Thread Subject: New File Exchange download counts

Dear community members,

This morning, we have made a change to how we define the displayed download count at MATLAB Central.

Historically, the number of downloads were captured and displayed. Several of our community members suggested that filtering would minimize the impact of repeat downloads by individuals, crawlers or bots. As of this morning, our new calculated count reflects one download per IP address per 30 days. Because of this, you will notice that current file and author numbers are lower than they used to be. We believe this new number better reflects interest in our contributed files. We hope you do too!

Thanks to each of our community members who took the time to contact us suggesting this improvement to File Exchange. We do appreciate your comments and feedback.

Best,
Helen Chen
for the MATLAB Central Team

"Helen Chen" <helen.chen@mathworks.com> wrote in message <gsi6pb$7uj$1@fred.mathworks.com>...
> Dear community members,
>
> This morning, we have made a change to how we define the displayed download count at MATLAB Central.
>
> Historically, the number of downloads were captured and displayed. Several of our community members suggested that filtering would minimize the impact of repeat downloads by individuals, crawlers or bots. As of this morning, our new calculated count reflects one download per IP address per 30 days. Because of this, you will notice that current file and author numbers are lower than they used to be. We believe this new number better reflects interest in our contributed files. We hope you do too!
>
> Thanks to each of our community members who took the time to contact us suggesting this improvement to File Exchange. We do appreciate your comments and feedback.
>
> Best,
> Helen Chen
> for the MATLAB Central Team

Excellent. Thanks Helen! I think this will improve the FEX vastly.

As far as user ratings, maybe a limitation where ratings from the same IP as the author are excluded? Something to think about.

Hey, Thanks for the feedback!

> As far as user ratings, maybe a limitation where ratings from the same IP as the author are excluded? Something to think about.

I hadn't thought of this - on first read I thought that this was covered since when you rerate something, your previous rating is grayed and is not included in the calculation.
But could help some situations where individuals create multiple accounts and use them all for playing"the rating game". Interesting....

I'll add it to the list of enhancement suggestions.

Thanks again for sharing!

Helen

"Helen Chen" <helen.chen@mathworks.com> wrote in message <gsi6pb$7uj$1@fred.mathworks.com>...
> Dear community members,
>
> This morning, we have made a change to how we define the displayed download count at MATLAB Central.
>
> Historically, the number of downloads were captured and displayed. Several of our community members suggested that filtering would minimize the impact of repeat downloads by individuals, crawlers or bots. As of this morning, our new calculated count reflects one download per IP address per 30 days. Because of this, you will notice that current file and author numbers are lower than they used to be. We believe this new number better reflects interest in our contributed files. We hope you do too!
>
> Thanks to each of our community members who took the time to contact us suggesting this improvement to File Exchange. We do appreciate your comments and feedback.
>
> Best,
> Helen Chen
> for the MATLAB Central Team

Helen,

SPLENDID. Many thanks to the team for getting this in.

John

"Helen Chen" <helen.chen@mathworks.com> wrote in message <gsi6pb$7uj$1@fred.mathworks.com>...
> Dear community members,
>
> This morning, we have made a change to how we define the displayed download count at MATLAB Central.
>
> Historically, the number of downloads were captured and displayed. Several of our community members suggested that filtering would minimize the impact of repeat downloads by individuals, crawlers or bots. As of this morning, our new calculated count reflects one download per IP address per 30 days. Because of this, you will notice that current file and author numbers are lower than they used to be. We believe this new number better reflects interest in our contributed files. We hope you do too!
>
> Thanks to each of our community members who took the time to contact us suggesting this improvement to File Exchange. We do appreciate your comments and feedback.
>
> Best,
> Helen Chen
> for the MATLAB Central Team

Hi Helen

That's good news for users of the FEX, I think.

Another suggested improvement (very superficial):
Currently when you update a file on the FEX, if that file acknowledges another file then the update causes an extra comma to appear in the list of acknowledging files of said acknowledged file. Look at http://www.mathworks.com/matlabcentral/fileexchange/10889 for a long list of commas (caused by my updates). Also, the authors get notified again (by email) that their file is being acknowledged, something they know already. I think both these features should be removed.

Regards,
Oliver

> Another suggested improvement (very superficial):

Thanks Oliver. I'll report the update issues.

Helen

"SN Orkshef" <nospam@spam.com> wrote in message

I think posting the download log daily makes the point, but you seem to be posting much more often than that!

"SN Orkshef" <nospam@spam.com> wrote in message <gsvbu9$bu0$1@fred.mathworks.com>...
<SNIP ... a too long log of MP downloads ...

> He is hammering away right now. I'm going to start posting his download log daily to help illustrate the problem.

I suggest you don't. Do not get involved in his childish behavior ...

Jos

Innumerable posts have been made on the newsgroup about how to manage the FEX, i.e. what rules to implement to clean it.

However, all the cleaning that was requested was because of Marco, and all the rules that were proposed were merely ways to avoid constantly stumbling on Marco's files, whatever search is launched on the FEX.

TMW could perhaps save a lot of time to all of the people suggesting these rules by simply banning Marco? I know the talk about freedom of speach, etc. but at a certain point, a decision has to be made...

In addition, we're talking about the administration of the FEX, not of a country... no one said it had to be democratic ;)

"Gavrilo Bozovic" <gavrilo.dot.bozovic@gmail.dot.ch> wrote in message <gt4av3$7nm$1@fred.mathworks.com>...
> Innumerable posts have been made on the newsgroup about how to manage the FEX, i.e. what rules to implement to clean it.
>
> However, all the cleaning that was requested was because of Marco, and all the rules that were proposed were merely ways to avoid constantly stumbling on Marco's files, whatever search is launched on the FEX.
>
> TMW could perhaps save a lot of time to all of the people suggesting these rules by simply banning Marco? I know the talk about freedom of speach, etc. but at a certain point, a decision has to be made...
>
> In addition, we're talking about the administration of the FEX, not of a country... no one said it had to be democratic ;)

Technically, I think it would still be a democratic decision, since I gather that the majority of FEX users would vote to have "him" removed. =)

However, I doubt TMW will ban anyone. We'll just have to bide our time until the appropriate filters are incorporated into the FEX so that we can all erase "him" from our purview.

"Gavrilo Bozovic" <gavrilo.dot.bozovic@gmail.dot.ch> wrote in message <gt4av3$7nm$1@fred.mathworks.com>...
> Innumerable posts have been made on the newsgroup about how to manage the FEX, i.e. what rules to implement to clean it.
>
> However, all the cleaning that was requested was because of Marco, and all the rules that were proposed were merely ways to avoid constantly stumbling on Marco's files, whatever search is launched on the FEX.
>
> TMW could perhaps save a lot of time to all of the people suggesting these rules by simply banning Marco? I know the talk about freedom of speach, etc. but at a certain point, a decision has to be made...
>
> In addition, we're talking about the administration of the FEX, not of a country... no one said it had to be democratic ;)

As much as I want to agree, this turns into one of
those slippery slopes. Where do they stop removing
files? This is a problem we had when we originally
tried to clean up the FEX a few years ago.

Suppose we decide that Marco is obvious spam. Yes,
his crap is exactly that. I'm sure that we all would
like to see it disappear. How about the homework
assignments? While I'd like to get rid of the many
crappy versions of Gaussian elimination, at what
point does a homework assignment become worth
deleting? Suppose that as a student, I find a simple
solution to the 4 color problem (or Goldbach, or FLT,
etc.) posed to me as a homework assignment. I'm
impressed with my solution, so I post it to the FEX.
Should it be deleted?

The problem is that the FEX admin team cannot do
the sorting. They cannot make the choice between
files that truly are crap and the rare gems between
them, because they may not have the expertise to
make that judgement, and they certainly do not have
the time required to carefully assess each submission.

Ratings and reviews make sense here, making the FEX
a peer reviewed site. Only because there are many
hundreds or thousands of people willing to make
an occasional assessment of a few of the many files
up there, can we try to sort out the gems. I've been
pushing for a few years that these ratings should
use some form of reputation measure, so that
ratings become a more valid assessment of the
worth of a submission. I like the StackOverflow
model. Check it out.

http://stackoverflow.com/

Something like this can work nicely on Matlab
Central, where your reputation can be built up
from a variety of ways that a person contributes
to this community.

John

"Kenneth Eaton" <Kenneth.dot.Eaton@cchmc.dot.org> wrote in message <gt4c76$d78$1@fred.mathworks.com>...
>
> Technically, I think it would still be a democratic decision, since I gather that the majority of FEX users would vote to have "him" removed. =)
>

I don't know. Suppose I don't like an individual
I meet on the street. In fact, this person is a
quite disagreeable curmudgeon, agreed upon by
many members of society. Can the members of
a democratic society (outside of Texas) vote to
have him put to death? I suppose if this person
has committed a felony, he might be locked up.
I.e., we might limit his ability to interact with
other members of society.

Has Marco committed a felony? Sorry, but no.
He is a nuisance. An irritating one. And he has
even committed a few misdemeanors, with his
self ratings. Even if he is the one who has figured
out how to do IP switching to puff his download
counts, this is not much more than a speeding
violation.

The best way to deal with Marco to ignore him.
Filter him out. If he knows that nobody cares
that he exists, his reason for existing on the site
will go away.


> However, I doubt TMW will ban anyone. We'll just have to bide our time until the appropriate filters are incorporated into the FEX so that we can all erase "him" from our purview.

Yes.

John

Helen,

How about a filter to make the rating stars on your own files un-selectable? There has been a rash of abuse recently:

http://www.mathworks.com/matlabcentral/fileexchange/24361
http://www.mathworks.com/matlabcentral/fileexchange/24334
http://www.mathworks.com/matlabcentral/fileexchange/24293
http://www.mathworks.com/matlabcentral/fileexchange/24281
http://www.mathworks.com/matlabcentral/fileexchange/24273

And these are only off of the first page of the current "Submitted" tab. It is disappointing to see a 5 star rating, decide to take a look at an "Excellent" file, only to find that the author thinks his/her own file is excellent. I would assume that is the case for most submissions. However, I am really only interested in an excellent file that has been judged so by persons other than the author.

Just a thought.