Difference between revisions of "ReaderCentricDesign"

Latest revision as of 15:46, 21 April 2010

This page was first written on 17 March 2006. Many of the problems in mentions may (hopefully) no longer exist by the time you read it. -- Wordman

This wiki is a great tool for authors to organize and share their Exalted creations and the current organizational scheme works very well for authors. Consider, however, what a reader who approaches the wiki for the first time wants to get out of the wiki. As a small subset of what he or she might want to do, consider the following examples:

• Find some celestial circle spells people have written
• Browse a list of three dot artifacts
• Browse a list of (say) solar melee charms.
• Take a look at the Temperance-related Limit Flaws people have contributed.

What kind of experience will such a reader have when doing each of these? They'll most likely start with the WikiContent page in all cases. For the first item in the list, he finds the Spells tag, which leads naturally to the Sorcery page. There, a nice list of spell names, grouped by circle, awaits. Not as many as I was hoping for, he might think, but so far, so good. He briefly looks at the "New Sorcery by Contributor" list, but largely ignores it, figuring it is probably the same spells just organized differently. But then he notices that the names in the list are different than those in the well-organized list up top. He clicks on one at random and sees a single page with several spells on it. WTF? Why aren't these listed in the main list on the page? He goes back and clicks on a few more. I have to plow through all of these like this?

He tries Artifacts next. He sees the same list of contributors, but there are some categories up top, so he tries one of these first. Ah, this is more like it! Grouped by rating... short descriptions... sweet! Then he clicks on one of the artifacts to get more detail, and is brought to a page that lists an artifact completely unrelated to the one he clicked. He quickly realizes that the page scrolls down, and ten items down, he finds the artifact he clicked. Why wasn't I taken directly to this one? And why are these all grouped together on this page?

He tries Charms next, clicking links to get to SolarMelee. Here, all there is is a contributors list, with no other organization at all. How do I just get a list of charm names that I can look at individually, like on the sorcery page? The hunt for Flaws is even worse, as the VirtueFlaws page isn't even broken out by virtue.

While most of the people who remain on this wiki have learned to deal with this organizational scheme, you find people on forums all the time who couldn't find what they wanted and never came back. Even for regulars, the organizational structure, while great for contributing, prevents getting the most out of the wiki. As someone who is predominantly a reader of the wiki, for example, I can say I've never even looked at the charms section, even when I had a specific need, because of the way it's organized.

The rest of this page suggests some ways the layout of the wiki could be made more reader centric.

IDs vs. Paths

One of the most powerful features of a wiki is that pages can be used for both content and organization. Multiple pages can organize the exact same content in completely different ways. In most wikis, pages are the sole mechanism for organization and page names are used more like unique identifiers. In this wiki, however, page names have been used organizationally for some reason, with the names acting sort of like "paths" in a file hierarchy. The problem with doing this is that a hierarchy that makes good sense for one purpose (like organizing pages by author) is not necessarily the best hierarchy for other purposes (like letting readers find things). The vast majority of wikis solve this problem by only presenting organizational hierarchy in page content, not their names. For example, page names in the Wikipedia form unique identifiers and are not hierarchal at all, not even for "disambiguation" purposes, yet the Wikipedia provides many different hierarchal organizations of their pages within the content of the pages themselves.

At the time of this writing, using flat page names as unique identifiers is somewhat foreign to the culture of this wiki. For example, I'm sure somewhere there is someone already complaining that the name of this page isn't Discussions/ReaderCentricDesign. Much of the discussion that led to the current system revolved around how subpages should be used, but the question "should subpages be used at all" didn't get much mind space. The BestPractices page reinforces the idea that page names, not links within content pages, are the best way to connect pages to their parents, in spite of the most successful wikis not using them at all.

But, so what? So this wiki uses subpages? Why does this make it harder on readers? The reason is that if you assume that page names should contain hierarchal organization, then you must pick a "favored" hierarchy by which to organize. In this wiki's case, the the organization chosen (by author) is the least likely to be helpful to readers. Very few come to the wiki with a mind to "find all the charms by Bob". The BestPractices says that splitting a page, for example, for a certain type of charms "by author would be an good use of subpages, since it is the lowest level of Charm taxonomy". This is demonstrably wrong because a) individual charms are the "lowest level of Charm taxonomy" and b) there is more than one possible "taxonomy" by which charms may be organized.

If you take a look at the way many of the main pages are organized, such as the sorcery example above, you'll note that those pages that actually do provide a decent reader-centric organization do so in spite of the current use of author-centric subpages, rather than because of it. Pages like this prove that it is possible to build reader-centric design even with hierarchal names, but the use of hierarchal names at all seems to produce a culture where this is the exception rather than the rule.

This idea is actually one that might be worth looking into since it's more intuitive then the scheme we have in place now and the hardest thing for most new posters it getting used to the path scheme rather then using making a new top-level page for every single thing they want to post. However doing this effectively would pretty much require the next item on your suggested revisions list be implemented and that idea has far, far more going against it.

This idea is worth checking out though. It might be something we can do, despite being a hassle to re-format. - Telgar

One Item Per Page

The artifact pages illustrate the second huge obstacle to reader-centric design. One of the powers of a wiki is the ability to organize data in multiple ways. The artifact pages seek to organize them based on artifact purpose, then by rating. You can easily imagine another set of pages organizing first by rating, then by purpose, and yet another organizing artifacts alphabetically. It becomes very difficult to do so, however, if the items being linked to are not on individual pages. The results are less than stellar when the target pages of artifacts have dozens of items on them (particularly when the links always lead to the top of the page).

The obvious solution here is to author pages with one artifact (or charm, or hearthstone, or whatever) on each page, with other pages that link to them in various ways. At the very least, authors should use header tags, as then at least one kind of mutant link can target sections directly.

This, however, raises its own problems. Authors seem to really like having the full text of all their stuff on a single page. It also makes certain natural organizations, such as martial arts styles, harder to read if you have to follow a link for each charm. Personally, I think that the flexibility provided by the "one item per page" philosophy to do things like link to individual charms within character descriptions and provide multiple other ways to find the same data is worth the added hassle. There are, however, some technical alternatives.

This is a rather large suggestion and has some rather large problems. First, bandwidth would increase drastically if you had to load 50 pages to read someone's brawl charms rather then 1. Also due to the lack of wikitrails in this software, it becomes very hard to navigate Charm Cascades when they're separated on different pages. Non-charm contributions fare better but Charms make up the majority of the wiki's content.

I have more Artifacts on the wiki then anyone else, I believe the total number is now nearing 200. Not only would it be incredibly annoying to have to navigate through 200 pages to read my artifacts but it would make it hell to update. Each index page would have to be updated every time a new item was added and there would be many of those. That's annoying, it's not at all helpful.

Overall, this is probably the most troublesome suggestion because this wiki, primarily, is for the writers not the readers. - Telgar

Includes

Some wiki software allows you to embed the full text of other pages "in place" within a page, instead of just linking to it. The bad news is that I don't think the software running this wiki does. As an example of what is possible though, consider the following links (these go to a private wiki used for tracking personal campaign information):

• a page containing solar combos
• a page containing thrown combos, grouped by exalt type
• a list of combos by name, group by exalt type
• a page containg combos I've written, grouped by ability, then exalt type

Note that the combo Crippling Firefly Infestation shows up on all of these pages, but only one copy of the text actually exists. The software running these links happens to be MediaWiki, the same software running the Wikipedia. It calls including pages like this "using templates". I'm sure other wiki software can do something similar. In the MediaWiki syntax, the complete wiki code of the Thrown Combos page is:

==Solar==
===Crippling Firefly Infestation===
{{:Crippling Firefly Infestation}}
==Terrestrial==
===Hurricane of Razors===
{{:Hurricane of Razors}}

This feature, or one like it, provides a good mechanism to allow both a "one item per page" philosophy and the ability to have "chapter-like" pages.

Going Forward

In spite of appearances, it is not my intention to start a revolution here. I do hope that if you've found any of this convincing, that you look towards pages that you have authored in a new light. In particular, changes made in pages you've created would go a long way in allowing the creation of reader-centric pages, even if you are not interested in building such pages yourself. Any of the following (preferably all of them) would be helpful (in order of usefulness):

1. Follow the "one item per page" method for things like charms, artifacts, hearthstones, spells, etc.
2. Use header tags (series of equals signs, see this page for an example) to but section headers in your pages. Such headers automatically contain anchors that allow linking to specific sections of your pages.
3. Treat page names as unique ids, not organizational paths.

Some technical changes would also help this, particularly introduction of "includes". This might be done with patches to UseModWiki, resurrection of the BurningWikiProject or moving to a different code base.

Comments

Your points are basically valid, I agree. However there's a small problem. Actually, two. First is quite simply that it would require a VAST effort, not to mention an extended period of forbidding people from posting new material, to reorganize the wiki into a new formatting structure. The work it took to create the Artifact Pages that exist now was considerable.

The second problem is that this wiki, unlike wikipedia, is not a collaborative effort to design a unified database. It's a place where authors can put up their work for other people to gawk at. And authors are notoriously large-egoed. It's their stuff, they wrote it, they want it seperate from everyone else's stuff.

Overall, the ExaltedWiki isn't designed for casual browsers and it really can't be. Sure, it'd be nice if there was a way to do that but there isn't one because Readercentricity is the opposite of Authorcentricity and, frankly, as an author, I'm more partial to the second. Wikipedia and other wikis like it are hell to add to and, I imagine, update and maintain, because they're geared for readability not editablity. Every time you add something you have to add it in a million places or it never gets noticed and that, frankly, isn't a happy thing. - Telgar

I, for one, think this is possibly the best time to begin a revamp of the wiki - no better time to put forth a new edition than a new edition, eh? Kukla

eh. Telgar

I agree with Telgar. These points are valid, and would it would be great if there was a way to combine the two approaches. Perhaps some kind of 'meta pages' creation software could be coded to run over the top of the wiki to provide more 'readercentric' viewing? I have no idea of the feasibility of that, but it's a nice dream. - nikink

Just noting that yes, the work to create the artifact pages WAS considerable... but people did it. Given sufficient time and motivation, I'm pretty sure you -could- convert everything we have now into a similarly readable format. I'd help with some of it... maybe a charm or two every time I browsed the wiki.
-- Darloth likes (and suggested somewhere utterly irrelevant) the 1-place many-displays idea, which although a pain to convert to could be made to work, especially if a slicing function rather than entire-page-replication function could be implemented

Telgar posted a number of objections, some of which are worth commenting on:

• it would require a VAST effort...to reorganize the wiki into a new formatting structure. It's true that if a "large push" was done to convert all content at once, the effort would be considerable; however, this is not necessary. Any authors who buy into the concept can change their own pages at leisure. To that end, I've added a bit to the "going forward" section above.
• not to mention an extended period of forbidding people from posting new material. This is simply wrong. Even if the (evidently unlikely) choice was made to move to a new code base, it might take a while to write code to move the data from one database scheme to the next, but this code could be run fairly quickly, resulting in downtime measured in hours, or even minutes.
• The work it took to create the Artifact Pages that exist now was considerable. Yes. That's kind of my point. It's far more difficult than it could be.
• this wiki [is] a place where authors can put up their work for other people to gawk at. Being able to "gawk" requires that you be able to actually find something at which gawk. The likelihood that you will gawk decreases as the effort needed to locate and use material increases.
• authors are notoriously large-egoed. Then you'd think they'd be interested in making it as easy as possible for people actually find and see their work, so that the fawning it will inspire can commence. If they're not, then why post at all?
• they want it seperate from everyone else's stuff. And nothing in this proposal prevents that. Again, the power of the wiki is that it can use multiple organizational schemes at once. They can certainly organize their own stuff into a little sandbox. The point of this proposal is that, if they build their pages a certain way, other people could also organize it in a way that is useful to them.
• it'd be nice if there was a way to do that but there isn't one. I just spelled out how it could be done.
• Every time you add something you have to add it in a million places or it never gets noticed and that, frankly, isn't a happy thing. The alternative is to do what is done now: never get noticed at all. It's true that you do save a lot of work for yourself by not helping your readers find your work. At that point, though, why post at all? It seems to me it's seems like the intellectual equivalent of an exhibitionist masturbating in a alley; sure, someone might see him, but what's the point? Additionally, I can imagine a number of technical fixes to the "have to add it in a million places" problem, though they'd all require coding and would be useless without the "one item per page" strategy.

Kukla also mentioned that with the second edition coming, there is "no better time" to turn over a new leaf. I agree. The second edition will add yet another axis on which data can be sorted, and having articles which lend themselves to whatever organizational scheme evolves can only help.
- Wordman

Don't know if it was mentioned yet.. while it would be nice to have an artifact on each of your pages, and then be able to link to the specific artifacts with a short description.. it's pretty much impossible to organise so much under the wiki.. simply because of the character limits and the fact you can't go further than like.. 2 subpages isn't it? Sure.. ideally we'd all have things like.. Artifacts/FluffySquirrel/SpecificArtifact .. but.. we can't really do that.. so pages with all the stuff crammed on them tends to be the way it gets done ^^ .. the multiple artifacts on the page thing could be solved by anchors probably.. but I seem to recall I hit some snags when I tried it before.. although it might have been bad memory on my part and getting the code wrong.

But yeah, if every author just does their own pages.. .. it wouldn't seem such a huge thing anymore.. because people will only be putting in the effort they've wanted to put in from the start :) FluffySquirrel

Problem is it isn't just a thing about everyone doing their own pages. Who sets up the indexes? Who creates the new structure itself and who tells the people that are gone, inactive or rarely active to change their pages? There's some great content from people that no longer post regularly or just don't care enough to change their pages (me). If you folks come up with a way to work out this new structure of yours with the base level of organization being specific entries, good for ya'll. I think it'll suck horribly for Charms and setting material. It'll work fine for Artifacts, Hearthstones and such that are self-contained but not for anything else due to the lack of wikitrails and the need to see cascades as a whole.

The whole thing is a good idea, if it didn't require insane annoyance when adding new items and didn't require a total re-write of the wiki structure and didn't happen to make Charms unreadable. - Telgar

I do agree with Telgar here. Projects have been launched before to make the wiki prettier, more useful, whatever, and almost all of them failed because of lack of participation. There's a completely insane number of authors here, and the majority of them won't be porting over their own work. I think that the idea of trying to apply the new standards to all 2nd Edition material is a much more manageable one, though it will lead to two parallel wikis -- where 2nd Ed material is in a different format from 1st. This isn't actually that bad, but it may lead to a lot of really great 1st Ed stuff being ignored.

Re: the author thing, a fair number of people prefer to have stuff organized by author because they prefer certain authors over others. It's not necessarily that hard to get noticed, either: if you just post stuff and then don't do anything else, sure, people won't necessarily look at it. But if you really want feedback, you should go to other peoples' pages and give them feedback, and they'll often return the favour. (At least, if they're still reasonably active.)
~ Shataina

I'm reading this with great interest. I have been thinking about setting up a second wiki for 2e. Undecided issues: what software to use, how to migrate from this wiki. Mediawiki is pretty sweet, and has the template inclusion advantages mentioned earlier, but the markup syntax is different, so y'all would have to learn new markup, and migration of pages would suck. - Xyphoid

Migration, if done, should be done with code. The translation wouldn't be that bad, as this wiki's syntax is fairly close to MediaWiki's already. I'm not sure how different the underlying databases are. Wouldn't surprise me if the code already exists somewhere. In my own campaign wiki, I've been looking at an extension (which turn out to be much easier to write than I thought) which allows you to automatically build lists of links using the category tags that the MediaWiki uses. I'm looking to use this to eliminate the list building drudgery that Telgar mentions. For example, if I define combos with tags like Category:Solar Combo, Category: Melee Combo, Category:Defensive Combo (or whatever) I can then add tags to the main combo page that automatically builds links to any pages that have those tags, in whatever combination. The code for this could probably be leveraged for use in UseModWiki, but I'm not sure if it has an extension scheme or, if so, how it works. - Wordman

Answering my own response, MediaWiki includes migration code for pulling data from UseModWiki. Looks like Wikipedia originally used UseModWiki, then migrated. - Wordman

Migration? Scrumptious! Bring on the new dawn! Kukla

The White Wolf Wiki has been launched and operates on MediaWiki. There is 0 content at present and thus the perfect time to design a new, readercentric method of organization exists. I may at some point even move some Artifacts over if the site becomes well populated and interesting. On that...my doubts remain , see WhiteWolfProposal for more. - Telgar

In the past, there have been a couple of pushes to make this wiki better-organized and more accessible. They all failed; the majority of people who post here seem to really like the idea of having their "own webpage," but strongly resist taking the effort to make it accessible or transparent. It's also not helped by the fact that no one can ever reach agreement on which standards should be promoted.

That stuff Telgar has been saying about writer-centricity is some of the most depressing shit I've ever seen posted here. It's the direction ExaltedWiki ended up heading in, yeah, but it is a banal and insular direction. It's not something to be celebrated. _Ikselam

Show me a way to make the wiki reader friendly and well-organized that doesn't involve a massive amount of work (once completed mind) to simply add a new Artifact or cause migraines when trying to navigate Charm cascades. Then I will dance around in joy and lead the way in implimenting it, no matter how much work is required to do so. I haven't seen one that doesn't involve a total migrate to the new White Wolf wiki. If you'll recall, Roger, I was at the head of several of the attempts to rework this Wiki into a prettier, more readable and accessable place. In fact, I've been near the head of ALL of them. I'm not against them by any means. But I haven't seen one that will work. - Telgar

re Wordman's proposal, I think most of it makes pretty good sense. I think you'd have to be pretty careful with the one-thing-per-page idea, though, especially when it comes to Charms; Charms' relationships to each other are very important, and that would be easy to lose track of. You would have to be scrupulous about linking to both prerequisite and dependent Charms from each Charm's blurb.
The current organizational scheme seemed like a good idea at the time, but it was largely a result of me, and the other people who worked on laying groundwork in the first couple weeks, being totally unfamiliar with how wikis function. It's also a result of the wiki software used here lacking a number of useful features, such as inclusion or trails -- those things would make accessible organization much less onerous. _Ikselam

Wordman, could you write a tutorial on how to do reader-centric design over at the White Wolf wiki? Get it in soon and it might become standard. Kukla

Ideally, what I'd like is a means to go Artifacts/Trithne/(Artifact). That way, the main Artifact page can be organised as it is now, into offensive, defensive, what-have-you. From there, it links to individual artifacts. People can have their own artifacts individually organised and set as one artifact per page, and then the page at, for instance, Artifacts/Trithne can link to all of my artifacts in one place, and the page at Artifacts can have a list of artifacts by contributor, at the bottom, for those who prefer a particular author's work. I've tried to do much the same as it is, but because I'm limited to one subpage, I've had to create the top level page TrithneArtifacts to do what Artifacts/Trithne would do under ideal circumstances. A wiki is a website, and therefore can have a series of links that connect the page in question to every related page. But, this will take far too much effort, and we'd have to make the link to BestPractices on the front page size 62 font or something. - Trithne, who knows how he'd -like- to arrange the wiki, but it's too damn big.

Not a huge WikiFanatic so I am not up-to-date on the state-of-the-art of wiki-design. I have, however, started using http://www.tiddlywiki.com/ as generalised notebook and world-building tool. It has a feature called "tags" which you attach to a page, and once saved become meta-identifiers of sorts. So, if I wrote a new Solar Bureaucracy Charm it would likely be tagged: "Solar" "Bureaucracy" "Charm" "Moxiane", and then you can ask it to bring up every tiddler (tiddlywiki's version of pages) with a given tag.

Taking advantage of this feature (which I don't know if exists outside of tiddlywiki) would mean that reader- and writer-centricity would become more or less the same thing. If you want a list of all Solar Melee Charms you simply go to the metapage that tells the wiki "Generate a list of every page that has been tagged with Solar, Melee and Charm". Or you could just bring up every page tagged with a given author's name. - musing Moxiane

A fun feature (that may or may not exist) would also be to not only include the tags "GreenLantern" and "Moxiane" but also "Edit: GreenLantern" and "Edit: Moxiane" so that it's possible to see where someone's edited without cluttering up the page of things they've written. Still, I think we're definitely on to something with the tag thing, and wish we could get it implemented on the traditional large scale required for the Exalted Wiki. Especially if it's possible to define tags that can simply be selected from a list, to prevent people using tags like "SolarBureaucracy" and "Bureaucracee" that mess things up. -- GreenLantern, who really could change the world with only 25 trained developers at his side

I love metatags, and heartily support this idea. Of course, I don't know if the modified UseMod we're using here can do it though. - Trithne

I'm still digesting a lot of these comments, but one thing that sticks out at me is many of them imply that there needs to be a One True Way to organize the wiki. This is not true and is not the point I was going for. As I've mentioned before on this page the power of a wiki is that it can be organized in many ways at once. Even if people started using my suggestions above, the wiki could still be organized in the exact same author-centric way it is now. The point is that such changes would also allow a more reader-centric organization at the same time. In other words, I'm not looking to replace one organization with another. I'm looking to add new organization in addition to the one that exists. Telgar asked "Who sets up the indexes?" As far as he is concerned, the only answer that should matter to him is "someone else". He doesn't want to do it, because he only cares about author-centric stuff. In my opinion, the best answer to the "who sets up the indexes" question is: "software". All the software solutions I can think of rely on authors somehow "tagging" their posts. I think it would be worth making technical changes to support this, and then lists get populated with the stuff that authors have tagged. If authors choose not to tag, then their stuff doesn't show up in the lists, and can be found the same way it is found now (i.e. not very well). As for charm cascade/readabiliy, the "template" idea seems like it would solve this just fine, allowing authors to group them on a single page just the way they like, but still allowing them to be linked to individually and grouped in other ways as well. As an example, all the charms on this page are actually on their own pages, one charm per page. -- Wordman

Wordman, where does the wiki stand on this topic of discussion? I know you've implemented some spidering tools for certain wiki elements. Are there procedures in place for "tagging" posts, includes, etc? I'd be inclined to do a bit of reorg if there were tools in place to facilitate it. - Quendalon

There are no changes. There are no tools. Considering we still can't even do InterPageLinks over a year after I supplied the single character code change needed to make it work, there's no reason to expect there ever will be. I've considered using, sort of "metablocks" information combined with the spider stuff I'm doing (see Wordman/MartialArtsRequest), but they are clunky. Since originally posting this page, I've been using a semantic plugin for my MediaWiki-based campaign site. Semantic wiki-ing is pretty flippin sweet. - Wordman

Not that I know the headaches you've been going through to get this working, I noticed on UseMod they specify a patch made to rectify that situation. It was at this -> http://www.usemod.com/cgi-bin/wiki.pl?WikiPatches/BracketLinksWithAnchor and http://www.usemod.com/cgi-bin/wiki.pl?WikiPatches/AnchorsInFreeLinkBugFix <- addresses. Been beating my head against it also, but I have to agree, this appears to be something that coding only will solve. - EwindaleMoss