Talk:Scholarpedia
Please, put your thoughts and suggestions here.
- Eugene.
Contents |
Nice
I just arrived here at this site, and I can see that this might eventually produce a real encyclopedia, unlike the random collection of facts and nonsense that can be found at wikipedia. I will have a look at some articles I know more about and see what I think..... KimvdLinde 20:20, 12 September 2006 (CDT)
Clubby
It looks to me as a very good initiative, but totally locked and clubby.
The possibility to invite "people" to become "curators" or else is clearly promoting friends without needed justification and is totally undermining the credibility of the scholarpedia project as a community project where people are elected or gaining scores according to their qualitative and quantitative participation in the project.
Editor's response: Getting invited is only one of many ways to become a curator - the quickest one. Getting elected is another one. Taking responsibility of an article after its curator retires is another way to become a curator. Finally, getting high score is yet another way.
Thanks for the clarification and for answering, i appreciate it. Anyhow, having different ways, the easy and quick, and the hard and slow doesn't look really democratic to me.
---
The idea to get 0.1 credit when one "Vote for the candidate who eventually accepts the responsibility to write the article, i.e., becomes its author." is totally insane : it clearly pushes people to vote for the one that has more chance to be elected than for people that they believe being the best to write the article.
Editor's response: This statement relies on the assumption that the one who is elected accepts the challenge. Each candidate has a "brief note" section where his/her writing skills and qualifications should be mentioned. In addition, there is a discussion page for every article where people scholars should discuss cons and pros of every candidate form point of view how good his/her article could be.
Yes, and in these descriptions you can already feel who has more chance to be elected than the other, the bias can be often easily unmasked (comments like, "already accepted to write the review if elected" for example !)
---
To have a vote multiplied by a credit that goes from 0 to hundreds is totally outrageous.
--
Editor's response: Many experts have Scholar Index over 10 now because they were involved in the review of articles. Thus, the election results are more biased toward the expert's vote. (experts elect experts). Finally, the vote is multiplied by an index that is not bound by hundreds; it could be thousands for some people 20 years from now. Yes, I agree that this is outrageous.
I am still not convinced that this recurrent excitatory network is the most suitable way of creating an interacting dynamic encyclopedia. It really looks to me as having a huge inertia (some inertia is good, but clearly not too much), where known people get to be more known. Anyhow, it is easy to criticize, i admit, even more when the system is well built like this one.
privacy
Writing an encyclopedia involves lots of responsibility (see the position of wikipedia in China). With the risk of being grumpy, let's get a bit paranoid:
- How are the rules of privacy protected? The administrator of the server may know all edits / logs / page views with IP...
Editor's response: The privacy and anonymity of reviewers is protected the same way as in printed journals with electronic submission/reviewing. Editor-in-chief, as well as server administrators, know the password and have access to the database. Editor-in-chief invites other experts to review articles and enforces the anonymity.
- Who can look at all comments and so? Wikipedia is community wide, scholarpedia is based on one editor...
Editor's response: All the comments are available to the public. They are not hidden. Only identities of reviewers are hidden, unless reviewers reveal themselves.
- How to be sure that editing were not suppressed (I mean from the history of edits)?
Editor's response: Yes, edits can be manually deleted from the database. This is a drastic measure aimed at removing offensive or illegal material. It is not going to be used to resolve scientific conflicts, if any.
- in Special:ElectionForum it is written "A soft-max (probabilistic) procedure will be used to choose the author based on the number of votes received." Why not majority simply? Who writes the rules? Is it automatic?
Editor's response: If the author is decided by the simple majority of votes, then everybody would know that one person got more votes than the other candidates. This would create an unhealthy atmosphere of winners and losers and may deter potential experts from participating in elections. In contrast, a soft-max procedure erases this problem. Indeed, even a person who got only one vote has a chance (though extremely small) to be elected. Thus, election of one candidate does not imply that the other candidates had fewer votes. This rules were invented by the editor-in-chief. Scholarpedia will have this procedure automated.
Building momentum
This wiki is a great idea. However the project doesn't seem to be moving very quickly. If one clicks around, all one finds are empty pages reserved for curators-to-be. How do we build some momentum behind this valuable project?
Here are some possible problems along with proposed solutions.
Possible problem: Reliance on high-profile curators. The reserved pages seem primarily to have high-profile curators assigned to them (people well-known for being an expert in a given area). The problem is that these folks may be too busy to write wiki articles. And who wants to invest time in writing for a website that doesn't have users yet? Also, these folks may not want their writing edited by others in the wiki fashion. Some solutions: Do an initial pass of article writing by post-docs. These folks may be more motivated to write about what they are just learning, have a fresh position-neutral perspective, and have enthusiasm for the topic. Let junior contributors to a page get credit for contributing, which gives them visibility in the field. Perhaps the lead role should be "editor" rather than "curator". Also, consider assigning multiple people to this role on a given page. One name can be a celebrity that doesn't need to do anything except proof the article periodically. The other can be a post-doc who does the writing and gets recognized for being a co-author with the celebrity researcher.
Editor's response: It is true that most of the authors are the best experts in their respective fields, and hence are quite busy. Some of the people were invited by other curators, but many were invited by me. In my invitation letter I explicitly mentioned that they could have junior co-authors who would take care of the article on a long run. What could be a better incentive for a post-doc or a junior faculty than the possibility to co-author an article with the best living expert? Many authors replied that they would write the articles themselves.
Possible problem: No content and no users. Who wants to write for a website that no one is reading? Who wants to read a website with no content? The question is how to get over this chicken-and-egg problem. Some solutions: Cut-and-paste articles from wikipedia. Wikipedia in neuroscience has gotten quite good (technical, detailed, and accurate) in the last year. It's writing style is quite good as well (favoring clarity and simplicity in writing). So to get the ball rolling, why not cut-and-paste wikipedia pages into scholarpedia? A special warning can be placed at the top saying that the page is a place-holder page, has not been reviewed, and is not citable. Having a basic bage that is already written might inspire the would-be curators/editors to revise and clean up rather than write from scratch.
Editor's response: Borrowing from Wikipedia would lead to another "Citizendum" project. Most contributors to Scholarpedia typically have their material written for other journals multiple times, so they only need to edit their existing review articles and tailor them for the Scholarpedia format. Besides, these scientists will not be willing to put their name as an author for an article written by a group of anonymous people, no matter how good the article is. My goal is to create something that withstands time through the process of curatorship. If an article, say on fractals, is written by Benoit Mandelbrot, it has a higher chance to be maintained by future experts even (and especially) 50 years from now than the same article written by, say, me.
Possible problem: Empty pages are too daunting to start. Some solutions: Promote the idea of writing stubs. Rather than asking an expert to write everything they know (who wants to even attempt that?), have informed early-career people write simple stubs for credit. The first version of a page doesn't need to be the last word. It's enough if it says something. The 3 - 6 sentence introductions that are used on scientific posters at conferences might be a good framework for where to start on a topic.
Editor's response: This is the Wikipedia style. It works great and I have no objection for that. However, the stubs most probably be removed once an expert takes care of the article. Since the expert will probably write the whole article from the scratch (recycling his own review articles) rather than improve the stub, stubbing is a temporal measure.
Possible problem: The site seems dead. There is no evidence of any activity. This makes the project seem abandoned, which makes writing for it seem like a waste of time. Some solutions: Do what wikipedia does and have a news page that appears as the main page. Get a few major section editors to update the main page with news of recently written articles, featured (high quality) articles, and featured (prolific) editors. This will humanize the site, making it more interesting to visit, more of a community, and more motivating to write for.
Editor's response: This is a good suggestion. Editor-in-chief has a regular correspondence with authors, participates in the peer-review process and helps to convert MS-word and LaTeX documents to wikitext (in addition to programming php code that runs the website). This, of course, is not visible on the website, giving an erroneous impression to non-participants that the site is abandoned. Changing the front page may be a good idea.
The activity of the site can be seen in 'recent changes' link. On a typical day, there are a few experts working on their articles at the same time. Because Scholarpedia hosts only 3 focused encyclopedias, there is no much traffic.
Followup response: Having just subscribed to the email updates, I see that more is happening than I thought. The 'recent changes' link is not very useful (it's too detailed and looks like machine output). However using the Main page to highlight new articles in the last month by topic area would make the site a fun place to check in on. Once I got my email update, I immediately read those articles and read about the authors. It would be great to see highlights on the main page with links to a detailed News Page. If there were a news page each month, once could read "past issues", almost like a journal. Right now, it is difficult to find the new articles because there are so many empty stubs.
Possible problem: How to retire non-expert authors. If one starts with lesser known authors, how does one "upgrade" when more prestigious experts start taking an interest in being involved? Some solutions: Encourage frequently rotating editorial teams. When someone is rotated off, add them to a list of Editors Emeritus at the bottom of the page. The rotated-out editors get immortalized with permanent credit for their contribution, and then they don't have to edit a wiki for the rest of their life.
Editor's response: There is a built-in mechanism of rotation of curators of Scholarpedia: Every time an article is modified, the curator approves or rejects the modification and evaluates its usefulness. When the curator decides to quit the curatorship (by pressing quit button) or just does not take care of submitted modifications for a long period of time, Scholarpedia looks at the archive and finds the person who submitted most useful modifications for that article (according to the judgment of the current curator) and offers curatorship to this person. The new curator evaluates all new submissions in due time and evaluates their usefulness. When he retires, the curatorship is offered to the next contributor, and so on. The idea is that if the original articles are written and curated by the best living experts, they will be maintained by the future best experts. Yes, it may take months or even years before the project is up and running even in the three focused areas. However, I would prefer the project to move at half the speed if it means twice the quality.
Followup response: With articles trickling in (as I see now), the issues previously mentioned may not be as great as previously thought. The key at this point is to make the activity more visible, which will motivate everyone and build excitement. At the moment, the inactive stubs overwhelm and dilute the evidence of activity. This could be addressed with some simple organizational changes as previously mentioned.
Editor's response: Thank you for suggestion. I will make "News" or similar link where I will post copies of the newsletters that I sent regularly to the registered participants of Scholarpedia.
Will standard style conventions ever be adhered to here as much as on Wikipedia?
Very interesting project.
Scholarpedia seems to have less stylistic consistency than Wikipedia. Two style issues that that leap out right away are (1) Lack of consistent style in regard to capitalization in article titles and headings. "See also" or "See Also"? Is there any standard? (2) Lack of consistent bolding of the title phrase at its first appearance.
Also, is there any standard as to singulars versus plurals in article headings? On Wikipedia you can't use a plural article title unless you state cogent reasons for making an exception (e.g. The Beatles); if you do then someone moves the article and cites the style manual and the community backs them up; the plural usually exists as a redirect page.
(I'm not too surprised that there's less rigid adherence to standards of that kind here than on Wikipedia. If you think about it, you see rather obvious reasons why that would happen.) Michael Hardy 20:22, 15 April 2007 (EDT)
.... OK, I just looked at the "Review Forum" and I found this:
- Mathematical Biology
- Mathematical logic
- Network Dynamics
- Neural networks
Capital "B" in "Biology", lower-case "l" in "logic", capital "D" in "dynamics", lower-case "n" in "networks". Michael Hardy 20:31, 15 April 2007 (EDT)
....and now I've done a lot more looking around, and it seems (as is also to be expected under the circumstances) the authors here are generally less familiar with their medium and how to use it most advantageously than are those at Wikipedia. I'll say more on this later, but for now, another issue---see below. Michael Hardy 22:51, 15 April 2007 (EDT)
Editor's response: Michael, you are right; different articles do have different styles, simply because many authors do not read the "instructions for authors", and some are not even familiar with wikitext format. I try to correct as much as possible, but I inevitably miss a lot. Eugene.
Initial context-setting
One article begins thus:
- An onset from regular to chaotic regimes through the break-down of a two-dimensional invariant torus has been observed in many natural and man-made systems including[...]
I've noticed a lot of the articles here are similarly deficient in initial context-setting. Unless I've misunderstood something, Scholarpedia aims to eventually include a very broad range of subjects. Once upon a time I came across a Wikipedia article titled schismatic temperament. Being familiar with the usual meanings of each of those two words---separately---I thought maybe it was an article about a psychiatric disorder. I read all the way through the first sentence without learning otherwise! The first sentence was written in a way that was perfectly clear to those familiar with the field (music theory), but failed to tell others just which field that was. I edited it so that instead of saying
- Schismatic temperament is.....
it said
- In music, schismatic temperament is ...
That's all it takes. Michael Hardy 22:56, 15 April 2007 (EDT)
Comments on copyright and patents
Some of the ideas behind this website are very exciting. For example, the notion of voting for the currently living foremost expert in a given field, and then using academic peer pressure to get that person to "curate" an article is impressive. That you would try to patent this methodology rather than share your modifications of MediaWiki as plugins that anyone can freely use disturbs me greatly. If the MediaWiki software had been put under a closed license in the first place, and not released under the GNU Public License, you would not have been able to create Scholarpedia as it stands. The argument that all text is under a restrictive copyright because "otherwise no publisher will print the three encyclopedias" is spurious. The GNU Free Document License assigns copyright to the person who authored the text. Your publisher will not have a difficult time with the copyright status of the text, but rather the fact that all of your articles were written by multiple authors whose identities cannot ultimately be verified. I was able to create an account on your website, and although I am who I say I am, you have no guarantee of this. See the Essjay controversy for more details.
My point is not that you shouldn't try to publish it as a book - I think it's a great idea. It's that any publisher would be absolutely nuts to NOT publish the book, no matter what the license. As it is you are giving nothing back to the community that has given you so much. Why can't anyone download an XML dump of all the articles in Scholarpedia? If you use the GFDL, they would be required by law to explicitly reference this website as the source.
I would ask that you seriously consider making your modifications available and giving authors the option of what copyright to assign to their text. The notion of having a curator vet seemingly anonymous contributions, so that they are judged on their merit and rather than ad hominem could perhaps be modified such that a community of curators votes on those bits of text instead, and used in Wikipedia. --Mingus 15:39, 29 April 2007 (EDT)
Editor's reponse: 'Dear Brian Mingus. Thank you for your thoughts and suggestions. Currently, Scholarpedia investigates the possibility of having creative commons licence. The MIT Press and Springer expressed interest in publishing focused encyclopedias generated by Scholarpedia, so we need to discuss relicensing details with these publishers before finalizing Scholarpedia license/copyright.
Authors of Scholarpedia do not relinquish the copyright. They remain copyright to their material (see the two-liner below the 'save' button at the bottom of the edit page). Thus, Scholarpedia is not an exclusive copyright holder.
I do not think it is a good idea of providing the capability of mirroring or dumping the database of Scholarpedia. The authors agree to contribute to this project under the condition that they become curators and have control over the content of their entries. Once their entries are outside of Scholarpedia, they cannot curate them. In addition, the database contains the information on who the anonymous reviewers are, so making the database public will reveal the reviewers.
