Rating One's own file
Divakar Roy
- This rating of one's own file in some cases unintentional though, but still there should be some automated system to "not accept" one's own ratings.
- If one is rating a file as one star that basically means dislike, he/she should give proper reason.
Hope points are heard somewhere.
Divakar
Divakar Roy
RJOllos
> I really wish that Mathworks guys would do something about authors rating their own files. I did search for threads with this same topic, haven't got anything, so thought to raise it.
Previously it was possible to rate a file multiple times. It appears
that has been fixed ... so hopefully they fix this as well.
For example, I posted multiple comments here and the second rating
seems to remove the previous rating:
http://www.mathworks.com/matlabcentral/fileexchange/22968-gridbaglayout
Previously I had posted there multiple times and each posting added
another rating.
John D'Errico
Actually, I had already sent an e-mail to Shari, asking
her to remove the self ratings as well as your response(s)
on this file.
Until they fix this, we can do no better. I've been
asking for that fix for the last year, so who knows
when it will happen. I hope soon.
John
Helen Chen
things that we would like to do in the future, although the exact date of
implementation is not yet determined. Until then, as John notes, feel free
to contact us directly or at fi...@mathworks.com if you need help with
anything.
John - Shari is on a well earned vacation, so probably hasn't seen your
message. Sorry about the delay in response.
Happy new year to everyone! :-)
Helen
"Divakar Roy" <devro...@gmail.com> wrote in message
news:hhh913$3cv$1...@fred.mathworks.com...
Jan Simon
> Previously it was possible to rate a file multiple times. It appears
> that has been fixed ... so hopefully they fix this as well.
>
> For example, I posted multiple comments here and the second rating
> seems to remove the previous rating:
> http://www.mathworks.com/matlabcentral/fileexchange/22968-gridbaglayout
>
> Previously I had posted there multiple times and each posting added
> another rating.
It "seems" to remove the previous rating.
But in the average rating of an author, mutliple ratings of one user for one file are counted. So if you rate a submission with 1 star and the author updates its submission such that you give 5 stars, the author gets an average of 2.5 stars.
Is this correct, wanted and useful?
Jan
Divakar Roy
Wish you and everyone else Happy New Year too!!
Divakar
"Helen Chen" <hel...@mathworks.com> wrote in message <hhiedj$kep$1...@fred.mathworks.com>...
Divakar Roy
But, in the total ratings thing, it's still counting the invalidated previous rating, it should be one less, though a minor thing, still a thing.
Hope that is taken care by Mathworks.
Best,
Divakar
"Jan Simon" <matlab.T...@nMINUSsimon.de> wrote in message <hhip88$h0b$1...@fred.mathworks.com>...
Matt J
> I really wish that Mathworks guys would do something about authors rating their own files.
============
Is this really such a big deal? Politicians vote for themselves in elections. We'd consider them crazy not to.
Divakar Roy
"Matt J " <mattja...@THISieee.spam> wrote in message <hhiu02$b42$1...@fred.mathworks.com>...
John D'Errico
Since you are someone who recently chose to rate all
of their own files, with amazingly glowing comments,
your statement seems strongly self serving.
Since a file gets assigned an average rating, to include
the authors self appraisal is inappropriate. Many of the
files on the file exchange get no ratings at all, some
only get at most one if they are lucky. So to include
your own ratings will strongly bias the result. We know
that YOU think your own work is good, as why else
would you post it?
A high rating (especially with glowing comments) on
a file does two things. First, it encourages others to
download that submission. Second, it acts as a teaching
tool. Good ratings tell other authors when a piece of
code is well done. It offers them something to emulate
for their own work. By allowing self serving, meaningless
ratings, the value of the FEX would be decreased.
In the case of a political vote, there will be thousands
or millions of votes cast. To include or exclude a single
vote will not bias the results significantly. And in that
case, it makes sense to allow a politician to vote for
themselves. After all, ALL citizens are allowed a vote.
The FEX rating system provides a peer review process,
where others are allowed to offer their opinions of
your work. If you wish, think of the FEX as an online
journal. But self ratings pervert the process.
While I wish for (and have pushed for) a system in
place to encourage more votes, with more active
participation in the review process, it is what it is.
There are other forums in place, such as slashdot
and stackoverflow, both of which prevent self
moderation of input, but also encourage good
behavior of the citizens who inhabit these domains.
Both of those domains seem to work reasonably
well.
John
Matt J
> Since you are someone who recently chose to rate all
> of their own files, with amazingly glowing comments,
> your statement seems strongly self serving.
No, I really don't care whether we get to rate ourselves or not. I just don't see the harm in it.
For the record, my "glowing comments" were meant jokingly, as anyone who read them (aside from yourself apparently) would have seen. I do have some misgivings in retrospect that they would have spammed the Inboxes of people watchlisting the submission. On the other hand, if anyone was watchlisting the submission they would presumably already be a fan of the file, and so would presumably not disagree with a high rating.
> In the case of a political vote, there will be thousands
> or millions of votes cast. To include or exclude a single
> vote will not bias the results significantly.
That principal would apply here as well.
It would be ludicrous for someone to judge an FEX submission based on its average score, whether that score be good or bad, when it is only based on 1 or 2 ratings. A score like that can't be trusted as statistically significant, as you can never know how thorough or fair those 2 people have been in their appraisal of the file. And that's true regardless of whether the author is one of the people submitting the rating.
Conversely, a high rating based on 20+ contributors is more rationally to be trusted, but then who cares if the author weighed in on that.
People should, and I believe do, download submissions based mainly on the description of the submission and whether it suits their needs, and not on the ratings. In the rare case where you have nearly identical submissions, maybe you'd use the ratings as a tie breaker, but again only if there is a statistically significant number of them.
> Many of the
> files on the file exchange get no ratings at all, some
> only get at most one if they are lucky.
Because most people can't be bothered submitting a rating. And that's possibly because they know the rating system is inevitably flawed, as I've contended.
Jan Simon
> Is this really such a big deal? Politicians vote for themselves in elections. We'd consider them crazy not to.
Right, Matt: It is not a big deal.
Selfratings are no real problem for the quality of the FEX submissions.
I even dare to claim, that ratings are not a big deal at all in the FEX. I see in the top 100:
James, >4000 downloads, 23 submissions, 5 ratings, Average: 5
Bruno, >5000 downloads, 28 submissions, 22 ratings, Average: 4.8
Matt Fig, >7000 downloads, 33 submissions, 101 ratings, Average: 4.7
Without doubt, these submissions have a very high quality. But how do we know?
1. The number of ratings is not meaningful (see Bruno).
2. The number of submissions imply, that the authors are active and hardworking. But Marco is more assiduous.
3. The average rating? Then the excellent 5961-magnify is one of the highlights with 28 * 5 stars (77 downloads in the last month).
4. The number of downloads? "Hough Transform for Circle Detection" of Tan Chin Luh 9898 was downloaded 336 times and got some strange ratings (1, 4 and 1 stars).
Nope. I know, that these authors work with a high quality because they have demonstrated it again and again over the last years.
There are also negative examples in the top 100 FEX authors - I don't feel like commenting this. No fun and no profit.
However, selfratings never caused problems for me. I don't like them, but who cares... I do not care for other ratings or thousands of downloads also, as long as John likes the function!
Thanks to the mentioned authors, thanks to all authors who published in the FEX in 2009, thanks to CSSM contributors, thanks to TMW!
Jan
PS. I found several selfratings with 1 and 2 stars - funny.
Matt J
> I do not care for other ratings or thousands of downloads also, as long as John likes the function!
============
Jan makes another good point about the difference between a peer-reviewed journal and the FEX: in a journal, the quality of the reviewers are controlled, whereas on the FEX they are not. Thus, we're only going to trust reviews from people we know well.
One more reason why the average rating won't (or at least shouldn't) be the primary factor in a decision to download....
John D'Errico
> "John D'Errico" <wood...@rochester.rr.com> wrote in message <hhj4g6$k71$1...@fred.mathworks.com>...
>
> > Since you are someone who recently chose to rate all
> > of their own files, with amazingly glowing comments,
> > your statement seems strongly self serving.
>
> No, I really don't care whether we get to rate ourselves or not. I just don't see the harm in it.
That someone else might actually believe something
you said, without noticing the similarity in names.
> For the record, my "glowing comments" were meant jokingly, as anyone who read them (aside from yourself apparently) would have seen.
Personally, I prefer honesty. Feel free to try it one day.
John
John D'Errico
I can't design or control the system that is there.
Our choice is only to work with the system as it
is in place.
John
Husam Aldahiyat
I don't understand. Downloading the files here is FREE, and takes seconds, so why not download codes based on their description, try them out and judge based on personal fondness?
The rating system serves, to me, as a thank you for the person who uploaded the code. I get some 1 ratings every now and then but it doesn't upset me if the person who gave the rating doesn't explain it. On the other hand, a 5 rating makes me feel all fuzzy inside, even though it never affects the number of downloads.
Divakar Roy
- With one star and no comment, it’s like you don’t know why the user didn’t like it.
Though I do understand one or two bad or good ratings would not make a lot of change on the viewpoint about our own coding/type of codes or whatever but still it’s something an author would wonder. I might not lose my sleep on it, but for authors who love MATLAB, it’s quite something they would care to know.
Happy New year to all By the way. I think it's better to leave this whole discussion of ratings issue to Mathworks, let them decide, and lets hope FEX continues to "HELP" everyone associated with MATLAB. Peace all.
clear all;help all;
"Husam Aldahiyat" <numa...@gmail.com> wrote in message <hhk2pj$3u$1...@fred.mathworks.com>...