User login

Navigation

You are here

How Should Journal Club Work at iMechanica?

This has been a topic that the editors of Journal Club have been discussing for some time now.  We would like some feedback from the entire iMechanica community.

iMechanica's Journal Club was originally proposed by Pradeep Sharma to be an online analog to a standard "Journal Club" that many research groups hold.  The basic idea is simple enough: a set of articles are identified by the group, everyone reads them, and then there's a discussion. In theory, there are many advantages to mimicking this process online and specifically at iMechanica.  These include experts leading the discussion, the discussion being saved online so that anyone can see it, and the use of different types of media.

In practice this has worked reasonably well for some of the issues of Journal Club, and perhaps less well in others. As one of the current editors, I would like to hear your thoughts.  Please, any and all ideas are welcome.  How should discussion topics be selected?  How about discussion leaders?  Is once a month too frequent?  Should we consider topics that are "controversial"? 

My thanks in advance.  

Comments

I read iMechanica via Google Reader.   The feed is usually cluttered with spam comments which don't disappear even when the comments are deleted.  I don't have the patience to go through the feed and filter out valid comments from junk.   Fixing that issue is the first step that needs to be taken.

Once a month seems to be the right interval for journal club articles.  I can take my time and read the articles (if I can access them, which isn't easy or cheap).   

Topics that are of current academic interest are far removed from what people in industry are working on.   Most journal club topics are only of academic interest at this time in history.  So students of engineering will find it hard to relate to these topics.   Also, journal club topics deal with quite advanced subjects.  

We need a mechanics expert to teach the us how any journal club article is relevant to the real world and how it is related to what we already know.  It may be safely assumed that we don't know much cell biology, quantum field theory, differential geometry and so on.

Given that we don't know much, a journal club author and editor needs to spend a significant amount of time explaining the basic concepts in a series of comments following the article.  That way the article can be read by people who are reasonably familiar with the field without unnecessary details.  And novices like me will also be able to understand what's going on.

-- Biswajit 

I'm not familiar with the issues concerning Google reader, but hopefully that is something that can be addressed.

I'm actually a little surprised at your implication that students of engineering will find it hard to relate to academic topics.  Is this because you believe many students are coming from industry?  

It seems to me that what you would like to see is perhaps to have some of the advanced subjects broken down into much simpler sub-topics, with all the details explained as clearly as possible.  Is that a fair description?  I think we can help to effect that.  Do you think this would improve participation?

John,

1) " Is this because you believe many students are coming from industry?  "

I have been thinking about the questions that are asked most frequently on iMechanica.   Most questions are on standard enegineering topics and tools.  Right now, the column on the left of the page shows the following topics of interest: beams, fracture toughness, applying pressures, contour integrals, cracks, creep, drawing parts, Abaqus and Matlab, contact, Timoshenko theory, temperature differentials, energy release rate, fixing the center of a gear, and friction stir welding. 

Are these questions being raised by students? Academics? People in industry?  I'd guess fom the questions that these are mostly students.  Are they undergraduate or graduate students?  My guess is that the tilt is more towards graduate students.  

2) "It seems to me that what you would like to see is perhaps to have some
of the advanced subjects broken down into much simpler sub-topics. .. Do you think this would improve participation?"

The level of participation reflects the interests of the self-selected group that inhabits iMechanica.  A lack of participation is a proxy vote; not on the worth of the topic being discussed but about the interests of iMechanicians.  A discussion of simpler subtopics will help some of us learn but I doubt that the level of participation will increase significantly.  There are only so many things one can think about at a time.

-- Biswajit 

Zhigang Suo's picture

Dear Biswajit:  Thank you for the comments about the spam users.  We are looking for solutions.  Let me know if you have any ideas.

Meanwhile, you can click on "comments at a glance" on the right side of iMechanica.  These comments are filtered by some of us, and are spam-free. 

You should be able to identify the IP addresses from which the spam is being generated and block those if necessary.  If that's too draconian then add a captcha to the comment field and then block users that continue to spam.  I  also delete junk messages on iMechanica when I find them but that doesn't seem to have solved the problem.

-- Biswajit 

This spammers are really annoying. Today I found 25 new comments in the rss feed, of which 21 were spam messages!! And sometimes it is even worse...

Jayadeep U. B.'s picture

I agree with all that the spam messages are really irritating.

Blocking users alone might not work, as it is quite easy to create new user profiles and send spam messages.  I think, these are not done by human beings even, rather the softwares are automatically creating user profiles and then use them for spamming.  So there should be better checks at the time of creation of user profiles.  I would suggest the following:

1.  Require new users to fill more details, while creating profiles (affiliation etc.).  It is nice to have this information in future also.

2. More importantly, some standard checks (like asking to fill the charactors from an image) should be there to prevent the automatic creation of profiles.

Jayadeep

Gopinath Venkatesan's picture

Added to that, we can also implement a confirmation by mail system, for example, if I wanted to post a comment, I type and submit, then the system sends an email to the email ID so I can confirm that I want the material to be posted. A time lag is unavoidable.

On the Journal Club:

To initiate interests, topics can be somewhat simpler, can also discuss at some length on "How did I do it ...".

Vgn

Graduate Student

University of Oklahoma

Hi,

0. Though I am replying to Zhigang, I mean this as a thought for everyone...

1. I think there is no way to prevent spamming. But you can control and minimize it. The only way to accomplish this is to have some kind of a validation system.

2. Given the open nature of the Internet, and of the forum, I know that many people would be against any such an idea. Yet I think "validation" need not form a very serious issue...

Here, I can think of a system that would make use of precisely the same connectivity as is afforded by the Internet---its famous six degrees of separation. The solution essentially depends on a chain of validations---the same idea as is used in network marketing (like, say, AmWay).

3. Solution:

3.1 The iMechanica admin himself directly validates, say, only 10 people, one of who is, say, Biswajit. Biswajit, in turn, should have no problem validating someone like me (say as going by my email, Web, mobile, etc. information that I share with him). Once a member, I, in turn, should have no problem helping validate users coming from, say, Pune region, say, my students at COEP... (Geographical area need not be the sole criterion here.) ... So on and so forth.

3.2 We thus require every new user to supply email and other credentials, say, phone/fax/affiliation/ID card scan info to a validator---and the latter himself may very well be only an intermediate validator in the overall link. 

3.3 Once spamming or bad usage is notices, the admin contacts the validator of the spammer. The validator gives, say, 7 days to the spammer to improve and/or stop. If not, the iMechanica admin publicly notifies and then, if necessary, deletes the spammer. Notice, an entire branch might have to be deleted one validator at a time. But the point is, spammer gets localized very soon.

3.4 Notice, iMechanica does not store these validations (like ID card scans) on its servers. This permits it to remain flexible and free. ... Indeed, there need not be any fixed set of requirements as to what criteria a given member may employ while validating others---the specifics could very well differ from case to case. After all, every member is a validator, and at the same time, every member stands to lose his own membership if he is too lax in getting other members...

3.5 Validation is a distributed task in this model; every ordinary member is a potential validator. This spares admins and moderators any excessive load/responsibility, and lets every individual member feel important and responsible.

3.6 Though it won't be necessary, there won't be any harm in displaying a link in each member's profile as to which other member has validated his membership.

3.7 Notice, all members are potential validators. So, none should have any issue with this arrangement.... It's just a 'netization of what we do in real life: depend on the word-of-mouth (and pull your friend if he brought along some nasty company to your party).

My two cents.

Zhigang Suo's picture

Thank you all for the suggestions.  After some thoughts and experiments, I've made two changes:

  1. A new user will see a longer registration form.
  2. A new user can log in only after being approved by a user manager.

We'll watch the effect of these changes in the following days.

 

There are many here amongt us who prefer not to use their real name, and not give personal information.

Zhigang Suo's picture

For these people all they need to do is to send a note to one of the user managers, and explain that they are not spammers.

Note that all the identifying information on the registration form remains optional.

How about listing a few potential topics and have a system that allows people to vote. Then the journal topic can be determined based on this. How's that for a change?

Pradeep Sharma's picture

Dear John,

 

I express here some personal views on what is going on and some suggestions on what we can do about improving the j-club. Not all my suggestions are likely to be practical but I hope they will stimulate discussion that will converge to ideas that are.

 

Currently, I believe we are now seeing a type of apathy borne out of busy schedules, wearing
of the novelty etc. A further issue is that it is dawning on (some) people that while
iMechanica is an informal forum, it has similar impact as archival media and can
therefore negatively impact reputations if comments are not written carefully. Well written
posts and comments take time. This I think prevents many serious mechanicians to
actively participate even though they may be interested and read the
posts.

 

Here are some suggestions:

 

(i) We reduce frequency

(ii) At least for the next year or so, line
up prominent people in our area. One simple way to do this is take picks from
the Timoshenko awardees, chief editors of journals, NAE members etc. Ask them to
cast a broad net in their j-club issue and use their knowledge of the field to
discuss what the key challenges are in a given area. This may be easier said than done....While I enjoyed and benefited from nearly all the journal club issues, implementation of this suggestion may perhaps lead to an increase in participation as well as attract a class of participants who normally don't interact on iMechanica. Regarding the latter, please see my next point.

(iii)  Some private exchange between us (Michelle, you, Teng and Zhigang), stuck a chord: if some of the posts are review oriented, there may be less
discussion. Controversies always attract more discussion.....is it worthwhile to
experiment the following?: we choose a guest editor that can invoke a discussion
about something very controversial and see how many people respond? If this
experiment works then we may have found the niche for iMechanical jclub. If not
then of course we are back to square one. This experiment may very well entice participation by some people who normally are not active on iMechanica.

(iv)Finally, the current model is not working as a
"journal club" but given the consistency of the number of reads it is clearly of
value to the community. May be we could simply rename it and redefine success to
mean '"number of reads" rather than "number of comments" (-----I know this
sounds lame but I am getting old and now learning all the nifty tricks used by university administrators!)

 

 

i) I actually don't believe frequency is an issue.  Once a month works for most "real" journal clubs.  I don't see why it can't work at iMechanica.

ii)While I think this would be valuable for iMechanica to do, I'm not sure it fits the concept of a journal club.  As I mentioned to you at one point, I'm also not sure people wouldn't be intimidated by such luminaries penning issues.  

iii)I think this is an interesting idea. The challenge would be to come up with a good list of current "controversies", but I think if we could do that then the rest would follow.

iv) I'm not a big fan of redefining success.  Besides, these posts are stickies at the top of iMechanica's main page.  Given the page hits the site generates, it's not surprising they get lots of reads.  I'm not sure lots of reads can actually be interpreted as something that's necessarily of value to the community.  

Pradeep Sharma's picture

John, I guess then suggestion (iii) may be worth trying out. "Redefining success" is half in jest but it does hide a serious part that I do believe in (and disagree with you on): the number of reads have been quite consistent across the various jclub issues (both well commented ones and ones that were not). If larger than average number of people are reading the jclub (as opposed to other regular posts) then those issues which are not well commented are still adding a lot of value. Of course, such relatively-less-commented issues do not conform to our initial view of a "journal club" but are servicing more as tutorials or information repositories. Perhaps the niche that imechanica jclub can fill in (which journals often don't do) is engage debate on controversial topics.

A few more thoughts off the top of my head.

1) Instead of making guesses about the audience of JC posts, we can data mine the IP address list to find out the set of people actually clicking on the post.  Note that one doesn't have to click unless one has some interest in the post.  If the clicks are mostly from new users or passers by then JC needs some serious rethinking.

2) JC articles point us in directions that we may not be interested in at that moment in time.  However, such articles make us aware of possibilities that we may have not thought of.   I have found JC articles useful while thinking of ideas for new areas of research.  

3) I assume that most readers don't comment because they don't have anything useful to say on the subject under discussion.  It's up to knowledgable persons to start a discussion.  Do they have the time?

4) History shows us that JC posts will rarely get comments from the lay reader.  And experts will hesitate to comment on posts written by people whom they perceive to be beginners.  I feel that JC posts by experts will draw comments from other experts.   Such comments may act as icebreakers for the rest of the readers.

5) If experts need to be paid to become motivated enough to write JC articles then some sort of ad-based approach could be used to generate the needed revenue. 

6) Avoid JC articles that are advertisements for the work of a particular person/group.

Julian J. Rimoli's picture

I think we could address both John's and Pradeep's concerns regarding the measurement of success. I agree that the number of responses is an important issue but I also consider that the number of reads is also a good indicator if used together with a web traffic analysis tool. As a way of example, by using google analytics we could not only count the number of readers of a given post but also the time they spend reading it, the percentage of new visitors, the bouncing rate, etc.

Teng Li's picture

The visibility and readership of jClub have remained steady, with 2000
some reads on average, and 3000-4000 reads for some with more comments.
Yes, many read but a few comment. So one simple and important remedy is
to increase the number of comments for jClub issues. To some extent,
it's a perception issue, for many or even majority readers of the jClub
issues. And quite often we see snowball effect on commenting a specific
post.

Here are some quick thoughts on commenting and suggestions from you all:

  1. Can we effectively encourage comments from students, postdocs? Majority
    users of iMechanica nowadays are graduate students and postdocs. Most
    comments on jClub are not from this group.
  2. It will be nice to have prominent people host a few issues, but I am
    not sure how successfully we can earn their time and effort to such a
    service. We already have a few prominent people generously serving
    through iMechanica. It could be hard to have more. Maybe we can leave
    this as an option, and plan a mixed scheduel of issues from both junior
    researchers and prominent people.
  3. (This originates from some off-line discussion) We can experiment with the idea of promoting ad hoc posts to the top of front
    page. Technically it's doable. Maybe we should limit to maximum two
    highlighted posts (regular jClub issue and an ad hoc one) at the same
    time.  This could be more encouraging for the iMechanica students/postdoc group.
    We can experiment on this to set the criteria of promotion (number of
    reads, comments, etc).
Xiaodong Li's picture

Thanks a lot Teng. For each theme, the discussion leader knows the area well and may invite friends and colleagues to give both general and in-depth comments. This will stimulate readers' interest and they may follow up. I think that personal invitation from the discussion leader may help. 

We've actually been doing this for awhile.  For whatever reason, it has not been particularly effective.

Subscribe to Comments for "How Should Journal Club Work at iMechanica?"

Recent comments

More comments

Syndicate

Subscribe to Syndicate
Error | iMechanica

Error

The website encountered an unexpected error. Please try again later.