DIY Wiki - articles Under Construction

[John Stumbles]

Following on from the discussion in the thread "Article on Pumps" I've been wondering if we could have an area of the wiki for articles that are so seriously 'under construction' that they should be kept separate from articles which are reasonably comprehensive and accurate. That way a casual visitor to the wiki doesn't get put off by dipping into articles which are half finished, but potential contrubutors who want to write an article on something can add to a partly-constructed article rather than starting from scratch.

As an experiment I've moved an article I've been working on, on showers, to ZZZ/Showers

's not quite as good as having a separate namespace in the wiki but it means that it appears at the end of the all-pages listing 've also created an 'Under Construction' template (imaginatively titled 'Under-construction' :-)) which I've applied to this article.

What do people think of this?

[John Rumm]

Seems ok, although I expect the template alone is probably enough... if you arrive at an article and the first thing you see is a message saying it is not done, then you can't expect too much from it regardless of its name.

[Jason]

It's a WIKI. Every page is *always* going to be under construction. There is no point stating the bleeding obvious ;-)

-- JJ

[John Stumbles]

I guess I wasn't clear in my post: the template does acknowledge that wiki pages are almost by definition always under construction. I was trying to indicate the difference between a page that's more like muddy field with a few survey posts knocked into the ground than a building which needs a few bits of second fit & decoration left to complete.

[Robin]

As a reader rather than a writer I'm inclined to think that there is some point in stating the obvious - especially when the subject may involve sharp tools, large currents, heavy weights, big holes, ...... The readers of the Wiki are not *all* widely experienced and deeply knowledgeable. And some may even have done stupid things despite them being manifestly stupid? (Me puts hand up reluctantly.)

[meow2222]

It may turn out to be just one more thing for argument to occur over, time to be wasted on, and contributors to not follow or revert. Telling others how to write articles is always a recipe for trouble. Its one of the ways mods unintentionally kill forums. Suggestions can work sometimes.

It may be better to have a link at the bottom of all pages to a disclaimer that explains why safety critical info from a wiki should never be relied on. That applies to all pages, not just a few. There is a para in 'House wiring for beginners' already explaining why. Maybe that could become its own article on wikis and safety.

Wikis are by their nature incomplete, in development, subject to differences of opinion and ever changing.

Another neater options may be to append 'under construction' or 'unfinished' to a page name when most of the info is missing, and remove it once most parts of the skeleton are filled in. However the great majority are not in the half-done stage, so I dont think there is a problem to begin with. Thing about wikis is that unfinished articles encourage visitors to contribute.

If someone has the impression that wiki articles are the Final Say-so then they need to know a bit more about how wikis work.

NT

[John Stumbles]

OK maybe I've got it arse about tit: rather than relegating stub/incomplete/unchecked etc articles to limbo (which let's face it ain't gonna happen because nobody's going to do that to their own work and it would seem offensive to do that to others') perhaps we should go the other way and award some brownie points to articles that /are/ reasonably complete and peer-reviewed? And as part of the QC/approval mechanism we could then include each newly-approved article in an index of the Great and the Good, which could be the Acceptable Face of the Wiki :-)

Maybe that would go some way to addressing concerns expressed when this wiki was first mooted about the anything-goes nature versus the peer-reviewd process of putting up articles on the FAQ.

For starters I'd suggest as candidate articles (to pick a few): cable crimping

heating (and sub-articles in this topic) outside washers include if they get a few Yeay votes.

In the longer run I'd hope that as people bring new articles to the attention of the ng and they're reworked in light of feedback from the group (as happened with the cable crimping article IIRC) they'd automatically get okayed as approved when everyone's reasonably happy with the end result.

[John Rumm]

Could have a little standardised graphic (sort of horizontal bar graph type thing perhaps - red when short going through to green at maximum) that shows the level of review and checking the article has had. Each article could simply link the appropriate graphic in the intro. Clicking the graphic could give some explanation of the meaning;

______________ |XX____________|

Article is still in the process of being authored - this version is a work in progress - disregard for now.

______________ |XXXX__________|

Article complete but has not been checked it may contain serious errors or ommisions

______________ |XXXXXX________|

Article is complete and has been proof read but has not been peer reviewed it may contain serious errors or ommisions. ______________ |XXXXXXXX______|

Article is complete and has been proof read and peer reviewed. There are some areas awaiting rework. ______________ |XXXXXXXXXX____|

Article is complete and has been peer reviewed. All review actions have been incorporated.

etc.

Yup, could work....

As a more general point, I get the feeling that some people are reluctant to contribute articles at the moment since they fear they will get hacked about with too much. Some sort of documented review process may go some way to reduce those fears.

[John Stumbles]

Yup, and that could be done with templates so all one would have to do would say (something like) {{new-article-outline}} {{under-construction}} {{not-checked}} {{not-reviewed}} {{approved}} and the template would include appropriate boilerplate

Difficult one. That's par for the course with wikis: "If you don't want your writing to be edited mercilessly ... don't submit it here." However in practice that generally doesn't seem to happen: if one person starts an article others tend to leave most changes to them, au contraire to Wikipedia's "be bold" culture and practice, for example.

[Andrew Gabriel]

There is no concept of approved in a wiki, by definition. Even supposing you could think of some way of defining an article as approved, it can't be assumed to remain as such unless it is locked against update. The nearest you can get to approved is proably a wiki article with, say, at least 10 contributors, with no one contributing more than some percentage, but even that needs refining (it's easy to engineer in a garbage article). I think you are perhaps trying to turn the wiki into something it can't ever be, and which our FAQ is.

Yes, that's happened to me. My time is in short supply, and I'm not likely to waste it again. I occasionally dabble, but if an article needs a lot of work, I now pass on it.

That's what the FAQ is. We could look to move generally acceptable articles into that, with the wiki being the more bleeding edge workspace.

[John Rumm]

The "changes after the fact" issue is hard to work around without some form of configuration control - at least in theory. In practice at the moment there does not seem to be much going on in the way of changes to established articles once they have reached some level of maturity.

Which article did you edit out of interest?

One thing the wiki does do well it maintain an audit trail, and ensure that all previous revisions of a document are retrievable. So any work you did will still be there (even if not in the latest version).

This is part of the reason I was keen on a content management system for the FAQ - to make updating it simpler and faster but also placing less demand on one person all the time for doing the updates. Perhaps a combination of the two would be an ideal - articles get developed on the wiki and eventually graduate to the CMS when enough people are in general agreement on the content.

I feel torn between the attractions of each at times. I read some of the stuff I have on the FAQ and get annoyed by typos and other kludgy bits of wording. If they were on the wiki I would just click edit and fix them. In the current FAQ they don't annoy me enough to download the page, edit it, and post it back to Phil just for the sake of a typo.

With the wiki there is the worry that comes with the lack of outright control. However I have to admit that so far, this has generally worked well - my articles are better than when I wrote them because the edits that were made were carefully done and have improved the readability etc. There have not been too many occasions where someone has changed something to make it less "right" either.

Perhaps one should just set notifications on articles that they have an interest in, and hence get oversight of changes.

[John Rumm]

Thinking some more on this...

One can tell[1] allot about the provenance of an article simply by looking at the list of names involved in its creation - just as one can make a reasonable judgement call on the likely accuracy of info posted to the newsgroup by the reputation of the posters.

The history capability of the wiki maintains one aspect of this very well, but what it does not record is names of the people who have read it, and agree or approve of the content, but have not contributed to it directly. As an *author* I would find it valuable to know that people with high levels of domain knowledge have read what I wrote and not found anything to complain about.

One are on the wiki I would personally like to see changed is the editing and account creation policy. I.e. have the default access be read only, and editing only be allowed by those who have created an account. That makes it much simpler to see who is changing what. It might also curtail the spammers a bit! Perhaps it should go so far as to have a sysop in the account creation loop. So that write access is only granted when approved by a human.

[1] Assuming one has been reading the group for a couple of years!

[John Stumbles]

I don't see why not: Wikipedia, for example, has a rating system for articles

from "stub" ("either a very short article or a rough collection of information") through "Start", "B-class", "Good article", "A-class" to "Featured article" (the ones you see on their main page).

I'm not saying that everything that WP does we should do, but they have been at it long enough and are big enough that we should at least look at what they do so we don't reinvent wheels.

No but you can indicate in the "approved" (ow whatever) notice what version got that rating. Then if major changes are made to the article it can be re-tagged. That does require that someone is monitoring articles for significant changes, and seems to me the fundamental difference between a wiki and a FAQ: when no-one maintains a wiki it tends to chaos, when no-one maintains a FAQ it withers on the vine. Neither is desirable of course and one can argue the pros and cons but I do think that a wiki has the potential to involve many more people in collaborative effort than a FAQ simply because of the ease with which one can contribute. Hopefully that also results in serious contributors taking an interest in maintaining the quality of the site by housekeeping and weeding - which does seem to have been happening with the DIY wiki.

AIUI (tho' IANAL) the GFDL does allow this. However there are ways of managing change within the wiki itself (again, Wikipedia uses these). Say for example there were problems with safety-critical articles (e.g. regarding electrics or gas) being recklessly edited in a way which compromised the quality of the safety information. In the wiki the page could be protected from editing so that any changes had either to be proposed on the article's talk page or made to a copy of the article (e.g. "Protected Article/Proposed new version").

[meow2222]

Lots of ideas have been discussed here for how to turn a wiki into a collection of authoritative articles. I dont think its possible, as wikipedia has found. It simply contradicts the basic nature of the wiki format. All suggestions made here have their good points, but also their problems, and all would require significant extra work (by who?) and generate controversy, neither of which is conducive to wiki development, or best use of people's limited time imho.

It is a wiki, and this fact shapes its nature significantly and inevitably. I really think the better option would be to educate any new readers to the nature of wikis. Awareness of the issues in life is the real solution IME. I've made at least a start on this with an article on wiki safety.

we wanted to copy some aricles over to the FAQ that would increase the database of authoritive articles, but its not a work-free process as the wiki articles will often change and be improved over time. And since theyre already on the wiki, I would question the true value of such extra work.

Maybe its just best to accept that the wiki is a wiki. And to continue to maintain other workspaces too, such as the FAQ.

NT

[John Rumm]

Use with the FAQ in that way would only be workable if the FAQ itself moved to a CMS platform, otherwise it is too cumbersome.

[John Stumbles]

I think Wikipedia has shown that it *is* possible to develop authoritative articles through a wiki-based collaborative process. I think that's where their peer review

aforementioned rating scheme for articles in.

In our own wiki I think cable crimping

outside washers good examples of informative, useful, accurate (as far as I can tell!) articles that - in the case of the safety-critical electricity-related articles - have been discussed in the ng and refined at least as much as articles that have gone into the FAQ.

[John Rumm]

cable perhaps... might be handy if someone else did some of those just for extra sanity checking.

[meow2222]

There's a difference between good articles and authoritative. Wiki can not be quoted as a reliable source, which means theyre not authoritative. They might be good, they might not.

NT

[John Stumbles]

I'm not sure I understand what you mean by "not authoritative". Wikipedia aims for articles which "have sufficient external literature references, preferably from reliable, third-party published sources with a reputation for fact-checking and accuracy (peer-reviewed where appropriate)"[1] Thus you might say that it aims to be an accurate guide to, and summary of, the authorities in a field, rather than being an authority itself. It is quite clear that the place for original research is elsewhere.

In the context of the DIY wiki this seems a sound principle for articles about topics such as plumbing, heating, electrics, building practice, safety etc, although I'm in favour of also having articles giving more personal accounts of projects and suchlike e.g.dehumidifiers for laundry drying, loft conversions and home-made heat banks.

In terms of authoritative-ness I think we should be aiming for articles on core topics which cover all areas in appropriate depth and convey information which can be shown to have foundation in external sources.

[1]

[meow2222]

Authoritative means one can rely on the information. This is not generally so with wikis - plenty of good stuff there, but rely on it? Not really, for many reasons.

An example is the rewiring tips article. Good yes, lots of good info there, but authoritive no because it contains questionable information about type c mcbs.

But an article meeting that is not always correct even at a basic level. Info from references can and at times has been put together in ways showing the writer missed some concepts and thus misinterpreted the info. Nor is it necessarily a fair representation of the knowledge in the subject. Nor does it necessarily get issues in proportion, nor are the linked expert infos necessarily accurate. Etc.

We can write articles of all qualities, as is occuring, but to put our wiki on a pedestal of being an authority is a bit questionable for all the reasons mentioned on

anyway

NT