Wikipedia:How Wikipedia looks to newbies

Invitation

Please feel free to contribute constructively to this essay.

If you have something controversial to add, please consider discussing it on the talk page first.

 

This essay is about How Wikipedia looks to newbies, which is very different from how it looks to experienced editors.

Editors are busy people with their own concerns, and, often, deep inside knowledge of how Wikipedia works, and (therefore) of what a Wikipedia article consists of - citations, external links, footnotes, sections, lead-in, imageboxes, categories, past discussions, previous visits to AfD, edits by sockpuppets and so on and on.

Newbies have no such knowledge. To the newbie, Wikipedia articles look like text: which, astonishingly, they can edit. So they do.

This essay considers how a newbie sees an article, and what Wikipedia should do to enhance the newbie's vision.

What's there and what the newbie sees

edit

The newbie sees text, text, and text

edit

To the newbie, Wikipedia articles are made of text. And indeed, the newbie is right: read it, and it's text, one word after another. Press the edit button, and a text editor appears.

As text, Wikipedia articles look (to the newbie) much like traditional encyclopedia articles, except that they are editable parts of a website, like comment fields in a blog. If the newbie is constructive (this essay isn't about vandalism, so let's Assume Good Faith), and something seems wrong or missing, the logical response is to fix the error by tweaking the text. Which is fine if it was a typo or a grammatical slip; less good if it adds a true fact (without a citation); and less good still if it adds a valid opinion (but others think otherwise).

For a Wikipedia article isn't a text at all (or only at the base level of how it's physically implemented, just as at the physical level, a telephone call consists of electromagnetic oscillations: while at higher levels it consists of data structured by signalling protocols, and at the top level it consists of human conversation).

It is, probably, a mercy that newbies don't know what goes on at an AfD or at NPP, or they'd probably never contribute at all, or not more than the odd comma. Trouble is, the newbie does need some knowledge of how an article is to be judged, so he or she is motivated to spend time working on non-obvious things like looking up references (and let's not even mention image copyrights) rather than just venturing facts and opinions.

What I thought as a newbie

edit

If a little reminiscence is allowed in an essay, I recall that I was astonished to find that a sentence of skilfully-handcrafted fact, summarizing another article and correctly linked to it, could be treated as a problem. After all, the other article had been carefully edited and fact-checked, no? (Surely it had been. Millions of other editors couldn't have been wrong. This was a global encyclopedia, after all.) But there it was, a big angry-looking message from a cross-sounding and presumably terribly experienced editor. You mean, you want me to prove that those facts are true? Why? I linked them already! And I knew what I was writing about. And I'd taken care to write it properly. Some people!

The trouble, of course, is that the newbie is often (usually?) totally unfamiliar with the idea that facts have to be traced to sources all over the place. Or, for that matter, that facts are not self-evident. Or that Wikipedia is not a reliable source. These things are not intuitively obvious to the newbie – quite the reverse, they seem utterly extraordinary.

On what the world thinks of providing sources, for instance, I recall a Technical Report that somebody presented in the software industry. It was a dull meeting, and I was naively interested in a bored kind of way in where all the information had come from, so I flicked through it looking for footnotes or endnotes or whatever. There weren't any. At the back, there was a nearly-empty page called References, supplied presumably by a template or standard, and which the author had felt obliged to include if only for ceremonial reasons. It had two entries. The first was the brief from the client, which did not support more than a tiny fraction of the claims made in the Technical Report (if it had, the Technical Report would not have been necessary anyway). The second was to the Microsoft Excel User Manual. Obviously it didn't support any of the claims either. And that was it. Nobody in the meeting seemed at all concerned about the lack of sources of evidence. Indeed, had anybody been foolish enough to ask, they would have replied that it was the job of the person writing the Technical Report to come up with the facts and figures that it presented.

Needless to say, the newbie with that particular worldview is in for a rude shock when they start editing Wikipedia.

What we know doesn't work

edit

Templates

edit

Would a template do the trick? Well, no, we know it wouldn't. How?

  • For one, newbies using the Article Wizard frequently leave the sensibly-provided 'References' section, with its 'helpful' list of asterisks, as blank as they first found it. Incidentally, advice like "incorrectly formatted and unreferenced articles are often deleted" doesn't help either; the statement is true, but it glosses over the very different policies towards formatting and referencing – one is tolerated and cleaned up, the other provokes Wikipedia's wrath; and more importantly, the people who need to know they should be creating references don't read, mark, learn and inwardly digest that bit of guidance: it's literally 'off to the side', out of the newbie's field of attention to the task in hand. Which is to create text. Now.
  • For another, we have all seen how newbie contributions to an existing article don't even glance down to the Notes, Bibliography, Citations, References, and External links sections – what a lot of apparatus we have! – since to the newbie, the article to his front is pure text. The text editor says so: Edit. After that, there's really no guidance, beyond the keyboard's invitation to QWERTY the text field, and the screen's simple command to Save page - and a page, obviously, is a container for text.

Advice and more

edit

Would more advice and policies and procedures and processes and rules and standards and committees and warnings and punishments do the trick? You know the answer to that already.

Suggestions

edit

Perhaps it's none of an Essay's business to say what policy to adopt, what changes to make.

However, newbies create most new articles; good editors are in short supply; and everybody was a newbie once. Anything we can do to give newbies a more realistic view of what an article is, and how an article may correctly be edited, would be an improvement. So, here are some general suggestions. No doubt you can easily think of many ways of implementing these suggestions.

Guidance, to hand

edit

What is missing is the approach (you could call it a philosophy) that says it's better to show people how to do a task right, than to let them get it wrong, and then to BITE THE NEWBIES. It's really nice that there's a policy not to bite the new guys and gals; just a pity that the way that the most fundamental tools are set up - as a Wiki - invites exactly that. For if the invitation really is "anyone can edit", but the user interface is an (almost) bare screen, then we shouldn't be surprised if things go awry.

This isn't the place for detailed suggestions on how to design a user interface. But if you think of the basic principles you observe when using a bank's automatic teller machine, that you are only able to make correctly-structured requests, that you are reminded of what to do at each stage, that you are given your receipt before you see your cash (and walk off without the receipt), then you will easily see that newbies could in principle be given a lot more support, right where they need it.

Help and guidance need not be purely a matter of software design, either. Links to the forms of human assistance that are available - asking questions on talk pages, chat, and so on - could equally be placed far closer to the locus of newbie action.

Recommendations

edit

Wikipedia should

  • help people to get things right, rather than letting them drift and then smacking them for getting things wrong.
  • provide tools (word-processing software, templates, context-sensitive help, ...) that guide the newbie to make correct edits.
  • steer newbies towards guidance (automated and human) when they attempt their first edits.
  • provide suggestions and examples of how to edit correctly, in-place rather than far away in some undreamed-of policy document.

Practical suggestions

edit

Rereading this essay 6 years on, I note that I described what was wrong, but did not suggest how to fix it, other than in such general terms that nobody noticed, I believe. Well, what should an article look like? I think there are two parts to this: new articles (not a blank slate, then), and every existing article, at least in theory.

A new article ought immediately to have an inevitable, inescapable, muscular structure, that says "I have a title, a lead section, a body made of more sections, numbered citations in the text, and a references list." And every time the newbie adds some text, the article replies "and now please describe the source that you derived that text from".

Backfitting the millions of existing articles to such an architecture is plainly a challenge, but not unlike many that have been faced already. Much of it is a matter of presentation, again, so that the bones and muscles of an article - its reference structure connecting claims to sources, to be quite plain about it - are fully revealed, and cannot be broken by snapping a "</ref>" off anything. In other words, editing must be by a structure-savvy mechanism, one clever enough to cope with every complicated twist and turn. For a "visual editor" that takes you a few steps on the way, and then dumps you in the road, is worse than useless: it's a trap. We need a structure editor that builds cited-facts, and maybe even helps us assemble cited-facts into cited-sections (with "main" or "further" links tied to the section headings). That would be progress.

See also

edit