Approved editions - now with added GRAVI

[Peter Hitchmough]

Could the highest quality versions of articles always be available to users by default?

Open source software has a planned release policy. This offers a lesson in managing quality of the Citizendium. There is a deep analogy here between open source software and Wiki publication.

The analogy between Open Source Software and Citizendium
Many OSS projects perform automatic builds many times a day, but they maintain a daily or weekly release schedule of higher quality minor releases. Major releases take place even less frequently.

- In the same way, any author could contribute to articles on the Wiki, but they would not have the status of released versions.

Who decides what goes into a software release? The key figures working on the components.

- In CZ, editors could approve articles to enter the "daily build" or "major release" edition of CZ. They in fact act as editors.

How are the various software builds used? Developers continue to work on unpublished versions of the software. The public users have a choice of using one of a number of approved releases or the latest unapproved release. These releases may be located at different web URLs.

- In CZ, users could use a standard URL to see "approved editions" of articles and still retain the chance to work on and edit unapproved versions. These versions would have an alternate URL., or could be selected explicitly through preferences.

I envisage the normal CZ pages always to show an approved edition unless no approved edition exists. The benefit is that the flow of author contributions is not disrupted but published quality remains high.

- Peter

[Jon Awbrey]

Zounds goot -- but who will debug the debuggers?

  Jon Awbrey

[Nicholas Kaye-Smith]

A bit from another thread (http://textop.org/smf/index.php?topic=51.msg520#msg520) with my response:
Posted by david91
This thread is now reaching a critical point. I suspect that a majority of those in this forum do not want to see a Wikipedia redux. The only point to this project is that it achieves what Wikipedia has signally failed to do: namely, to produce some degree of stability if not growth in quality. We cannot expect to gain the acceptance of the academic world if, ten seconds after a page has been approved, it is reduced to rubble by a passing person who has no expertise. To me, it is absolutely inconceivable that we should be preparing to repeat the fundamental design error that has prevented Wikipedia from achieving some degree of long-term respectability. I believe that we need a dynamic system. My area of expertise used to be law, legal policy and philosophy, criminology, and the management of information systems and risk. All of these fields are under continuous development. There are major new case precedents and statutes every day around the world, new security problems and solutions bubble in the IT world, etc., so keeping up-to-date is a major headache. That updating in line with current development might be achieved through the serendipity of people passing by would be of inestimable value. However, I am with Phil in spirit. I think that there should be two versions of each page: the approved version and the dynamic version that anyone who has registered can edit (really, just a retitled talk page with little additional storage or programming required to create or manage). The passing authors will get the instant gratification of seeing their words on the system which is part of the reward system for encouraging people to be active. But trusted copy editors will catch the typos in the approved versions, and the task of editors will be to review contributions to the dynamic pages and transfer material as and when appropriate to the approved version. If not this, then there should be some other system that represents a filter before anything other than a spelling correction goes live — and even that may be doubtful because changing from, say, a positive to negative could completely change the value of the given expression or sentence.


I think there should be two versions, but the "default" version (i.e. the version that is linked to) should be the dynamic version, and the approved version should be something like: pagename/approved. If we do not do this, the dynamic versions will not be visible enough, and growth will slow.
At default versions which have approved counterparts, we could put something notifying users that a more comprehensive and accurate version exists, and that if they doubt any fact in the article, they should check the approved revision.
As for protecting the approved version...as it stands, we could restrict editing of approved pages to constables, and have non-constables post on the talk page or on the dynamic version, which will gradually be accepted into the approved version, the next time a constable has a look at the dynamic version and approves it (i.e. no long-winded bureaucratic process). This, however, is at risk of being too complex. What do you think?

Posted by david91
I wholeheartedly approved Larry's new launch plan which allows a period in which a core of willing volunteers beavers away to bring as much of the material up to standard. But no-one in their right mind is going to spend days and weeks of their time producing maximum perfection for launch if that will all potentially be lost from public view in the first few hours of assault by the world. Putting a tag on a page that it is CZ approved makes that page a target. Those are the pages that will get vandalised first. The motives for such an attack will be many and varied. Some people will attack just because they take malicious satisfaction in destroying the work of others. But a growing volume of quality material in CZ will also threaten the commercial interests that publish hard copy material or run commercial web-sites offering material for students. A significant number of people will have an interest in seeing CZ fail. If we do not put up barriers from the outset, we will preside over a foundering ship. We are claiming to be the professionals who can deliver a quality product. If we make this claim, and fail to take even elementary precautions, the resulting failure will damage our collective reputations. I no longer care. I am long retired and have no reputation to protect. But why should anyone with a career ahead of them risk such a loss of face?

[David91]

Interesting.

Given the choice, why would anyone prefer to see an unapproved version which may be littered with errors and mistakes?

and

Will constables have the expertise to make editorial decisions on the quality of content?

Surely, if we want to be a safe haven for naive users, we need to offer them first sight of material guaranteed to be correct. Otherwise, a naive user may only have the patience to read the first presented page, assume it to be authoritative and then leave. We have to protect users from themselves. They are not going to read the dynamic version alongside the static version to determine the differences and then make qualitative decisions about which is the better content. If they could do that, they would not be using the site as a reference point in the first place (which I qualify in that an expert may be passing through the assess whether to recommend students or others to use the site).

This is the first time I have seen it suggested that the roles of editor and constable overlap in this way. I agree that we need a quick mechanism to approve the move of material from dynamic to static version. When I had pages on my watch-list at Wiki, I made these decisions on the hoof because I understood the qualitative issues. Since we are no longer just talking about a default action of reverting to a prior version, this reviewing activity shoud be left to the safe hands of those with accredited expertise (since I assume that individuals will be able to hold both roles, constables with the expertise relevant to pages should be allowed to edit them).

[Nicholas Kaye-Smith]

Interesting.

Given the choice, why would anyone prefer to see an unapproved version which may be littered with errors and mistakes?

Because the dynamic version has been edited more and thus is potentially better?
Making the dynamic version default is a way of getting people to edit. I prefer to think of it this way - the Dynamic version is what would be there if there was only 1 version (no compromises on stability just because there is a stable version) and the stable version is an article that would be put in a "Citizendium" cd, only maintained constantly.
Also, it's more likely that academic institutions will allow students to cite Citizendium if stable versions are available.

[David91]

Well, this is the first time I find myself in complete disagreement with the spirit of a post.

My apologies. In the dim and distant days when I worked in universities, I would never ever have considered putting a Wiki site on a list of recommended reading for my students. The only time I might have used the resource would be as an exercise to see how many mistakes they could find. The development of the powers of critique and critical thinking is indispensable for the better students and a Wiki site is a wonderful source of material for that purpose.

We are supposed to be engaged in reversing this attitude which, I would suggest, is the normative reaction among professional academics. The only means of doing this is to ensure that the naive user is protected and, with respect, I believe that the only means of achieving this is by making the approved page the default page. We label that page clearly with the legend, "If you think that this page could be improved in any way, we would welcome your input here." or some similar formula which points the user to a page which can be freely amended. So long as one of the editors reviews the suggestion and makes a reasoned response with reasonable speed, nothing good is ever lost, new contributors are more encouraged because their work has been confirmed by an expert (which is always good for people's egos), and CZ is on a steady growth path to genuine quality.

[Kim van der Linde]

So, what you get is:

{name}

and

{name}/draft

The latter is free for everyboy to play with, and the first is the pne the reader gets to see. If the draft has become better, an editor approves it and we go on.... Something like that?

[Larry Sanger]

Responding to this post (http://textop.org/smf/index.php?topic=51.msg519#msg519) from David91 and some of the discussion since then.

The only point to this project is that it achieves what Wikipedia has signally failed to do: namely, to produce some degree of stability if not growth in quality.

The short-term point of the project, I think, is to improve Wikipedia's content and build a new community around the task of improving it.  That means growth in quality.  And growth in quality does not absolutely require stability (the standard around which the quality of an article revolves might be higher).  The openness and bottom-up, self-assigning feature of work on a wiki explains both the dynamism of Wikipedia and, like it or not, also the instability of the WP system.  What I've proposed is that we place editors in control of content decisions, that we place the community under the guidance of a group of mature "constables," and that we use our real names.  I think the combination of these things will improve both the quality and the stability of articles.  In addition, we can select particular versions of articles to link to and even, as you and some others suggest, present to the public first.  But none of this will guarantee that editable versions of articles remain perfectly stable (which, I suppose, is necessarily true!).

We are supposed to be engaged in reversing this attitude which, I would suggest, is the normative reaction among professional academics. The only means of doing this is to ensure that the naive user is protected and, with respect, I believe that the only means of achieving this is by making the approved page the default page. We label that page clearly with the legend, "If you think that this page could be improved in any way, we would welcome your input here." or some similar formula which points the user to a page which can be freely amended. So long as one of the editors reviews the suggestion and makes a reasoned response with reasonable speed, nothing good is ever lost, new contributors are more encouraged because their work has been confirmed by an expert (which is always good for people's egos), and CZ is on a steady growth path to genuine quality.

Bear in mind that many of the CZ-edited pages will not be approved for some time.  But, if an approved page does exist, I'm not sure I have any problem with making the approved page the default page, labelled as you have it.  As to having work "confirmed by an expert," that all depends on what you mean.  If you mean that each edit would individually have to be approved by an editor, I can assure you that that would simply break the process.  There would not be enough editors to make the process work.  Rather, the suggestion--which I have been making from the start--is to have editors work shoulder-to-shoulder with authors, correcting their work on the spot, making decisions about structure and content on the article's "discussion" page, but still editing the article right alongside them.  On my most recent proposal, when editors actually want to approve a version, they edit the article as they wish, and then link to that version of the article in the history.

David, I wonder if you have seen the Wikipedia process in action, "on the ground," when it's working.  It isn't quite as horrific as you might think.  I was closely involved in designing it, so I know exactly where it works and and where it's broken--and I think the changes we'll be making will precisely address the areas where it's broken.  Unless you think that simply making available an imperfect article is a problem--which I doubt you do think.

Putting a tag on a page that it is CZ approved makes that page a target. Those are the pages that will get vandalised first. The motives for such an attack will be many and varied. Some people will attack just because they take malicious satisfaction in destroying the work of others. But a growing volume of quality material in CZ will also threaten the commercial interests that publish hard copy material or run commercial web-sites offering material for students. A significant number of people will have an interest in seeing CZ fail. If we do not put up barriers from the outset, we will preside over a foundering ship. We are claiming to be the professionals who can deliver a quality product. If we make this claim, and fail to take even elementary precautions, the resulting failure will damage our collective reputations.

Well, perhaps those pages will get vandalised first, but they will also be on many people's "watch lists" and so the vandalism will be undone instantly.  Furthermore, actual vandals (who just write curse words for example) will be instantly banned.  This is how it works with Wikipedia.  It will work similarly on CZ, although it will be even harder for vandals, because they'll have to supply a working e-mail address and there will be even less tolerance for their antics.  You might also bear in mind that bylines will not be on articles, though as editor you might have your name on the discussion page when articulating a decision about some point of controversy regarding the article, or when approving a version of the article.  So it seems unlikely that individual reputations will be much at stake, any more than the reputations of academics who work on Wikipedia--there are quite a few, believe it or not--are at stake.  For them, it's merely a hobby.  Nobody's going to be denied tenure for working on Wikipedia.  Where reputations are at stake is when a person's name is placed on a statement of endorsement, either in the form of a byline or of editorial approval.  Well, there's nothing in the proposed system that requires that editorial approval be attached to any particular person's name (and I can think of a few arguments against doing that); and besides, no one will be twisting editors' arms, asking them to approve of stuff against their better judgment.

[Nicholas Kaye-Smith]

So, what you get is:

{name}

and

{name}/draft

The latter is free for everyboy to play with, and the first is the pne the reader gets to see. If the draft has become better, an editor approves it and we go on.... Something like that?

Yes, although I am suggesting that the draft be the default version and the approved version be at {name}/approved.

[Kim van der Linde]

So, what you get is:

{name}

and

{name}/draft

The latter is free for everyboy to play with, and the first is the pne the reader gets to see. If the draft has become better, an editor approves it and we go on.... Something like that?

Yes, although I am suggesting that the draft be the default version and the approved version be at {name}/approved.

Why? Isn't it more logical to serve the approved version to the readers?

[Larry Sanger]

Kim, you've got an interesting suggestion there.  One problem I see is that some editable articles would live on /draft subpages while others would live on the main page--unless you wanted to upload Wikipedia articles to /draft pages.  But then, you want to fork piecemeal, so this might not be such a problem for you.

An alternative, of course, would be to make the frozen versions on subpages, such as {articlename}/approved.  That would neatly overcome the problem I just set out.

The question clearly comes down to whether CZ is devoted first and foremost to developing content, or to publishing content.  Developing, I say, at least for the first few years.

The feature you're after, anyway, is precisely the feature that was written into the MediaWiki software for the Encyclopedia of Earth.  I do think you should apply there.  See http://www.eoearth.org/  You will like what they're doing rather better than what we're doing.  I can say that definitively because I've seen your objections and I helped design the EoE system.  And they are more publishing-oriented than development-oriented, perhaps.

[Kim van der Linde]

You are wrong, I am very much development oriented. Actually, my ideas on how to use a wiki in combination with expert approval would be more flexible than the current ideas of CZ. I do not believe in a final version of an article; dynamics is what rocks. BUT, I do also believe in quality control, and I am very very sure the two can be combined. In my ideas, anybody can contribute (anonymous, under name, whatever), but contributing is limited to the draft and other open pages. You have to create an account only to get access to those editable pages. Visitors/readers only get access to the approved section.

At the draft pages, the normal community processes as on wikipedia can just go on there. However, the main incentive for vandals, POV-pushers, history revisionists etc is gone, namely that your vandalism/.../...  becomes visible for the world to admire, because anything you want to add has to go along the approval. Any trusted editor gets access to the approved pages, and with trusted, I mean that they know where to go and where to stay away. And these trusted people are roughly to be divided in four groups: experts, copyeditors, translators, and managers (category organisation, template updating whatever is not under the previous two).

I do not believe in importing Wikipedia. There are so many articles that need to be merged, split, deleted, moved, whatever, and there is so much junk and just plain historical revisionism, POV-forks, etc, it just does not work. But if there is a large group of editors, many pages can be worked on very quickly, and we should not be affraid to have holes i our coverage in the beginning. This provides a much better incentive to actually get good pages at the site insteadof letting the imported wikipedia pages untoutched.

So, how are the dynamics. Simple, anybody can start pages as {name}/{language}/draft, either by copying it from wikipedia, or starting from scratch, translate from {name}/{other_language} or whatever. Anybody can contribute. As soon as an expert feels the article is good enough for publication, it gets approved (copy editors and managers do not have that option, while translators can approve work of others). From then on, development continues at the draft page and as soon as a new addition is ready, substantial update, a rewrite to make it more comprehensible, or just the latest news on a topic, it gets approved again, and the cycle continues. With this, I envision also to have current events (expert group journalists).

Will people work on this? I think so. What is nicer if you work on a draft version and your work is appreciated and gets approved. If you produce junk, well, the fun will be gone very quickly, and if you are a vandal, well, all trusted contributors can block those people, but as your vandalism is not becoming visible, why should you. Someone commented that she did not like it because of the approval. Well, I looked at her wikipedia work, and that person would be most likely get expert status anyway. And for me as an expert, if someone is interested in a topic, writes a good article, great, lets do some more.

Who gets trusted status? Effectively anybody who shows the required skills for contributing on the development of an encyclopeadia. Ph.D. can suck in their writing, have no clue how to present it to the public, and some should never ever be trusted. My idea is to have a central board with some open-minded, broadly interested people who can approve experts in new fields. As soon as there are some experts for a topic, those groups of experts can approve other experts within their own group. And anybody can oppose experts at the board level, and the board can take away those privileges (and believe me, that will be needed at times). The board also functions as a content ArbCom.

Expert fields are loosely organised groups, and experts can be part of multiple groups if that is desired. Articles have a primary expert group, but often will belong to several groups. Some aspects of articles (etymology for example) requires a complete different group of experts, and those expert groups are more general covering the whole encyclopaedia, but limit their approval work to the etymology sections of articles.

Ok, enough ramblings for now.... (and this post probably should be moved to anther forum...)

BTW, I have applied to EoE (or was that DU, it was a while back), but never heard anything. Besides that, both are to static to my liking, and way to top-down oriented....

[Larry Sanger]

Kim, let me reply to your proposal point-by-point, although I'll surely be repeating myself.  I'm curious what your responses are.  I'd like actually to have a dialogue and perhaps understand where we might make headway.

In my ideas, anybody can contribute (anonymous, under name, whatever), but contributing is limited to the draft and other open pages.

The reason I oppose anonymous contribution is that real-world identities--whether or not they are known to an application committee--act as a crucial moderating influence, as they do in the real world.  People who are act without taking personal responsibility for their actions quite frequently do, on Wikipedia and Internet forums of various sorts, behave in all sorts of ill-mannered ways that they would not act in if their real-world reputations were on the line.  Many of the content problems on Wikipedia And these trusted people are roughly to be divided in four groups: experts, copyeditors, translators, and managerscan be traced back to behavioral problems, and many of those behavioral problems would not exist in a system in which people knew that their poor behavior was connected to their real-world identities.

At the draft pages, the normal community processes as on wikipedia can just go on there.

We might agree about this, but, just to be sure, I would add that there are parts of the normal WP community processes that we most certainly want to change.  For one thing, the availability of expert editors makes it possible for them to resolve content disputes without people having to come to a consensus or compromise in which the truth, or high quality, actually loses out.  Furthermore, editors will be working shoulder-to-shoulder with ordinary authors on articles.  This will change the dynamic in important ways, and I hope for the better.  But, again, I don't mean to imply that you're disagreeing with any of this.

I do not believe in importing Wikipedia. There are so many articles that need to be merged, split, deleted, moved, whatever, and there is so much junk and just plain historical revisionism, POV-forks, etc, it just does not work.

"It just does not work" just does not follow logically from your premises--which I grant.  Why doesn't it work?  Granted, there are piles of problematic material on Wikipedia.  That does not imply that there are not also piles that can be saved.  The question--which I have asked before--is whether it would be more efficient in the long run to upload all the articles and weed out or clean up the bad ones, or instead to upload stuff piecemeal.  I've argued at length that all-at-once forking will greatly increase the number of people at work on the project, and it will be more efficient for those people as well.  See: https://lists.purdue.edu/pipermail/citizendium-l/2006-September/000483.html

You have not responded to these arguments.  I'm a reasonable person.  I can and do change my mind when confronted with a better argument, and have done so in the past, in public.  So I'm curious to see how you respond to my arguments, and what substantive arguments you have.

And these trusted people are roughly to be divided in four groups: experts, copyeditors, translators, and managers...  So, how are the dynamics. Simple, anybody can start pages as {name}/{language}/draft

The language part I'm not going to comment on, except to say that it is actually a nonstarter to suppose that any one language, English or otherwise, could be the master language for an encyclopedia article.  The French, for one thing, simply would not translate the English language article--and you could leave it at that.  There are many other problems with the proposal.  Another one is that you leave everyone who does not write the language of the "master version" of some article out of the article development process.  They'll clamor for independence, and quite right, too.  I've thought about it many times, since Nupedia days, and I've never found a way around such problems.  Also, it does not logically follow from the mere fact that people can have different versions of the facts in different languages that there must, therefore, be only one master article for each topic, translated into the different languages.  Explain this reasoning more carefully, please, if you are really persuaded by it.

Also, if--contrary to your hopes--we do import many articles for which we do not have "approved" versions, then essentially you'll be asking contributors working on those not-yet-approved articles to click through to the "draft" version whenever they want to work rather than read.  This might sound trivial (just a couple of extra keystrokes and server calls) but I can guarantee that it would make the whole project, to contributors, seem to be moving in quicksand.  The better way, in the end, is to autogenerate a stand-alone website from approved versions that are marked in some other way (either via a template pointing to a version in the history, as I've suggested, or via a new plugin, the ManyOne-authored one or maybe the German one if it's suitable).

In this way, we can have two stand-alone websites: one for article development, and one for readers-only.  Let me grant that we should eventually have a readers-only website containing all and only approved articles.  Is there some reason we should not also have a widely-accessible website for article development?

By the way, one decisive reason against putting "draft" articles in a "draft" area that only signed-in people can view is that it is precisely by seeing demonstrated the fact that the website can be edited that people come to edit the website.  In other words, our single most powerful recruitment tool is an easily viewable article-in-progress and the statement: "You can edit this article now!"

Will people work on this? I think so. What is nicer if you work on a draft version and your work is appreciated and gets approved. If you produce junk, well, the fun will be gone very quickly, and if you are a vandal, well, all trusted contributors can block those people, but as your vandalism is not becoming visible, why should you. Someone commented that she did not like it because of the approval. Well, I looked at her wikipedia work, and that person would be most likely get expert status anyway. And for me as an expert, if someone is interested in a topic, writes a good article, great, lets do some more.

We certainly agree on the merits of approval.  That's always been part of my proposal for CZ.  In fact, the only difference between my proposal and yours, here, is whether the editable (or draft) version appears first, or instead in a subpage, as you have it.  On both proposals, work will be collaborative.  So it won't be a matter of whether your junk is approved.  Your junk will probably be rewritten or removed before there is a chance for an editor to decide whether to move it to the approved version.

I do want to grant that you have at least one good argument here, and that is that vandals would have less incentive, because their vandalism wouldn't be widely visible.  But neither would the article in progress, which again is a great recruitment tool.  You've got to take the bad with the good.  Besides, vandalism will be fairly easy to take care of if we require real names and a working e-mail address.  But, yes, it will exist.  That's OK.  It will exist on pages that are marked as unapproved, in a project labelled as a workspace, that also have definitely unvandalized pages that are approved.  The possibility of very temporary vandalism isn't nearly as bad as it might sound.

Well, this is all I have time to review for now!

Make sure it's the EoE you've applied to.  You definitely should have heard from them.  The DU is a separate group that hasn't been able to get started yet.

[David91]

I see nothing incompatible between development and publishing. You have every article under continuous development but always make sure that the naive user gets to see the approved page first. Let me put my expert hat on for a minute. Law depends on words — those written by legislative acts and those written in a judicial forum — and their detailed interpretation by academics, advocates and judges. The source of law, those written words, have to be carved in stone as the starting point of all legal processes. In many legal systems, academic works are relied on as a part of the process of developing and reforming the laws, and are admissible in evidence in the judicial fora as to what the law is or should be. Statutes and all judicial decisions of significance must be published and made freely accessible to all. The knowledge that the words cannot be changed is fundamental to the credibility of the system. Even though an encyclopaedia editor may simplify the language to make it more accessible to the lay reader, it is absolutely critical that the sense of what is written is as close to the "formal" meaning as it is possible to get it. Anything less than that may completely mislead the naive reader as to what the law allows or proscribes. I also had some expertise in management information systems and risk management which depend upon accuracy of analysis and evaluation, frequently in a mathematical context. The idea that accounting or other source data to be used for this purpose can be fraudulently manipulated, threatens the foundations of the capitalist system. Investors and all other interested parties have to be able to rely on the accuracy of the data and its evaluation. I assume that the same is true in the areas of expertise more formally considered scientific. Hence, wherever possible, encyclopaedia pages relating to all these subjects must be as accurate as possible to have any real credibility. Even in non-scientific contexts, failing to produce an accurate discography or failing to give an accurate listing a characters in an anime series can seriously annoy those interested in those topics. Why should any information be put at risk of unauthorised amendment when it can so easily be protected?

If CZ is to distinguish itself from the anarchy of Wikipedia (and I made several thousand edits of substance over a reasonably long period of time there so I do understand the Wiki systems from actual experience which taught me only to write about non-controversial topics in a technical style to avoid upsetting the local wild life), it has to focus on establishing and growing a quality product. To my mind, anything less than that is a complete waste of effort. If you want to engage the academic world's support, I believe that you have to run development and publishing in parallel, with the publication face turned to the world wherever CZ standards have been met. Adopting such an approach will cumulatively distinguish CZ from the competition and slowly convince the academic world to support us. Hopefully, a snowball will find a slope to run down as more experts join for the joy of being involved in something internationally worthwhile.

I envisage that every editor should make a routine of evaluating proposed changes to the pages within their expertise for a guaranteed number of hours every day. There will be people in different time zones all around the world who can, once expert teams are formed, offer 24/7 coverage of all proposed changes. The movement of material from proposal to approved could be made within minutes in areas where there is little traffic. I see absolutely no problem in this system that cannot be solved with good will on all sides.

I hope that we are only differing in the words used rather than substance. I still like to dream about a perfect world. But if the world is not as perfect as we would like, then we have to make the best set of compromises that will preserve the opportunity to realise the dream.

[Larry Sanger]

Here is something I'm willing to stipulate (I don't mean this to be a complete reply to David91 by the way): if there is an approved version, then that version should be first displayed to people who are not signed in to the system.  But, otherwise, even unsigned-in people should be able to see the unapproved version.

I toy with the idea of autoforwarding pages to the most recent page in the history that was approved...