Thread Subject: MATLAB Central is proud to annouce our new File Exchange!

Hello MATLAB Central Community Members –

I am pleased to announce that early today MATLAB Central replaced our existing File Exchange interface with a new design. This entirely new application was built and is hosted by the MathWorks, Inc. We will no longer be using our existing hosted solution. The new File Exchange has many new features for your improved use. These features include:


- Ability to filter discussion content by a variety of different time periods: all, 7 days, 30 days, last year.

- Ability to track discussion threads using your own personal “watch lists”

- Ability to organize threads by tagging with a word or phrase that are meaningful to you

- Ability to select content for viewing using single or multiple tags

- Ability to select content by author

- Ability to set—up RSS feeds to have your site updates delivered to your desktop using your favorite RSS feed reader.

- Decreased spam (due to required login to your MathWorks Account)


Please log in in order to submit files or to tag. Click on the “Submit a File” button to submit new files. If you are reading a submission and want to leave a comment, please click on the “Please login” link in order to log in and leave your comments. To tag, simply enter your tags in a comma-separated list in the Tag block. Send your questions, comments or feedback to us at Files@mathworks.com. Please also be aware that there is a FAQ available at
http://matlabwiki.mathworks.com/Using_File_Exchange


Enjoy!
Cristina McIntire
MATLAB Central Administrator

Thanks Cristina! I would like to invite everyone to check out the new File Exchange and let us know what you think about the new functionality.

If you are new to File Exchange, or even if you are an old hand, check out the short videos that Matthew has shared with us on cool topics like submitting a file, finding files and finding authors.

Enjoy!

Helen Chen
MATLAB Central Team

Well done on a much more appealing look to the File Exchange, Mathworks! Hopefully it will be even easier to find the files I'm looking for now, too.

A few comments:
It's great that I can change the tags on my files without having to resubmit the file. It's also great that we can add tags to other people's files.

It would be nice if we could update the description, perhaps even the title of our own submissions too. We could do that when resubmitting/updating the file previously. Now we can't do it at all. It would be nice to be able to change the title, description, tags and screenshot as and when we like (not only when resubmitting/updating).

2 of my 6 comments and ratings got lost in the port to the new system. They were the longest and most informative comments too. :(

Since you guys are now hosting this I hope we'll see new features and improvements to the system more often!

"Oliver Woodford" <o.j.woodford.98@cantab.net> wrote in message <gdo5p9$f8j$1@fred.mathworks.com>...
> Well done on a much more appealing look to the File Exchange, Mathworks! Hopefully it will be even easier to find the files I'm looking for now, too.
>
> A few comments:
> It's great that I can change the tags on my files without having to resubmit the file. It's also great that we can add tags to other people's files.
>
> It would be nice if we could update the description, perhaps even the title of our own submissions too. We could do that when resubmitting/updating the file previously. Now we can't do it at all. It would be nice to be able to change the title, description, tags and screenshot as and when we like (not only when resubmitting/updating).
>
> 2 of my 6 comments and ratings got lost in the port to the new system. They were the longest and most informative comments too. :(


I did a quick check of my own files. It appears that
between 20% and 50% of the reviews disappeared.

One file lost 15 out of 44 reviews. Another lost 5
out of 9 reviews. Almost every one of the 47 files
I've put on the FEX lost at least few of the reviews
it had received.

My guess is that only reviews by individuals who
have a matlab central account were retained.

John

Another comment on the new file exchange is that the order of the ratings is no longer shown chronologically which makes it difficult to see the progress of enhancements. This also makes it difficult to notice when someone corrects their own rating or comments.

Likewise the update list also appears in a randomized order (which is even more weird.

~Paul

Thank you MATLAB for trying to improve it. I like it, and have some comments:

1. Many options were deleted (like the file statistics and Authors Map, etc), but, less is better. GOOD
2. When I click on a LINK (blue) it doesn't turns to readed (violet)???. BAD
3. Yes, the new way to display comments is wrong I think. Should be better to put it cronologically descending like everything else. BAD
4. The authors and files web page are now very well organized because of the hosting (now I must update all my files to update the new good-looking link). GOOD
5. Code Metrics was fixed. GOOD
6. Delection of spam comments is good, although at this point, a bit confusing of how this was/is/will be done. GOOD
7. Now, we cannot know if the downloadable file is an .M, .P, .PNG, .PDF, .EXE, etc, file!! Only ZIP contents! BAD
8. I still dont know if the updated files will be marked as new (updated) submitions, which will be good so the community knows about it. Although I see that there is know a release and update date next the file description. GOOD
9. BLOGS on MATLAB Central Index is not working ??
10. In resume, the change is great, and I hope there will be more File Exchange releases.

Thank you,
Carlos Vargas

"Helen Chen" <helen.chen@mathworks.com> wrote in message <gdo1ve$kuo$1@fred.mathworks.com>...
> Thanks Cristina! I would like to invite everyone to check out the new File Exchange and let us know what you think about the new functionality.
>
> If you are new to File Exchange, or even if you are an old hand, check out the short videos that Matthew has shared with us on cool topics like submitting a file, finding files and finding authors.
>
> Enjoy!
>
> Helen Chen
> MATLAB Central Team


There is no display of FEX files by category? I think this is a huge omission. Sure, the "Files by Product" will be useful but I think there must be a category (science or engineering category) sort available. Perhaps I have missed it?

> My guess is that only reviews by individuals who have a matlab central account were retained.

I don't think this is the case because I'm registered and a comment of mine was lost on someone else's file (it was a long comment, and a shorter one from me in the same list was saved), and also because there are plenty of reviews remaining which are most likely by unregistered users (e.g. fake name/email).

My guess is that long comments or comments with a carriage return have been lost.

Can we get them back, MathWorks?

Oliver

Missing comments are a problem. Those remaining often make no sense outside of the context of the missing comments. The lack of chronological order has a similar effect.

> It would be nice if we could update the description, perhaps even the title of our own submissions too. We could do that when resubmitting/updating the file previously. Now we can't do it at all. It would be nice to be able to change the title, description, tags and screenshot as and when we like (not only when resubmitting/updating).
>

It does require that you repost the file to get to that page. The second page lets you edit any of your other page content.

Thanks for pointing that out.

> 2 of my 6 comments and ratings got lost in the port to the new system. They were the longest and most informative comments too.

Yes, we've been looking into these reports and found the issue. There was an issue with longer replies (hence why the responses that folks really worked on), which resulted in those not being posted. We are working on a patch and these will reappear soon.

> Since you guys are now hosting this I hope we'll see new features and improvements to the system more often!

It is our goal to regularly review the feedback and enhancement requests and do more smaller updates to the various applications on File Exchange. So keep sharing your thoughts and suggestions!

Thanks,
Helen

Paul -

On the ratings, the order is oldest to newest, which is
intended to show the conversation over time. (This is consistent with Newsreader). If the user corrects a rating, the original rating will be dimmed so that you know it has been changed. The new rating will be displayed chronologically with the date of the new rating. If you see this not happening, could you please send the url where you are seeing it to either Cristina or myself, please.

The updates is an excellent observation. I have logged the issue for the engineers.

Thanks,
Helen



"Paul Mennen" <nospam@mennen.org> wrote in message <gdoa26$b5d$1@fred.mathworks.com>...
> Another comment on the new file exchange is that the order of the ratings is no longer shown chronologically which makes it difficult to see the progress of enhancements. This also makes it difficult to notice when someone corrects their own rating or comments.
>
> Likewise the update list also appears in a randomized order (which is even more weird.
>
> ~Paul
>

Hi Carlos - Thanks for all the feedback. See comments inline.

"Carlos Adrian Vargas Aguilera" <nubeobscura@hotmail.com> wrote in message <gdq0ra$e7g$1@fred.mathworks.com>...

> 2. When I click on a LINK (blue) it doesn't turns to readed (violet)???. BAD

I will share your feedback with the designers.

> 3. Yes, the new way to display comments is wrong I think. Should be better to put it cronologically descending like everything else.

The reversal was to try to get people to read what is already posted. Out of curiosity, if it is most recent first, do you actually read what others have said or do you just post your comment?

> 7. Now, we cannot know if the downloadable file is an .M, .P, .PNG, .PDF, .EXE, etc, file!! Only ZIP contents! BAD

Looking into this
> 8. I still dont know if the updated files will be marked as new (updated) submitions, which will be good so the community knows about it. Although I see that there is know a release and update date next the file description. GOOD

Looking into this too
> 9. BLOGS on MATLAB Central Index is not working ??

It was broken for a while this morning. What are you seeing now?

Helen

> > 2 of my 6 comments and ratings got lost in the port to the new system. They were the longest and most informative comments too. :(

We have just rolled out an update to the system. Please take a look at your comments. They should all be there now.

Also, the author rankings have changed to display downloads for the last 30 days rather than all-time.

Keep up the comments!

Helen

> We have just rolled out an update to the system. Please take a look at your comments. They should all be there now.

They're all there. Excellent work, super response time. Many thanks.

Nice layout!! However, is there a way to change the font size? They are super tiny and I can't change it using the regular View->Text Size-> Largest. Thanks!

Hi Helen!
 
>The reversal was to try to get people to read what is already >posted. Out of curiosity, if it is most recent first, do you >actually read what others have said or do you just post your >comment?
Yes, I always try to do so, because for sure JOHN or US (where is he?) had already point out already something good/bad! But then, once I comment something, I'm just interest in the author response.


> > 9. BLOGS on MATLAB Central Index is not working ??
>
> It was broken for a while this morning. What are you seeing now?
Is OK now!

The authors ordered by last 30 days is fine (although weird to see me on the top 50!, so no complains je je je).

Finally, somebody said something about the letter size, I have a little complain about the color of the links. It's a little hard to read this soft blue and grays.

Well, thank you very much!

Carlos

"Carlos Adrian Vargas Aguilera" <nubeobscura@hotmail.com> wrote in message <gdqoq6$qiq$1@fred.mathworks.com>...
> Hi Helen!
>
> >The reversal was to try to get people to read what is already >posted. Out of curiosity, if it is most recent first, do you >actually read what others have said or do you just post your >comment?
> Yes, I always try to do so, because for sure JOHN or US (where is he?) had already point out already something good/bad! But then, once I comment something, I'm just interest in the author response.
>

OK, good feedback thanks!
>
> > > 9. BLOGS on MATLAB Central Index is not working ??
> >

> The authors ordered by last 30 days is fine (although weird to see me on the top 50!, so no complains je je je).
>

Hey, Congrats on making it on that list! Way cool!

> Finally, somebody said something about the letter size, I have a little complain about the color of the links. It's a little hard to read this soft blue and grays.
>

Noted

> Well, thank you very much!
>
:-)

"Judy" <sauwen.jl@gmail.com> wrote"
> Nice layout!! However, is there a way to change the font size? >They are super tiny and I can't change it using the regular >View->Text Size-> Largest. Thanks!
~~

Hi Judy,

We'll be making minor adjustments over the next week or so and this is definitely on the list of things to be looked at.

My apologies.

~Matthew.

Hi,

another detail: the update list is not chronological and some (or most?) of the update description text is lost (see e.g. http://www.mathworks.de/matlabcentral/fileexchange/18430 )

All in all, congratulation to the new layout and features!

But one question:
Can you sort the Files list generated for a certain tag by the number of downloads during the last 30 or 7 days? In addition to the ratings, this may help you find 'high quality' submissions, too. If I click on 'Time Frame -> Last 30 days' it just filters out the older submissions, even after sorting by Downloads.

Regards
Andres

Please go back to a larger, black on white, font for most of the data. Small font, poor contrast printing - it looks nice until you have to read it, then it's an invitation to eye strain. I imagine that those with one of the more common forms of color-blindness will find it impossible to read the light blue font on a white or pastel background. I have a nice high res LCD screen - woe unto those who have a smaller resolution or a CRT.

I know you are trying to get a lot of information in a small space, but sacrificing viewability is a mistake. I used to look forward to going to MATLAB Central - I find the current interface very uncomfortable and unpleasant. I will have to find out how to use the visually impaired settings in Windows - this has never been an issue. I usually spend at least 15 minutes a day catching up at this site, but that seems unlikely to continue with this interface.

There seems to be little point in posting blogs, files, reviews, and questions if the interface is painful to look at.

Scott Miller

<snip>
> There seems to be little point in posting blogs, files, reviews, and questions if the interface is painful to look at.
>
> Scott Miller

I fully agree with Scott Miller. It looks nice, but please do something to make it easier to read.

Do something to the graphics to group the header and the bodytext of a post. Now the text is grouped with the header below.

The editbox in which I'm writing this message would gain from being larger - much larger.

/ per
Eizo Flexscan L795

Andres, you should be able to find the most downloaded files for a given tag by clicking on the tag or searching for "tag:whatever", then sorting on the download columns. The "Time Frame" column, as you note, only filters the list to files that have been updated or submitted recently. It shouldn't change the count in the download column, which is always last 30 days. We've changed download counts all over the site to always and consistently be "last 30 days". There were a couple bugs with this when we launched, which may be why you weren't getting the results you expected, but it seems to be working now.

"Matthew Simoneau" <matthew@mathworks.com> wrote in message <ge4jin$35g$1@fred.mathworks.com>...
> [..] It shouldn't change the count in the download column, which is always last 30 days. We've changed download counts all over the site to always and consistently be "last 30 days". [..]

Thanks for your reply, Matthew. But still the numbers showing up in the "Downloads" column seem to be the total number of downloads, not just from the last 30 days. E.g. search for "tag:gui" and sort by "Downloads" and you'll get download numbers of more than 10,000 for the first twenty files. ("Matlab GUI Tips" by Roy Schestowitz has 900dls/30d on http://www.mathworks.com/matlabcentral/fileexchange/5439 , but is listed with 52,911).

Andres, you're right. I thought we had fixed that already, but the download count looks like it is still all-time. Soon!

I think that you should add a mechanism that should prevent users to rate their own submissions like this example
http://www.mathworks.com/matlabcentral/fileexchange/21896

Having the total number of downloads available somewhere is useful to me. As an academic, it helps me to persuade my college that time spent developing code is well spent.

When posting a update to a file I have noticed
1. All HTML files are now listed as "Untitled doc"
2. The reason for the update appeared immediately, and before the update itself was approved and posted by TMW.

Overall, congratulations on the new site design.

"per isakson" <poi.nospam@bimDOTkthDOT.se> wrote in message <gdvu3j$idq$1@fred.mathworks.com>...
> <snip>
> > There seems to be little point in posting blogs, files, reviews, and questions if the interface is painful to look at.
> >
> > Scott Miller
>
> I fully agree with Scott Miller. It looks nice, but please do something to make it easier to read.
>
> Do something to the graphics to group the header and the bodytext of a post. Now the text is grouped with the header below.
>
> The editbox in which I'm writing this message would gain from being larger - much larger.
>
> / per
> Eizo Flexscan L795

I find the current newsreader extremely painful. The light blue background in the text fields decreases contrast and the white borders draw my eyes to the sides. Seems something as simple as making the background of the sides and headers light blue and the background of the text boxes white would be an improvement.

~Adam

"Ruben Faibish" <gsx1100f@nana.co.il.remove.this> wrote in message <ge6phe$oru$1@fred.mathworks.com>...
> I think that you should add a mechanism that should prevent users to rate their own submissions like this example
> http://www.mathworks.com/matlabcentral/fileexchange/21896

I notice that user now has an author ranking of 48. He was 600 something before the changeover. What gives?

gil

"Gilbert " <no@spam.com> wrote in message <ge8upm$m2i$1@fred.mathworks.com>...
> "Ruben Faibish" <gsx1100f@nana.co.il.remove.this> wrote in message <ge6phe$oru$1@fred.mathworks.com>...
> > I think that you should add a mechanism that should prevent users to rate their own submissions like this example
> > http://www.mathworks.com/matlabcentral/fileexchange/21896
>
> I notice that user now has an author ranking of 48. He was 600 something before the changeover. What gives?
>
> gil

The new rating system is based on the number of downloads all your files have had over the last 30 days. Before, it was simply the total number of downloads. The author in question has spammed the FEX with an exceedingly large number of poorly-crafted programs, posting 2-3 per week it seems. Somehow, these files actually get some downloads, and since he has so many files he gets a high ranking. I personally think the star ratings of files should be incorporated into the ranking system as well. This would balance out the effect of FEX-spammers like said author.

One great improvement to the FEX was made to the star rating system. Even though an author can rank his own code, he can only do it once. In fact, only the most recent rating any user gives to a submission is counted in the total average score. No user can rate a piece of code more than once, because the new rating supercedes the old one. This helps a bit with regard to ego-driven authors or spiteful reviewers.

Hal

"Judy" <sauwen.jl@gmail.com> wrote in message <gdqmoi$sr0$1@fred.mathworks.com>...
> Nice layout!! However, is there a way to change the font size? They are super tiny and I can't change it using the regular View->Text Size-> Largest.


We just pushed out a new build. You should be able to change the displayed text size now.

Please let me know if you have any other issues with this.

Helen

"Helen Chen" <helen.chen@mathworks.com> wrote in message <geclco$3jd$1@fred.mathworks.com>...
> "Judy" <sauwen.jl@gmail.com> wrote in message <gdqmoi$sr0$1@fred.mathworks.com>...
> > Nice layout!! However, is there a way to change the font size? They are super tiny and I can't change it using the regular View->Text Size-> Largest.
>
>
> We just pushed out a new build. You should be able to change the displayed text size now.
>
> Please let me know if you have any other issues with this.
>
> Helen

Yep, everything is rainbows and happy sunshine goodness (sorry Ken, lol) over in the new FEX:

http://www.mathworks.com/matlabcentral/fileexchange/21922

"Gilbert " <no@spam.com> wrote in message <ge8upm$m2i$1@fred.mathworks.com>...
> "Ruben Faibish" <gsx1100f@nana.co.il.remove.this> wrote in message <ge6phe$oru$1@fred.mathworks.com>...
> > I think that you should add a mechanism that should prevent users to rate their own submissions like this example
> > http://www.mathworks.com/matlabcentral/fileexchange/21896
>
Thanks, I'll bring this up with the team. :-)

> I notice that user now has an author ranking of 48. He was 600 something before the changeover. What gives?
>
The rankings used to be calculated since the start of MATLAB Central (circa Fall 2001). We changed that to only count downloads in the past 30 days so that we are really looking only at current activity. Previous feedback was that feedback from years gone by doesn't really give us an accurate picture of the author's current relative rank within the community.

BTW, Matt Simoneau has posted a nice FAQ about downloads at http://matlabwiki.mathworks.com/Using_File_Exchange#Total_Downloads
Take a look at his description and see if that helps to clarify how downloads works now.

Best,
Helen

"Gilbert " <no@spam.com> wrote in message <gecnju$j2t$1@fred.mathworks.com>...
> Yep, everything is rainbows and happy sunshine goodness (sorry Ken, lol) over in the new FEX:

No apologies needed, Gil. I'm glad it gave some people a laugh. =)

I wanted to throw my 2 cents in on the new ranking system. I like the new change that uses downloads from the last 30 days. It allows newbie submitters to feel like they have a better chance at recognition (which, let's face it, everyone likes).

However, I think it could be improved. The focus of the ranking is still on quantity, not necessarily quality. I think it could be improved by incorporating the user ratings ("stars"). Perhaps it could be implemented this way: for each submission, compute a weighted score of:

score = (30-day downloads) *(average "stars")/5

Then sum the scores across all the users files to get their total ranking score. A 5-star file would have 100% of it's downloads counted towards the ranking, while a 1-star would only have 20% of it's downloads counted. For files with no stars (i.e. unranked), you could just consider it as a 5-star by default, or just not count those downloads at all.

I doubt many rankings would change as a result of using this new system, except for a few cases which I will not identify here... with the exception of myself! MY rating will probably drop using this new system, thanks to one of my more controversial submissions which Gil alluded to =). And yet I still think it is a good idea to change to a ranking system that weighs both quantity and quality.

Maybe that was a little more like 5 cents than 2 cents =).

"Kenneth Eaton" <Kenneth.dot.Eaton@cchmc.dot.org> wrote in message <gecu41$orj$1@fred.mathworks.com>...
> "Gilbert " <no@spam.com> wrote in message <gecnju$j2t$1@fred.mathworks.com>...
> > Yep, everything is rainbows and happy sunshine goodness (sorry Ken, lol) over in the new FEX:
>
> No apologies needed, Gil. I'm glad it gave some people a laugh. =)
>
> I wanted to throw my 2 cents in on the new ranking system. I like the new change that uses downloads from the last 30 days. It allows newbie submitters to feel like they have a better chance at recognition (which, let's face it, everyone likes).
>
> However, I think it could be improved. The focus of the ranking is still on quantity, not necessarily quality. I think it could be improved by incorporating the user ratings ("stars"). Perhaps it could be implemented this way: for each submission, compute a weighted score of:
>
> score = (30-day downloads) *(average "stars")/5
>
> Then sum the scores across all the users files to get their total ranking score. A 5-star file would have 100% of it's downloads counted towards the ranking, while a 1-star would only have 20% of it's downloads counted. For files with no stars (i.e. unranked), you could just consider it as a 5-star by default, or just not count those downloads at all.
>
> I doubt many rankings would change as a result of using this new system, except for a few cases which I will not identify here... with the exception of myself! MY rating will probably drop using this new system, thanks to one of my more controversial submissions which Gil alluded to =). And yet I still think it is a good idea to change to a ranking system that weighs both quantity and quality.
>
> Maybe that was a little more like 5 cents than 2 cents =).

Agree complelely. One problem that TMW seems unable to recognize is that the author in question is running around the FEX giving retaliatory ratings to established authors who have offered a revew/rating. It is unacceptable. It seems to be taboo for discussion. I will ensure it is discussed regardless.

Another unaddressed problem is equally unpleasant but must be considered. The author in question is racking up very large download numbers for files that don't work, are very poorly coded etc. Anyone keeping an eye on this fellow can see that the unreasonable download numbers occuring overnight don't seem legitimate. TMW seems to be carefully looking the other direction.

The new FEX design is flawed and should not have been rolled out until more thought was given to various FEX mechanisms. If you are going to maintain a file archive it should be done correctly or not at all. Peer review should be anonymous. Duh.

"Kenneth Eaton" <Kenneth.dot.Eaton@cchmc.dot.org> wrote in message <gecu41$orj$1@fred.mathworks.com>...
> "Gilbert " <no@spam.com> wrote in message <gecnju$j2t$1@fred.mathworks.com>...
> > Yep, everything is rainbows and happy sunshine goodness (sorry Ken, lol) over in the new FEX:
>
> No apologies needed, Gil. I'm glad it gave some people a laugh. =)
>
> I wanted to throw my 2 cents in on the new ranking system. I like the new change that uses downloads from the last 30 days. It allows newbie submitters to feel like they have a better chance at recognition (which, let's face it, everyone likes).
>
> However, I think it could be improved. The focus of the ranking is still on quantity, not necessarily quality. I think it could be improved by incorporating the user ratings ("stars"). Perhaps it could be implemented this way: for each submission, compute a weighted score of:
>
> score = (30-day downloads) *(average "stars")/5
>
> Then sum the scores across all the users files to get their total ranking score. A 5-star file would have 100% of it's downloads counted towards the ranking, while a 1-star would only have 20% of it's downloads counted. For files with no stars (i.e. unranked), you could just consider it as a 5-star by default, or just not count those downloads at all.
>
> I doubt many rankings would change as a result of using this new system, except for a few cases which I will not identify here... with the exception of myself! MY rating will probably drop using this new system, thanks to one of my more controversial submissions which Gil alluded to =). And yet I still think it is a good idea to change to a ranking system that weighs both quantity and quality.
>
> Maybe that was a little more like 5 cents than 2 cents =).

Agree complelely. One problem that TMW seems unable to recognize is that the author in question is running around the FEX giving retaliatory ratings to established authors who have offered a revew/rating. It is unacceptable. It seems to be taboo for discussion. I will ensure it is discussed regardless.

Another unaddressed problem is equally unpleasant but must be considered. The author in question is racking up very large download numbers for files that don't work, are very poorly coded etc. Anyone keeping an eye on this fellow can see that the unreasonable download numbers occuring overnight don't seem legitimate. TMW seems to be carefully looking the other direction.

The new FEX design is flawed and should not have been rolled out until more thought was given to various FEX mechanisms. If you are going to maintain a file archive it should be done correctly or not at all. Peer review should be anonymous. Duh.

In article <gecu41$orj$1@fred.mathworks.com>,
 "Kenneth Eaton" <Kenneth.dot.Eaton@cchmc.dot.org> wrote:

> "Gilbert " <no@spam.com> wrote in message
> <gecnju$j2t$1@fred.mathworks.com>...
> > Yep, everything is rainbows and happy sunshine goodness (sorry Ken, lol)
> > over in the new FEX:
>
> No apologies needed, Gil. I'm glad it gave some people a laugh. =)

Count me among those who appreciated the happy sunshine goodness comment.


> I wanted to throw my 2 cents in on the new ranking system. I like the new
> change that uses downloads from the last 30 days. It allows newbie submitters
> to feel like they have a better chance at recognition (which, let's face it,
> everyone likes).

I agree, though we might argue over an appropriate time span. I'm
thinking something a little longer might be better -- maybe 6 months or
a year, but it's not that important.


> However, I think it could be improved. The focus of the ranking is still on
> quantity, not necessarily quality. I think it could be improved by
> incorporating the user ratings ("stars"). Perhaps it could be implemented
> this way: for each submission, compute a weighted score of:
>
> score = (30-day downloads) *(average "stars")/5
>
> Then sum the scores across all the users files to get their total ranking
> score. A 5-star file would have 100% of it's downloads counted towards the
> ranking, while a 1-star would only have 20% of it's downloads counted. For
> files with no stars (i.e. unranked), you could just consider it as a 5-star
> by default, or just not count those downloads at all.

I don't agree with this, however. The vast majority of submissions have
very few rankings compared to the number of downloads. We simply have
no idea what most people think of any given submission because they
haven't rated it. I don't think it's reasonable to extrapolate the
thinking of a few. It makes one vote count too much.

The mere fact of many downloads is itself an indication of quality. The
stars are there to help people decide whether they want to bother with
that download so they certainly already affect the ranking. Let's not
let them count twice. Just my opinion.

--
Doug Schwarz
dmschwarz&ieee,org
Make obvious changes to get real email address.

Doug Schwarz <see@sig.for.address.edu> wrote in message <see-27D435.14493130102008@news.frontiernet.net>...
> The vast majority of submissions have
> very few rankings compared to the number of downloads. We simply have
> no idea what most people think of any given submission because they
> haven't rated it. I don't think it's reasonable to extrapolate the
> thinking of a few. It makes one vote count too much.
>
> The mere fact of many downloads is itself an indication of quality. The
> stars are there to help people decide whether they want to bother with
> that download so they certainly already affect the ranking. Let's not
> let them count twice. Just my opinion.

You've made some very good points here, Doug. The fact that people download files much more than they rate files does complicate matters. I've often wondered if there are ways to encourage more review/rating of files, since so many of the FEX submissions are not rated at all. On other public forums I've been on, you can actually boost your forum ranking by relatively small amounts by supplying ratings to other users. Maybe that system would work here?

Kenneth Eaton wrote:
> Doug Schwarz <see@sig.for.address.edu> wrote in message <see-27D435.14493130102008@news.frontiernet.net>...
>> The vast majority of submissions have
>> very few rankings compared to the number of downloads. We simply have
>> no idea what most people think of any given submission because they
>> haven't rated it. I don't think it's reasonable to extrapolate the
>> thinking of a few. It makes one vote count too much.
>>
>> The mere fact of many downloads is itself an indication of quality. The
>> stars are there to help people decide whether they want to bother with
>> that download so they certainly already affect the ranking. Let's not
>> let them count twice. Just my opinion.
>
> You've made some very good points here, Doug. The fact that people
> download files much more than they rate files does complicate
> matters. I've often wondered if there are ways to encourage more
> review/rating of files, since so many of the FEX submissions are not
> rated at all. On other public forums I've been on, you can actually
> boost your forum ranking by relatively small amounts by supplying
> ratings to other users. Maybe that system would work here?

But, who would rate the raters?

Unless the whole exercise is to be moderated it relies on fair and
honest players.

Failing that, ratings and even numbers of downloads as statistics of
quality are capricious at best.

Unfortunately, it only takes one or a few rogues to create havoc for the
rest. :(

--

I love the new design! Looks so... scientific!

"Gilbert " <no@spam.com> wrote in message <gecumq$4rd$1@fred.mathworks.com>...
> Agree complelely. One problem that TMW seems unable to recognize is that the author in question is running around the FEX giving retaliatory ratings to established authors who have offered a revew/rating. It is unacceptable. It seems to be taboo for discussion. I will ensure it is discussed regardless.
>
> Another unaddressed problem is equally unpleasant but must be considered. The author in question is racking up very large download numbers for files that don't work, are very poorly coded etc. Anyone keeping an eye on this fellow can see that the unreasonable download numbers occuring overnight don't seem legitimate. TMW seems to be carefully looking the other direction.
>
> The new FEX design is flawed and should not have been rolled out until more thought was given to various FEX mechanisms. If you are going to maintain a file archive it should be done correctly or not at all. Peer review should be anonymous. Duh.


A problem with anonymous peer reviews is
they only work if the reviewers are themselves
honest.

Anonymously, one might be willing to spam
inappropriately good or bad reviews at some
favorite target. But if your name is attached
to the review, then it matters, at least if you
care about your good name. There is at least
one author on the FEX that cares not the least
about his name.

A year or so ago, I floated an idea to use the
concept of karma on Matlab central. Karma
could be a measure of your good citizenship
on Matlab Central. So if you submit files to the
file exchange, you gain karma. If those files
are well liked and get good reviews, you get
more karma yet. If they get roundly booed by
all, then you lose karma. Likewise, your own
comments and reviews can earn you karma
or cost karma, because they too will earn
agreement or disagreement.

What good is karma? (By the way, I've loosely
stolen the karma idea from slashdot.com.)
You get a daily ration of mod points, that you
can apply to any files or reviews that you see.
High karma individuals get more mod points.
Low karma individuals get few mod points.

Doing things that show you to be a good
citizen on Matlab Central would be rewarded.

John

"John D'Errico" <woodchips@rochester.rr.com> wrote in message <geer9h$gqt$1@fred.mathworks.com>...
> A problem with anonymous peer reviews is
> they only work if the reviewers are themselves
> honest.
>
> Anonymously, one might be willing to spam
> inappropriately good or bad reviews at some
> favorite target. But if your name is attached
> to the review, then it matters, at least if you
> care about your good name. There is at least
> one author on the FEX that cares not the least
> about his name.
>
> A year or so ago, I floated an idea to use the
> concept of karma on Matlab central. Karma
> could be a measure of your good citizenship
> on Matlab Central. So if you submit files to the
> file exchange, you gain karma. If those files
> are well liked and get good reviews, you get
> more karma yet. If they get roundly booed by
> all, then you lose karma. Likewise, your own
> comments and reviews can earn you karma
> or cost karma, because they too will earn
> agreement or disagreement.
>
> What good is karma? (By the way, I've loosely
> stolen the karma idea from slashdot.com.)
> You get a daily ration of mod points, that you
> can apply to any files or reviews that you see.
> High karma individuals get more mod points.
> Low karma individuals get few mod points.
>
> Doing things that show you to be a good
> citizen on Matlab Central would be rewarded.
>
> John

Interesting but I'm sure it is academic. This author is evidently not subject to peer review by anyone regardless of qualifications. It seems TMW is satisfied with the current state of affairs in the FEX and is not going make any specific changes that address this. My guess is that TMW considers any negative ratings to be the problem and not this author. I find it unbelievable that TMW would stand by while a neophyte author sabotages the file collections of highly regarded mathematicians and engineers. I'll bet they think that after a certain amount of "dust-up" an equilibrium of sorts will be achieved. What TMW has really done here is they have killed the rating system. They don't want any negative ratings at all because they have a happy utopian view of what the FEX should be with nobody offering criticism etc. This type of thinking is familiar to me and I recognize its origin.

It’s a disappointment to me because I rather enjoyed the fiddling pleasure of seeing a few people download and use my code. I felt a certain balance between taking from the FEX and then subsequently contributing to the FEX in equal measure. I took pride in my ratings/reviews and signed them with my name. Often I restricted my ratings to those that were positive, merely offering a rating for code that I liked.

That has all changed. Now I will take but will not give. I will rate but will not be rated. I will regard TMW with deserved suspicion. I will seek alternative technical solutions at work where possible.

How is that for utopia?

"a programmer" <no@spam.com> wrote in message <geffrm$ppc$1@fred.mathworks.com>...
 
> What TMW has really done here is they have killed the rating system.
 
Couldn't agree more.
 
Besides, the new interface is only for astronomers. We need a telescope (or is it a microscope?) to read it.