User talk:Trent Toulouse

From Citizendium
Jump to navigation Jump to search

User talk:Trent Toulouse/Archive

Some ideas for contributions

Hi Trent,

nice to see you here! Since we currently have no direct channel for feedback other than the non-member forum (which is basically not used), I think the commentary provided at RationalWiki is valuable. Indeed, I agree with many of the points of criticism raised there, but my conclusions differ: Rather than abandoning the project or declaring it moribound, I prefer to try to help it get on its own feet and find its own way. In doing so, I keep academic perspectives in mind, and although I was not convinced of the necessity of a Charter at this point, I think the current draft provides (at least theoretically) for a more fertile ground for expertise than pre-Charter policies. Whether the Charter will be adopted, and whether it will make a difference in practice remains to be seen, but even if the whole thing were to come to an end soon, I am confident that the idea of coupling wikis and expertise will live on (possibly even at Wikipedia), while detailed knowledge about Citizendium could probably still help future endeavours in this direction to take off.

If you are interested in contributing to topics other than wikis, perhaps

and related articles may be good places to start.

Cheers, --Daniel Mietchen 19:15, 28 July 2010 (UTC)

Thanks, and general comments

Thanks for the information on Wikimedia loading. I don't think that necessarily rules out shared hosting, but it may call for a "cloud" intended for computationally intensive applications rather than web loads. Obviously, there's no simple answer. We may have to build benchmarks.

We now have a Charter, which obviously is only a beginning. It's fair to say, I think, that many people here have strongly divergent views on the fringe articles. Recently, I likened homeopathy to a litter box: lots of unpleasant things going in, and a continuing need to clean it up. I suspect some people who support the fringe articles like the challenge of arguing, others have a particular view of neutrality in which such things must be present, and, in some cases, single-issue advocates.

Critical mass is indeed a problem. I know that I have difficulty in getting even nonspecialist readers to do copy/flow editing on articles that I think are important topics in subjects ranging from current and historical military affairs, computing, and health sciences. I'd be interested in any ideas you might have to encourage participation. My personal opinion is fringe, and metadiscussions, eat up a lot of resources. Howard C. Berkowitz 05:26, 25 September 2010 (UTC)

Hosting solutions

By shared hosting I meant the more specific example of hosting the site on single machine or machine instance that was being used to support multiple websites, such that one website sucking CPU cycles decreases the available amount for other sites on that machine. That kind of hosting is pretty darn cheap, but really only designed for the smallest of web applications and is just there to support static webpages.

Shared hosting based around instancing is probably the preferred solution, if you go the commercial route, as a dedicated machine is not needed. Virtual private servers are definitely appropriate, I think the equivalence of around 2ghz dual core, 2-3 gigs ram and 80-100 gigs of HD would be about the level you would want to look for. That was what I was able to comfortably run RW on when we used commercial hosting. Our issue is that I could not reliably raise the money every month to pay for it, and I didn't want a bad month of fundraising to mean losing the site. I choose to build my own webserver and host it from my apartment as the "cheap" solution. A one time cost of $400 for the machine, and $30 more a month for the static IP and a bit more bandwidth and we have done okay. Trent Toulouse 17:08, 25 September 2010 (UTC)

Observations on CZ

As for CZ and participation, I could write pages and pages on my thoughts and observations over the last 5 years on online community management, and actually plan to soon enough, but I am willing to share with you my observations ideas about CZ. Critical mass is certainly key. The "lurker phenomenon" where people are able to imagine an army of anonymous, faceless masses reading what they create and learning from it. It is the driving force bedhind user content creation whether blogs, forums or wikis. Content takes work to create and people have to believe that others are reading it and paying attention to it. The bigger the perceived size of the army of lurkers the more obstacles and adversity someone is willing to put up with to change or add content. That is why WP works, despite having to put up with a lot from many different angles, the hugeness of the lurker army is such that people feel compelled to fight through the BS.

Frankly, the perceived size of CZ's readership is not very big. That means that people are not really willing to put up with many obstacles. And you have two big obstacles that are going to be very difficult to deal with: barriers to entry, and a poorly defined niche. First are barriers to entry, which exists in two forms. The obvious one being account creation and real name policy. I have watched RW and a lot of new users start out as merely readers, then might make a small correction or comment as an IP, that goes okay, so they make a few more, eventually deciding to sign up for an account. Or maybe some random person comes to your article sees something they want to change or add and can just do it. Most of the time they go away never to heard from again, but the actual act of editing and changing the site sticks with a few and they come back as a user a month down the road. A new online community is a scary place no matter how familiar you are with online communities, people like to test the waters, and enter slowly when moving from lurker to participant. Signing up with your real name through a vetted registration with a personal biography posted on your user page is not a way to "ease" into the CZ community. The next major barrier to entry is the wiki software itself, this is your catch 22, the people most savvy with wiki software, most comfortable with and most able to jump right in and start doing stuff are likely to be the most overtly antagonistic to your real name policy and expert guidance. I grew up with the hacker community of the late 80s and 90s who are sort of wiki's natural demographic, and there are still a lot of those early values instilled from that time. Things like "information wants to be free" and the power of a pseudonym, not being anonymous, as your pseudonym's reputation was often as valuable as your real life reputation, but it was about building an online identity separate from your real one. CZ's policies have shifted the natural demographic away from the 20-30 something computer nerd, to a demographic that is older, and more in the liberal arts domain of academia. These are the people most inclined to see the value in your policies and want to contribute to the project. Unfortunately, these are the people that are also more likely to find the wiki software, mark up, and the arcane nature of all the little things like indents, templates and signatures to be confusing and intimidating.

Beyond the barriers to entry in CZ itself you have another problem related to niche ecology. You haven't really answered the question of how a single users experience here is different from WP. With WP being the big dog on the block, a user with limited time and motivation has to decide why they should edit here and not at the place with the millions and millions of readers, and the awesome Google ranks. You have an answer in the most general and abstract of sense, the real name policy, expert guidance, etc. These are legitimate differences but the thing is they are not differences for the average user who just shows up to write an article. Their experience is really pretty much identical to what it would be in WP. The only difference is that they have to go through the above barriers of entry, and are subject to "guidance" from a faceless mass of experts who might descend on them at any time. I know that's not how it works but that's the perception of someone who is not closely familiar with CZ policy. Expert guidance is frankly, intimidating in its own right. I can tell you about my early thought process when I signed up for CZ in 2006 and made my first article. I was not an expert, while I knew about it I didn't have confidence in my expertise, I didn't want to put up what little I knew only to embarrass myself when the real expert showed up, and that embarrassment was attached to my real name. So right or wrong the only thing different, the only niche CZ offers compared to its competition is actually scary and intimidating to new users.

Going back to my first point about the perceived sized of the army of lurkers, CZs lurker size is perceived as small so people are not willing to put up with very much to edit. You have policies and branding in place to actually discourage editing. If millions of people were reading CZ daily those policies would probably work fine as people would be willing to put up with them because what they write here has a large perceived effect. Without that you wind up with low participation.

To increase participation I personally think you need to do two things. The first is the most important, and that's find a way to make editing CZ a manifestly different experience than editing WP for your average user. At RW we do this by allowing POV, original work, encouraging community growth and discussions that go beyond building the project, and don't have the notability requirements of WP. Basically you can come to RW and write something that you would never be allowed to write at WP. That makes us something totally different, and when someone is thinking about spending some time writing for a wiki we offer them something WP can't. You need to do the something similar. As a general encyclopedia though that is tougher. I could make several suggestions, such as drastically decreasing notability requirements, allowing certain forms of original research, and here is a big one, allow for article POV. Do what WP never had the guts to do and embrace the "scientific point of view" for your articles. That does two things with one stroke, it gives you a niche WP does not have and a massive potential user base of people that want to write general purpose encyclopedia articles but hate the NPOV of WP as it relates to fringe ideas. It also cuts short the huge amount of time wasting happening on those articles right now. You could also go the other way and decide that your "fringe friendly", that is an even larger audience, if you decide your altie friendly you could have an army of people over here editing all sorts of stuff. Though I include it merely as an example and would argue very much against that approach but that is the idea.

The second thing is to cut back drastically on the barriers of entry. Allow for limited editing by pseudonymous, or even anonymous editors. Use flagged revisions that established named users approve, or you all ready have the behinds the seen draft vs. the actual article. Allow open editing on the draft, something.

Summary: solidify and express a niche that differentiates the average users experience with CZ from WP, drastically cut back on account creation and editing restrictions. This will encourage growth, and growth is a feedback loop that can encourage more growth. Trent Toulouse 17:08, 25 September 2010 (UTC)

Hi Trent, I hope you don't mind my listening in, but Howard posted this link on the Talk:Homeopathy page. I just wanted you to know that I agree 100% with what you've said here and hope that you will help us make those choices over the next few months/years. I do agree that freeing up the presentation of original type research/synthesis is part of the answer and I think we've left that possibility available in the new charter, giving the authority to the Editorial Council to work out the details. Unfortunately, there is a fine line between allowing cutting edge research and synthesis by experts and advocation of crank ideas. I think we are all getting better at identifying the difference, but allowing one without the other is hard to put into practice. I still think that the idea of expert review and approval is part of the answer, but one missed possibility is that experts usually have motives themselves and ownership becomes an issue. These are all things that need discussion and rational debate to iron out. Please stay involved. D. Matt Innis 21:36, 25 September 2010 (UTC)
Trent, a great deal of wisdom here. As Matt points out, we have a serious opportunity to address them. I don't think experts always have motives, but, as they say in arms control, "trust, but verify". I think we are increasingly open to saying "XXX is the predominant scientific view (or YYY and ZZZ) if there are multiple schools, but AAA [not at length] believe the tooth fairy is responsible."
Critical mass is a major challenge. For example, I think intelligence interrogation]] gives some examples of proper synthesis, but the article is far too long and needs to be split into smaller, more accessible articles. In that article, however, what to me is a good example of synthesis puts three things in side-by-side context: the relevant language of the UN Convention against Torture, that which was ratified by the U.S. Senate, and the internal doctrine of the Administration. A different kind of synthesis was Vietnam wars, showing a much broader context than what was on American television. I did a better job splitting up Vietnam wars than with the interrogation policy, although the latter still has lots of subarticles and higher-level articles; see intelligence interrogation/Related Articles]]
I suppose I get along with the early MIT-style Hacker Ethic, but I do have problems with some of the later "information wants to be free", the challenge being balancing privacy and context. I will certainly admit to hacking my way into machines for my own knowledge, but never publishing, and, for other reasons, I don't hack without permission any longer.
Newspapers have been requiring signatures, on letters to the editor, for a long time. I, and I think others here, don't trust anonymous commentary much more than the talk of most politicians. Asking rhetorically, where is the balance? Perhaps pseudonyms with digital signatures?
I am personally appalled by homeopathy and have direct knowledge of people who won't participate in Citizendium while it goes on in its present form.
As far as learning curve, while I don't have a specific formula, I'd do much as does the IETF and other professional lists to which I belong: self-registration with verification of email, but a karma system to protect against immediate vandalism. On some of the software mechanics, there's recently been a very good entry form for metadata, developed by Chris Key, such that I use rather than the manual methods I fully understand because it's a lot faster.
Expert guidance is a challenge. Personally, I've usually had mentors, and seek guidance. Perhaps the issue is the way it is presented. Howard C. Berkowitz 00:28, 26 September 2010 (UTC)
CZ was created so that experts can present their POV as well (everybody's is welcome). If one looks at other wikis, it seems anyone can write anything about anything. Take the article on Homeopathy on most other wikis - they are all attack pieces (despite the fact that there is evidence that homeopathy works). If one takes the article on CZ on 'rationalwiki', we can see there are false allegations against Matt and Dana Ullman. I suggest other wikis also give some weightage to what the experts have to say. If that is not possible, there should be a fork for every article, where all the criticism can be bundled. An encyclopedia article can't be an attack piece.—Ramanand Jhingade 14:30, 26 September 2010 (UTC)
Regrettably, the above involves a lack of broad CZ knowledge, which might be related to working only on homeopathy and closely related articles. As a minor point, CZ actively discourages the use of POV and other Wikipedia abbreviations. Second, CZ, as I understand it, as one of the most prolific author in multiple areas, CZ was never created to let experts present their views without question.
With the new Charter, and then the new Editorial Council, quite a few things will be reviewed. Previously, the Editorial Council was ineffective for a variety of structural reasons, which no longer apply. The Workgroup structure in general, not merely in this area (e.g., Engineering is simply too broad), will certainly be reexamined.
A special problem applies here: that the Healing Arts Workgroup became a complementary and alternative medicine (CAM) workgroup, so its experts claim equivalence to Health Sciences Workgroup experts. I'll preface my observation with the comment of Marcia Angell, the former editor of the New England Journal of Medicine: "there are two kinds of medicine: medicine that works and medicine that doesn't work." Some techniques once considered CAM are now mainstream, while previous mainstream techniques have been abandoned.
We have a Religion Workgroup that has atheism within its scope; we don't have an Atheism Workgroup. We have a Military Workgroup that contains just war theory and critiques of wars that were started; we don't have a Pacifism Workgroup. Alchemy is a historical topic within Chemistry. The Biology Workgroup accepts evolution, a broader term than Darwinism, and there is no Creationism Workgroup.
Homeopathy has been the most persistent fringe article. As an Engineering Editor, for example, I went through some of the assumptions in Apollo Moon landing hoax theory. Many of the claims failed with the application of knowledge of basic science, and on which I simply issued an Editor Ruling. In a few other cases, rather than go point-by-point through refuting, I wrote a new article on how a technique actually does work (e.g., electro-optical tracking). There were no Hoax Workgroup Editors to assert an alternative expert view from a position of authority. Homeopathy, however, is a plague on CZ when the homeopaths consider their field works and any criticism is skepticism or attack -- and they fall back on the experts of Healing Arts.
Unfortunately, the current guidance on professionalism prevents me from questioning qualifications, conflict of interest, and objectivity of individuals. I suspect this might eventually be a formal process under the Editorial Council, with appropriate review. I'd note that several "instant" Editors were later banned for behavior, others dropped out when asked to follow CZ conventions, and many have never contributed other than on their user pages. In informal terms, I believe a CZ Editor needs to "pay dues" before being granted full authority. Howard C. Berkowitz 18:55, 26 September 2010 (UTC)
"Take the article on Homeopathy on most other wikis - they are all attack pieces", Ramanand? Try Wikinfo. Its policy is that the main article on homeopathy should be sympathetic, with a prominent link to a criticism article. That's the way it resolves all disputes. Everyone just presents their own points of view in separate articles, with links to each other at the top. That may seem a strage way of resolving disputes, but at least it works. I cam ehere from Wikipedia because they haven't got a proper dispute resolution procedure. After a couple of months I found there isn't one here either except in areas with a resident editor, so I moved on to Wikinfo. Now CZ is in the throes of constitutional reforms that I hope will resolve this, as Larry urged in his farewell message. If so, I can resume activity here, though without giving up on WI. Such a procedure would be a reason for someone to work here rather than Wikipedia. I think it's also important to break up WP's monopoly, which has predictably led to complacency. Peter Jackson 17:21, 29 September 2010 (UTC)
Interesting thoughts, Peter. Clearly, there is no single approach to dealing with, especially, fringe articles, but each wiki must have a clear one if it has any hope to be useful and event to survive. Yes, we are trying to develop one, and while not every EC candidate addresses the issue, enough do that it should be clear that homeopathy, as representative of a class of controversial articles with essentially irreconcilable sides, is an existential problem for CZ. I like to think of it as a triage issue, triage being the real-world medical problem of dealing with mass casualties and insufficient resources. Perhaps, in some future time of rich resources, there might be revised methods, but if one marginal article is sucking up more resources and generating more anger than anything else, it's not helpful to the project. Howard C. Berkowitz 17:44, 29 September 2010 (UTC)
It is only sucking up resources AT THE MOMENT because there is no mechanism in place to remove the non-scientific element from contributing to it interminably. Once the new Councils are in place and, hopefully, functioning, we will have a method of dealing with articles like this. And the non-scientific element will go away and the article will be rewritten to more exacting standards. And will STAY like that. Hayford Peirce 18:15, 29 September 2010 (UTC)

Response to RW question

Trent, you've given information here, and I'll try to answer some apparently serious questions (compendium and contextualization) you raised at RW. I hope this can be taken constructively, and not just go into snark there. I'm speaking for myself here, not the Editorial Council.

Compendium: I'm not sure why it was ever used, and it's an artifact from the beginning (I came in 2008). The best I can say is someone thought it alliterates with Citizendium.

Contextualization: This is a general approach taken by some people here for a while, with various implementation techniques, and I hope much more in the future. Think of it as the antithesis of "orphan articles" and "walled gardens": an emphasis on multiple kinds of linking, even to discover non-obvious relationships. The Related Articles page was the first attempt to structure things beyond "see also".

Until more sophisticated software techniques, such as Semantic Web, are available, I sometimes think of CZ more as a interconnected system of Related Articles pages, which then link the articles. Not all think the same way, and it wasn't an idea in the beginning. You mentioned CZ finding a niche, and that's one.

Another aspect of contextualization is building high-level articles and then zooming into detail. My own efforts in this produced Vietnam wars and its many related articles; the original Vietnam War article, now superceded, considered it an American war, perhaps with Vietnamese extras.

Lemma articles were extensions of the idea of Related Articles, especially when they have their own related articles pages. The content of the lemma article itself isn't always as important as its links, usually from the Related Articles page. For various reasons, we don't have enough automated tools, but that's a way to go.

Sometimes a lemma is no more than a way to put a definition into Related Articles pages, but they become much more powerful when they have subordinate Related Articles. When a knowledgeable person constructs a Related Articles page mapping the context of a subject, they may not worry if the terms they introduce are redlinks, or lemmas to help someone read the page.

"Social network" is ambiguous, but think what happens if you load a variety of brief articles on politicians and political organizations, and then explore how they link to one another: commitees, contributors, lobbies, interests. Not infrequently, using "what links here" can give some interesting and unexpected relationships that can be built into increasing content beyond lemma stage. Right now, this is essentially manual, but I definitely can see loading membership lists, etc., into databases, from which even lemmas can be generated, and additional list information added -- perhaps to the Related Articles page..

There are a lot of discussions, not always easily found from the outside -- less in the Forums than CZ space. and a certain amount of manual experimentation. Related Articles needs to have some additional constructs, some of which were blocked in the past. I plan to introduce ideas on this into the Editorial Council, as I experiment myself. Yes, there's sniping on some of this, but quite a few of us want to change that.

Let this be a start of discussion. Howard C. Berkowitz 06:33, 8 November 2010 (UTC)

I think I am still a little fuzzy on the implementation but better understand the concept. Question: for truly novel/interesting relationships to fall out of the interconnection of articles doesn't that interconnection have to be pretty much automated? If it is up to individuals to piece together the connections then aren't the connections always something that was a priori entered into the system? Trent Toulouse 17:57, 9 November 2010 (UTC)
Yes and no. Automated systems aren't smart enough to decide, but are good to suggest - so the "What else links here" cue on Related Articles is good for suggesting things that might be added there that you hadn't thought of. For instance, you work on dopamine, you might or might not have anticipated the connection with Food Reward which should come up on the "what links here" prompt. You'd need to decide whether that was a link worth adding depending on the nature of your article on dopamine. Gareth Leng 18:13, 9 November 2010 (UTC)
I mean I can see that one or two steps down the chain, at most. But where Howard's idealization started to sound interesting was the point where every there was a vast network of interconnection that could produce completely novel relationships as an emergent property of that network. So Person A and D can be seen to be related because A worked for Company 1 which funded Foundation 2 who hired Person B to lobby to Person C whose chief of staff is Person D. Or some other complex chain. If that chain relies on actual human connections at each point in the chain then it seems like there is no emergent property, as the connections were all known before the infrastructure was put into place.
With the dopamine example, someone had to put the link in food reward, whether I personally knew it or not, it was known a priori before the start of the interlinking. Trent Toulouse 18:24, 9 November 2010 (UTC)
Lemma/article, dumps into Related Articles, manual correlation. You are absolutely right that the current system requires a lot of manual input. I'm not sure, however, that the relationships are known a priori. There are a fair number of political articles that have only a stub as definition, or may only be a lemma, but have a Related Articles page into which I entered boards of directors, member lists, etc. This is all painfully manual. I'll note generally that there are lots of tools and applications for this in the intelligence community.
Let's take Congressman X and Congressman Y, and assume they are both members of the Congressional Ukrainian Caucus. True, the relationship of X->CUC and Y->CUC both exist as elementary relations, but now let's assume that some topical article on a political fight between X and Y takes place, and they accuse one another of not being willing to work together. When I later loaded a board of directors of the Heritage Foundation and of the International Crisis Forum, and both again are on it, using "what links here", and manually merging the memberships at this time, suggests they can work together in various forums and the public article may mean posturing.
The relationships may not be known, but aren't necessarily all in one place. I've been shocked by finding, for example, that someone I respected turned out, once I put in the guest list for a hate show, to have been in it. Now, I'm certainly not going to condemn them with that knowledge alone, because they may have been on that show to argue its premise. Still, it's an opportunity for further research.
Contextualizing articles definitely need to be manual and can be hard to write. With respect to Vietnam, it can't really understood unless you know they've had two millenia of wars -- as it is, Vietnam wars really deals with the effects of the first Western involvement, 1860-ish, or perhaps going back to the Nguyen Dynasty article that takes you to the first missionaries. Rightly or wrongly, Dai Viet, a more medieval period, is linked but not in the main hierarchy.
I went top-down from Interrogation, to national-level articles on Intelligence interrogation, U.S., to Intelligence interrogation]], and down even further into specific interrogation techniques, prisoners, etc. The level of prisoner might well be in the extrajudicial detention hierarchy as well as the interrogation hierarchy, and you should be able to trace this through Related Articles. Related Articles are a sort of index to help browse, as well as something to assist in defining articles. With existing articles, I still periodically do a "what links here" to find more things to put on the Related Articles page. Torture, for example, is related to interrogation but not always part of it, and torture isn't always for interrogation.
"Hierarchy" is an awkward term. Larry, as far as I know, invented the strictly hierarchical Related Articles page, but later objected to others changing it to include relationships that aren't strictly hierarchical. I hope to change that, on the way to semantic web tools. Howard C. Berkowitz 00:19, 10 November 2010 (UTC)
Retrieved from "https://citizendium.org/wiki/index.php?title=User_talk:Trent_Toulouse&oldid=944664"