keima] and [ref:keima].
• The user who inserts a link into a text can just use [keima]. If the page exists in more than one domain, then SL will create the necessary links. This means that the browsing user who comes after can chose whether he/she wants to follow the link to the reference page, the teach page or the lobby page. We need to discuss how this is done; spelling out the different links one after the other? A drop-down list ? Spelling out one link with e.g. 3 different superscripts or subscripts? Another way?
• If e.g. the keima page is linked to from other sites, if there are more than one [keima], by default the page displayed should be the one in the reference domain, followed in priority by the teaching domain etc (prioritised list to be agreed). Alternatively, external links can incorporate a specific intruction, but this is optional (e.g. http://senseis.xmp.net/?keima:people.) This should make sure that existing links work and enable future specific links.
• For domain names, suggest "Lobby" instead of "Social", "Edu"/"Educational" instead of "Teaching"
• Additional domains:
  • 'People' (e.g. names in Go etc.)
  • others?
• We should look at some sort of mapping between existing pages and the 'new' pages in order to make a semiautomatic 'transfer' where possible. E.g., all 'discussion' pages could become 'lobby' ?
• We need to start working on the 'rules' which will apply for
  • Reference - RefDomainRules
  • Education - EduDomainRules
  • Homepages - HomepagesDomainRules
  • People - PeopleDomainRules
  • Lobby - LobbyDomainRules
  • Meta - MetaEditingRules - an improved collection of the 'metarules' which we already have on SL (WikiEtiquette, TextFormattingRules, etc. etc.)

• Also, we may want to agree on the use of the term 'domain' before we go further. (It's fine by me, but before it gets settled in, others may have another opinion)

Just some initial cents... finally, based on feedback, this page needs a WikiMasterEdit.

Bill: Speaking of which, I have been doing some thinking about WMEs that may be relevant to this discussion. Briefly, it seems to me that current complaints include complaints about too much editing and about too little. I agree on both counts. There is too much editing that is disruptive, too little that is constructive, in particular, WMEs. I am formulating some ideas about procedures for WMEs that will address both prongs of the dilemma. See WikiMasterEdit/Discussion for my suggestions.

Charles I don't like that. We don't need be timid! as an SLogan.

Bill: Charles, is it that you don't like my suggestion? If so, what does timidity have to do with anything?

Charles I dislike the discouragement to editing, implicit in emotive language like disruptive. Vandalism could close any wiki if it were a real issue; but we know it is not. Otherwise, words like disruptive suggest bad-faith edits. Timidity is what is induced by constant suggestions that the right to edit - the fundamental tenet of wikis - must somehow be rationed. (What I said also refers to the Wikipedia slogan be bold, which is justified by the ability to revert.)

Bill: Well, a good part of the suggestions for SL II involves restrictions to editing. My suggestion is in response to that. It provides restrictions to editing by buffering it, which is less drastic than other forms of restriction.

I did not mean to suggest bad faith by the term, disruptive. I was trying to characterize complaints people had expressed. One enthusiastic beginner, referred to above, inundated us with good faith edits that are still causing problems.

Gronk: I, for one, am very encouraged by the creative suggestions of Bill and others to try to address what has clearly become an issue for a large number of serious SL contributors and users: namely the rendering unreadable of lots of otherwise useful pages through lack of serious editing. I don't see the attack of Bill's language as very productive to this effort. If there are serious issues with the actual content of his proposal, that is a legitimate area of discussion, IMHO. I hope this effort to improve SL is not brought down by the same conservatism that has continued to allow so much negligence to go on for so long. Freedom with structure and limits might very well produce a more enjoyable site. Let's find out!

Charles This being a meta-discussion, I am saying what I like and dislike. I dislike language that 'warns off' potential editors. What can one prove with anecdotal remarks about individual cases? We all know the answer to that one: whatever one chooses. People have to be able to learn how to do wiki editing, without being beaten over the head with words like 'disruptive'. It's not as if SL has enough people editing it.

Bob McGuigan: I understand Charles's point about "disruptive" being pejorative. However I also agree with Gronk's description of pages being made less readable by edits. There are too few people doing editing on SL. Thus, when questions or incorrect comments that interrupt the flow are inserted into a well-written page, for example, it may happen that no one re-edits the page and the overall quality suffers. The changes under discussion are intended to minimize the occurrence of this sort of thing. Precisely because there are not enough editors on SL we do not have many people "watching" and re-editting pages, as often happens on WikiPedia. And the few regular editors simply can't fix everything.

Charles Well, I gave up editing here for reasons that I have discussed elsewhere, mainly off-site. I made the choice, basically, because I felt attitudes were against editing, in a rooted way.

Bill: By offering a measure of stability to certain pages (a minority, I think), I wanted to encourage editing, particularly WMEs. I also wanted a way to address the complaints of those calling for restricted editing that would not ultimately restrict editing.

Ian: I find the new system makes recent changes very noisy on the eye, that doesn't annoy me but I thought it would irk some. I think what people like about a wiki is that anyone can edit it. There's something special about the first time, wondering if it's really true or not. Will somebody turn around 10s after you've finished and subject you to flaming mellotron, or will nobody notice. For my wme nobody notices, i must choose boring topics.

Arno: @Ian: Anyone can filter out the talk pages. The default filters are there and if one spends 5 minutes on UserPreferences/HelpRecentChanges it should be clear how to make these settings permanent. Also, if SL continues to grow we will reach a point where RecentChanges will lose the usability it has now.

@Bill and Charles: I think that the guidelines we try to establish for the different domains do not restrict editing in any way, but instead will allow unanimous editing, moving or removing of content (e.g. the mentioned newbie questions or discussions on main pages). It helps to establish a common goal to work towards to. "Be bold" is nice, if all agree on the direction of that boldness. By introducing topical domains (reference, teaching, etc.) such a common goal is established. At least, this is what I hope. We will see how it turns out. And as pointed out, much of the content on SL can never (due to its intrinsic nature) be as factual as a encyclopedia entry (e.g. BQMs).

Charles BQMs work - one of the best parts of SL. I have been clicking around. Pages like shut in show that copyediting has been neglected - this should have been edited ages ago. Jowa is a mess; 50% should be on the discussion page, and the rest should be unsigned.

Steve This came up in the Romanization of Korean discussion: "DrStraw: I also know nothing about Korean. My sole exposure to it is through the names of Korean players and, recently, a few baduk terms. So I take the pragmatic viewpoint that SL is of no value to me if I cannot use the search function to find what I need. I really don't care which version of romanization is the primary one but I do care that any version I might encounter in my reading is readily found. Therefore considerable use of the alias must be made."

On this note, perhaps it would be possible to have a box on the "edit page" page, in which one can enter aliases for a page. The current method of creating aliases is rather inconvenient if you wish to add a Korean name or term, and then alias all the alternative Romanizations.

dnerra: I am wondering about the exact list of domains we should have. Splitting up in different domains only makes sense if the differences are clear, and they are not to me for some of the domains listed by Morten above. Reference section is obvious. Why do we need a separate people section? Shouldn't articles on people also by factual and objective, and could therefore be part of the reference section? Do we need to distinguish between Homepages and Lobby? Maybe, maybe not. I would really like to have the playing section, though. I think the discussions as on the BQM's etc. are really worthwhile, but I don't think they would fit into the teaching section (this of course depends on what we expect from the teaching section and is worth discussion). The difference between Playing domain and Lobby is easy -- playing pages are strictly about go. Playing go, discussing moves, posing problems, asking questions is on-topic, everything meta isn't.

Bob Myers: The questions of which domains to have is meaningless unless we ask: what is a domain. To make any sense, domains must be more than another layer at the top of the hierarchy, or a search/filtering convenience that could be handled just as easily with keywords. Here's my first cut: Each domain has separate:

1) roles and rules

2) look and feel

Such roles and rules might include, for certain domains, restriction of editing to particular classes of deshi. With discussions, that's a less onerous restriction than it would have been before. We need to be careful about slavishly following the Wikipedia model. The dimensions of the space of go knowledge and the shape of the pool of expertise differ from the sort of general knowledge that Wikipedia deals with, as has already been pointed out.

Another question possibly worth asking is: can anyone set up a domain? Then we would have a "hundred flowers" situation which could be interesting, if a bit chaotic. This model could almost be looked at as SL being a hosting service for go-related wikis and discussions.

Speaking of aliases, I am against the proliferation of aliases (and therefore against the proposal above to make creating them easier). The ThousandYearKo page, for example, has as aliases both MannenKo and MannenKos. I find it particularly unuseful to list every single transliteration variant of a foreign term as an alias. For one thing, that encourages authors of other pages to use inconsistent romanizations--which for page names, at least should be standardized (see Wikipedia's approach to Korean). I believe part of the problem people are trying to solve with aliases could be dealt with, as someone mentioned earlier, with more smarts in the search function, for example, "folding" o-macron with o.

Charles Matthews Well, on aliases, that's the kind of comment that really brings home to me how little people here really understand wiki. It's supposed to be hypertext, you see. Aliases reduce the overhead in creating hyperlinks. So they're good. They are a solution to inconsistent romanization - anyone who cares strongly enough can actually check back along the links, and fix it. Once the link is there. If it's not there, yes, you'd have to use the search function, but you'd have to imagine all those wacky alternate romanizations. Makes no sense to me.

Bob Myers: My goodness. Given my demonstrably poor understanding of the real meaning of wikis, I obviously have no choice but to defer to your wisdom. Thanks for your insight on hypertext. It happens to be an area I'm especially interested in--I own two original books by Ted Nelson, the originator of the term and inventor of the concept. But I'm sure you've already moved beyond his thinking.

Charles Quoting from Wikipedia on Ted Nelson:

''The Web owes much of its inspiration to Xanadu, but Nelson dislikes the World Wide Web, XML and all embedded markup, and regards Berners-Lee's work as a gross over-simplification of his own work:

HTML is precisely what we were trying to PREVENT— ever-breaking links, links going outward only, quotes you can't follow to their origins, no version management, no rights management. – Ted Nelson (Ted Nelson one-liners)''

I'm with Berners-Lee on this. Well, mostly. Backlinks are very good, on a wiki particularly, as on SL, and transclusions are also good once it gets to heavy lifting.

2021 Thread Revival

DuEm6: I have been browsing https://en.wikipedia.org/wiki/Comparison_of_wiki_software lately to find some suitable engine. Because I think it'd be nice if there's an SL II or similar project that handles CJK text better. After all, the game is most popular in China, Japan, Korea, and Taiwan. SGF integration will be probably a lot of work, or is there some ready-made solution? Like some embedded SGF editor.

Anyways, it'd be nice to have an updated and structured wiki/database of players, tournaments, and books. With correct names/titles in Chinese (simplified and traditional), Japanese, and Korean. Listing the different pronunciations/romanizations, editions, publishers, etc. They are currently a mess in SL. This is my main interest/concern, which seems to be different from what SL II aims.

Wikipedia contributors are doing a great job for Go-related stuffs. But they only care about high-profile players and tournaments.

tapir: One thing my years at SL taught me: It is pointless to wish for things that one is not prepare to do oneself. SL is older than wikipedia, afaik, and it seems very optimistic to hope for new features at this point. On the plus side SL runs stable and never begged for money.

bugcat: Yes, SL is older than Wikipedia. It began development in 2000 and "went public" in January 2001, eleven days before Wikipedia was launched.

tapir: There is btw another discussion at SL As An Ageing Wiki (quite some years after this one) that is probably more in tune with our current situation. Don't agree with all I said there anymore, but the gist of "stay open for various fresh pages, but clean the playground once in a while" still holds.

DuEm6: I personally think the Epic 2021 Thread Revival heading is unnecessarily snarky. But meh.

tapir: You could have just edited it.