Thread Subject: ML spring 09 contest is bothersome and offensive

i am well aware that i am not going to make a lot of friends with this post...

however, i am extremely annoyed and - offended - by the blueprint of this contest and cannot but utter the utmost disappointment:

1) this is a mere (and scientifically as well as conceptually completely useless) beauty contest; apparently due to some hardware not up and running (WHAT! we are talking TMW...); so, why bother...

2) the FEX is heavily abused(!) to serve as a (convenient...) vehicle to convey information, which is in NO way asked for and clearly shall be perceived as an impertinent intrusion by the company…

i shall referee all such submissions with this one-star comment:

this submission is a) NOT suitable for and b) clearly does NOT live up to the goals and expectations of the FEX by which other authors are relentlessly judged by fellow MLers...
frankly, submitting stuff related to a (rather insignificant) ML-beauty contest to the FEX is an abuse of this (already heavily) taxed site and should be moved to another place...
please remove it...
urs

just my anger...
us

"us " <us@neurol.unizh.ch> wrote in message <gr137d$lts$1@fred.mathworks.com>...
> i am well aware that i am not going to make a lot of friends with this post...


Well when do we ever make friends by posting to this newsgroup! ;)


> however, i am extremely annoyed and - offended - by the blueprint of this contest and cannot but utter the utmost disappointment:
>
> 1) this is a mere (and scientifically as well as conceptually completely useless) beauty contest; apparently due to some hardware not up and running (WHAT! we are talking TMW...); so, why bother...


I agree about the beauty contest bit, and the value assessment.


> 2) the FEX is heavily abused(!) to serve as a (convenient...) vehicle to convey information, which is in NO way asked for and clearly shall be perceived as an impertinent intrusion by the company…


If we can judge from the length of the threads generated on the NG due to past contests, the FEX better get a disk upgrade.


> i shall referee all such submissions with this one-star comment:
>
> this submission is a) NOT suitable for and b) clearly does NOT live up to the goals and expectations of the FEX by which other authors are relentlessly judged by fellow MLers...
> frankly, submitting stuff related to a (rather insignificant) ML-beauty contest to the FEX is an abuse of this (already heavily) taxed site and should be moved to another place...
> please remove it...
> urs
>
> just my anger...
> us


Understandable, but it doesn't get me angry. It is annoying, but I have long since learned that the FEX is not the resource it could be and there is nothing that can be done about it (by us, the 'regular' ML users). It has changed from the nice, uncrowded and pretty beach it was 5 years ago to the Jersey Shore of today. It can be useful, but it takes a lot more work to find quality.

My .02.

"us " <us@neurol.unizh.ch> wrote in message <gr137d$lts$1@fred.mathworks.com>...
> i am well aware that i am not going to make a lot of friends with this post...
>
> however, i am extremely annoyed and - offended - by the blueprint of this contest and cannot but utter the utmost disappointment:
>
> 1) this is a mere (and scientifically as well as conceptually completely useless) beauty contest; apparently due to some hardware not up and running (WHAT! we are talking TMW...); so, why bother...
>
> 2) the FEX is heavily abused(!) to serve as a (convenient...) vehicle to convey information, which is in NO way asked for and clearly shall be perceived as an impertinent intrusion by the company…
>
> i shall referee all such submissions with this one-star comment:
>
> this submission is a) NOT suitable for and b) clearly does NOT live up to the goals and expectations of the FEX by which other authors are relentlessly judged by fellow MLers...
> frankly, submitting stuff related to a (rather insignificant) ML-beauty contest to the FEX is an abuse of this (already heavily) taxed site and should be moved to another place...
> please remove it...
> urs
>
> just my anger...
> us

Yes. TMW refuses to moderate the FEX, virtually
encouraging the spammers to dump their dreck
upon it, both as files and as "ratings". At most
they will send a quiet e-mail to those same
spammers, asking them to improve their behavior.

Worse, they are now encouraging hundreds
of new pieces of useless dreck to clutter it up.

I'm expecting to follow Duane soon now. It is
their FEX, and they can do as they wish.

What is sad is the FEX could have been a good
repository for code.

> i shall referee all such submissions with this one-star comment:
>
A reminder to everyone in this community that File Exchange comments should be focused on the code, and should help the submitter improve their MATLAB coding skills. We do review all comments posted to File Exchange. We remove any comments that do not meet this criteria as being off-topic.

This guideline was established following requests from our community members. It is intended to keep the discussions focused and professional.

Sincerely,
Helen Chen
MATLAB Central Administration

"Helen Chen" <helen.chen@mathworks.com> wrote in message <gr30tp$o5s$1@fred.mathworks.com>...
> > i shall referee all such submissions with this one-star comment:
> >
> A reminder to everyone in this community that File Exchange comments should be focused on the code, and should help the submitter improve their MATLAB coding skills. We do review all comments posted to File Exchange. We remove any comments that do not meet this criteria as being off-topic.

Since you do NOT review submissions to the FEX, why review comments? Many of the FEX submissions are orders of magnitude worse than the comments.

>
> This guideline was established following requests from our community members. It is intended to keep the discussions focused and professional.

And what your "community-members" requests to keep submissions "focused and professional" rather than the "pure dreck" submissions as John D. calls them?
>
> Sincerely,
> Helen Chen
> MATLAB Central Administration

> Yes. TMW refuses to moderate the FEX, virtually
> encouraging the spammers to dump their dreck
> upon it, both as files and as "ratings". At most
> they will send a quiet e-mail to those same
> spammers, asking them to improve their behavior.
>
> Worse, they are now encouraging hundreds
> of new pieces of useless dreck to clutter it up.
>
> I'm expecting to follow Duane soon now. It is
> their FEX, and they can do as they wish.
>
> What is sad is the FEX could have been a good
> repository for code.

I concur completely. John speaks with wisdom.

Can someone set up a simple moderated FEX somewhere? Preferably one that only lets the moderators comment on the code. Allowing public comments always brings out the lowest common denominator. I will gladly contribute my time as a moderator.

The site need only have a few categories and allow some searching. The presented description of each submission could/should be the help text at the top of the submission, i.e., >>help filename output. In that way good help text conforming to MATLAB formatting would be required and would be consistent across all submissions.

This site would not have thousands of submissions, but rather no more than a couple hundred of the most useful. We don't need a dozen submissions that perform Gaussian elimination with or without pivoting.

Duane

"Duane Hanselman" <masteringmatlab@yahoo.spam.com> wrote in message <grd4j9$ifr$1@fred.mathworks.com>...
> > Yes. TMW refuses to moderate the FEX, virtually
> > encouraging the spammers to dump their dreck
> > upon it, both as files and as "ratings". At most
> > they will send a quiet e-mail to those same
> > spammers, asking them to improve their behavior.
> >
> > Worse, they are now encouraging hundreds
> > of new pieces of useless dreck to clutter it up.
> >
> > I'm expecting to follow Duane soon now. It is
> > their FEX, and they can do as they wish.
> >
> > What is sad is the FEX could have been a good
> > repository for code.
>
> I concur completely. John speaks with wisdom.
>
> Can someone set up a simple moderated FEX somewhere? Preferably one that only lets the moderators comment on the code. Allowing public comments always brings out the lowest common denominator. I will gladly contribute my time as a moderator.
>
> The site need only have a few categories and allow some searching. The presented description of each submission could/should be the help text at the top of the submission, i.e., >>help filename output. In that way good help text conforming to MATLAB formatting would be required and would be consistent across all submissions.
>
> This site would not have thousands of submissions, but rather no more than a couple hundred of the most useful. We don't need a dozen submissions that perform Gaussian elimination with or without pivoting.
>
> Duane

Duane,

Sounds a good idea. If such a site can be set up and run, I wish to join. I am fed up with the downloading and rating wars in the FEX.

Yi

> >
> > I concur completely. John speaks with wisdom.
> >
> > Can someone set up a simple moderated FEX somewhere? Preferably one that only lets the moderators comment on the code. Allowing public comments always brings out the lowest common denominator. I will gladly contribute my time as a moderator.
> >
> > The site need only have a few categories and allow some searching. The presented description of each submission could/should be the help text at the top of the submission, i.e., >>help filename output. In that way good help text conforming to MATLAB formatting would be required and would be consistent across all submissions.
> >
> > This site would not have thousands of submissions, but rather no more than a couple hundred of the most useful. We don't need a dozen submissions that perform Gaussian elimination with or without pivoting.
> >
> > Duane
>
> Duane,
>
> Sounds a good idea. If such a site can be set up and run, I wish to join. I am fed up with the downloading and rating wars in the FEX.
>
> Yi

I have free webspace available, but I do not have the software to manage the site. If someone had the capability to do that, we would be all set.

Duane

"Duane Hanselman" <masteringmatlab@yahoo.spam.com> wrote in message <grd6u9$k1h$1@fred.mathworks.com>...
> > >
> > > I concur completely. John speaks with wisdom.
> > >
> > > Can someone set up a simple moderated FEX somewhere? Preferably one that only lets the moderators comment on the code. Allowing public comments always brings out the lowest common denominator. I will gladly contribute my time as a moderator.
> > >
> > > The site need only have a few categories and allow some searching. The presented description of each submission could/should be the help text at the top of the submission, i.e., >>help filename output. In that way good help text conforming to MATLAB formatting would be required and would be consistent across all submissions.
> > >
> > > This site would not have thousands of submissions, but rather no more than a couple hundred of the most useful. We don't need a dozen submissions that perform Gaussian elimination with or without pivoting.
> > >
> > > Duane
> >
> > Duane,
> >
> > Sounds a good idea. If such a site can be set up and run, I wish to join. I am fed up with the downloading and rating wars in the FEX.
> >
> > Yi
>
> I have free webspace available, but I do not have the software to manage the site. If someone had the capability to do that, we would be all set.
>
> Duane

I am a relatively new Matlab user and have found the FEX to be an invaluable tool in analyzing existing data and learning to develop code. I can understand why the experienced users of the board (people I am indebted to) have become frustrated with some of the unnecessary and distracting posts that have become more prevalent (over the last months, it seems). But my request would be that you don't throw the baby out with the bathwater. There are many newbies who find the FEX helpful and many of those will go on to become the next senior contributors - closing the door to comments or code from those people will be counter-productive.

As for the contest, surely the Matlab team could find a separate but visible portal for it, so that people who are interested can look, and those that aren't can ignore. Also, my opinion would be that if you feel that something/someone is simply trying to attract attention, then your best tactic is to deny it/them the attention they desire - continual scoring or comments simply launches the post onto the front page of Matlab Central.

With best intentions,

Mike

Dear all,

> > > Can someone set up a simple moderated FEX somewhere? Preferably one that only lets the moderators comment on the code. Allowing public comments always brings out the lowest common denominator. I will gladly contribute my time as a moderator.
> > >
> > > The site need only have a few categories and allow some searching. The presented description of each submission could/should be the help text at the top of the submission, i.e., >>help filename output. In that way good help text conforming to MATLAB formatting would be required and would be consistent across all submissions.
> > >

>
> I have free webspace available, but I do not have the software to manage the site. If someone had the capability to do that, we would be all set.

Before seeing this thread, I basically came to the same conclusion but looked towards sourceforge et al., see f.ex.
http://en.wikipedia.org/wiki/Comparison_of_free_software_hosting_facilities

Best,
Lars

 

 

i own the domain openkoderz.org do you want me to setup a managed forum? i would gladly do it if there are people willing to moderate it. i have bandwidth and the domain but not time to moderate it. if you wish just to help send me an email to icaoberg@gmail.com

ivan

 
> I have free webspace available, but I do not have the software to manage the site. If someone had the capability to do that, we would be all set.
>
> Duane

I wish you all the luck with that initiative.
If you go through wit it I'll resubmit my (modest) contributions that meanwhile I have also removed probably because I'm less patient and got fed up with TMW attitude already sometime ago.

J. Luis

"Ivan E. Cao-Berg" <icaoberg@gmail.com> wrote in message <grdm3v$rp5$1@fred.mathworks.com>...
> i own the domain openkoderz.org do you want me to setup a managed forum? i would gladly do it if there are people willing to moderate it. i have bandwidth and the domain but not time to moderate it. if you wish just to help send me an email to icaoberg@gmail.com
>
> ivan

I am willing to act as one of the moderators. John D., Jos, Urs, and others will probably volunteer as well. There was a group of volunteer folks who worked on with TMW in weeding out dreck FEX submissions and in giving special "Select" status to some. I no longer know how to get in touch with that group, but they would probably all volunteer (except for TMW employees, that is.)

Duane

I find it really a pity that making such a spin-off to the FEX has become necessary, but it clearly would be a great idea.

"us " <us@neurol.unizh.ch> wrote in message
news:gr137d$lts$1@fred.mathworks.com...
>i am well aware that i am not going to make a lot of friends with this
>post...
>
> however, i am extremely annoyed and - offended - by the blueprint of this
> contest and cannot but utter the utmost disappointment:
>
> 1) this is a mere (and scientifically as well as conceptually completely
> useless) beauty contest; apparently due to some hardware not up and
> running (WHAT! we are talking TMW...); so, why bother...
>
> 2) the FEX is heavily abused(!) to serve as a (convenient...) vehicle to
> convey information, which is in NO way asked for and clearly shall be
> perceived as an impertinent intrusion by the company…
>
> i shall referee all such submissions with this one-star comment:
>
> this submission is a) NOT suitable for and b) clearly does NOT live up to
> the goals and expectations of the FEX by which other authors are
> relentlessly judged by fellow MLers...
> frankly, submitting stuff related to a (rather insignificant) ML-beauty
> contest to the FEX is an abuse of this (already heavily) taxed site and
> should be moved to another place...
> please remove it...
> urs
>
> just my anger...
> us

I've read through this thread and the other threads about the contest and I
must say I'm not happy with the situation myself. Some of this post may not
seem so relevant at first, but please bear with me.

Here at The MathWorks, one of our core values is titled "Continuous
Improvement and Pursuit of Excellence". You can read a little bit about it
here:

http://www.mathworks.com/company/aboutus/mission_values/values/continuous.html

One of the ways we practice this core value is by taking "at bats" -- trying
new things, with the knowledge that they won't always be right or work out
well. This contest is an "at bat" -- an attempt to come up with a different
type of MATLAB contest. From the reaction of some of the CSSM regulars,
it's obvious that this was in no way a home run. But I'm not sure what it
was, so I'd like to ask a few questions to figure out if it was a single
(good), a foul ball (okay), or a triple play (very bad). Even though I'm
asking them in response to us's post, anyone please feel free to offer your
own opinions.



us, if we had released this contest data and asked people to post their own
analysis on the File Exchange, without putting it in the context of a
contest, would you have been more or less angry than you are now? [Are you
angry that we asked people to post contest entries to the File Exchange?]


If we had released this contest data and someone had posted their analysis
to the File Exchange without anyone asking them to do so, would you have
been more or less angry than you are now? [Are you angry about the fact
that people are posting submissions analyzing a data set to the File
Exchange?]


If we had released this contest data and run a contest where people posted
their analysis to a separate section of MATLAB Central, not the File
Exchange itself, would you have been more or less angry than you are now?


If you have been given this data and asked to design a contest around it,
what type of contest would you have designed?


Should we even hold MATLAB contests? After all, the previous contests were
(in part) an excuse for people to show off their MATLAB coding and tweaking
skills, making it very much like a "beauty contest".


I thought I remembered seeing you answer this question in another thread,
but now I can't find it so I'll ask it again: you used the phrase "goals
and expectations of the FEX" above. What are _your_ "goals and
expectations" of the File Exchange? Is it to be a repository of useful (for
some definition of useful) code for others to use and expand upon? Is it to
teach new users how to use MATLAB? Is it a place for people to post their
code in hopes that others will improve that code? Is it somewhere for
people to post their code to have it reviewed and judged like some sort of
dog show? Or is it something else?



Frankly, I don't see this contest as a "beauty contest". I view it as an
opportunity for people to show off how they actually use MATLAB to analyze
data. Admittedly the data isn't what most of you would work with on a
day-to-day basis, but finding a "real-life" data set that would be
accessible to everyone would be difficult at best. [Although maybe an
expanded version of the census data that ships in the MATLAB demos directory
...] And while the contest entries may not be very useful for you or me,
for someone who's new to using MATLAB for data analysis examples showing
what types of analysis people actually use on a "real" data set could be
useful, and would probably lead them to look at functions and capabilities
they had not used before.

As I said above, I'm not happy that an at-bat like this caused so much anger
and frustration among some of the newsgroup regulars that I consider friends
(even though I've never met most of you in person.) I hope we can find some
way to make the File Exchange, the contest, and all of MATLAB Central be
what you want it to be as well as what we at The MathWorks want it to be (or
some combination of the two.)

As always, this is just my opinion, but I wanted to make that explicitly
clear this time.

--
Steve Lord

"Steven Lord" <slord@mathworks.com> wrote in message <grfsit$84f$1@fred.mathworks.com>...

> As I said above, I'm not happy that an at-bat like this caused so much anger
> and frustration among some of the newsgroup regulars that I consider friends
> (even though I've never met most of you in person.) I hope we can find some
> way to make the File Exchange, the contest, and all of MATLAB Central be
> what you want it to be as well as what we at The MathWorks want it to be (or
> some combination of the two.)

Steve,

I really feel this was the straw that broke the camel's
back. What you see is an outflow of frustration
against spammers like Marco, against volumes of crap
appearing on the FEX, that we cannot rate downwards
without incurring the wrath of a person who does not
care about his own reputation.

It is an unwillingness to see potentially many new
files placed there that have no real value, and a fear
that this establishes a new precedent, accelerating
the influx of files to sort through. A problem with the
tagged FEX as it has been built, is all files are placed
in one large bin. We can use a tag to select some
subset of those files, but that tagged selection is not
terribly accurate. It typically returns many files where
we will have no interest. Often, any search contains
at least a few files by Marco.

There needed to be better tools in place to deal with
the spam and the homework assignments, in place
before this experiment began.

Surely you had to see something like this coming?
There are at least a couple of notable authors who
had already removed their work from the FEX as a
sign of their frustration with the direction the FEX
has taken in the last year.

I'm very much hoping that we can resolve this before
others choose to follow that path.

John

"Steven Lord" <slord@mathworks.com> wrote in message <grfsit$84f$1@fred.mathworks.com>...
>
[snip]
>
> I thought I remembered seeing you answer this question in another thread,
> but now I can't find it so I'll ask it again: you used the phrase "goals
> and expectations of the FEX" above. What are _your_ "goals and
> expectations" of the File Exchange? Is it to be a repository of useful (for
> some definition of useful) code for others to use and expand upon? Is it to
> teach new users how to use MATLAB? Is it a place for people to post their
> code in hopes that others will improve that code? Is it somewhere for
> people to post their code to have it reviewed and judged like some sort of
> dog show? Or is it something else?
>
For what my answer might be worth...

1, My expectations of the file exchange: A repository of useful code, for others to use and improve upon or provide "better than" submissions. Here there is a decided evolutionary ladder, in that I expect newer submissions solving the same task as already existing submissions to be better in some way (speed, range, adjustability or some other feature). When I see another RBF tool, or the next Savitzky-Golay filter I'd want to expect it to be better than the existing. If it would not be the authors would have spent their time better solving another problem, or improved their solution to become better. If TMW provides Delaunay triangulations, J. d'Errico provides gridfit, and others RBF-tools and other fitting routines, I'd use them, maybe improve or tailor, but I wouldn't waste my time writing something that isn't as good.
2, My goals, well since I'm decidedly lazy, my goal is to use tools downloaded from the FEX, since that will save me days and weeks of work, that I now and hopefully in the future spend procrastinating or solving other problems in my field, thereby reaching further and faster than if I had to do it all by hand. In the odd cases where I find that I've solved some problems someone else might find useful I'll try to contribute with those, so that others might benefit in the same way as I do.
3, FEX is not to teach people to program in matlab, it is for others to learn from. There is a significant difference.
4, Improve code - if someone want to improve the code, fine. But there is nothing saying that the improved code will necessary reappear on the FEX. If it does, the code have improved without "my" work, and I'll have a better tool for no extra effort.
5, Dog show, there is no way to stop people from being vain and pathetic. Maybe one could or even should try not to encourage that streak in those inclined.

The reason I get annoyed when I see a "yet another" submission is that the time spent on solving that problem does not contribute to much general benefit to the community. That way the time spent is wasted. It would have been so much preferable to get a submission solving another problem thereby widening the capability and power of the tools available. The submitee cannot argue that a poor "yet another" contribution lead to any more important learning than a good contribution solving an unsolved problem. It is this wastefulness of the poor "yet another" submissions that makes people irritated - the time could have been spent producing something useful for everyone when it was spent on programming anyway.

My 5p
/Bjeorn

> Are you angry that we asked people to post contest entries to the File Exchange?
The File Exchange should not be host to multiple submissions doing the exact same thing, especially coming from the same authors. Also one of the rules of this contest was 'users should not update their files'. This goes directly against the FEX 'ethic' of improving code as suggested by others. I was also unimpressed by the censorship conducted by TMW on comments.

> Are you angry about the fact that people are posting submissions analyzing a data set to the File Exchange?
The File Exchange should be for general files, which are not reliant upon the vagaries of a particular data set in order to work. However some exceptions to this are likely inevitable.

> If we had released this contest data and run a contest where people posted
> their analysis to a separate section of MATLAB Central, not the File
> Exchange itself, would you have been more or less angry than you are now?
I would not have had a problem with this, beyond my earlier criticism that this contest is like homework.

> If you have been given this data and asked to design a contest around it,
> what type of contest would you have designed?
I would have asked myself this question. What is fun about data mining? And then I would know not to hold a contest.
 
> Should we even hold MATLAB contests? After all, the previous contests were
> (in part) an excuse for people to show off their MATLAB coding and tweaking
> skills, making it very much like a "beauty contest".
I have enjoyed the previous contests, both as they took place and just reading the analysis of older ones. I particularly enjoyed the new direction of the last contest. As an 'at bat', that was a home run, though with room to improve. I had hoped that that improvement was going to develop in this contest.

> What are _your_ "goals and expectations" of the File Exchange? Is it to be a >repository of useful (for some definition of useful) code for others to use and
> expand upon?
Yes. I think that two broad categories of files should be supported.
1) Functions which TMW cannot/ have not implemented.
2) Replications of TMW functions for with older versions of ML (e.g bsxfun) or special toolbox functions (e.g vpi by John D.)

A third category would be collections of tips and tricks, such as the Fast Matlab notes by Pascal *Getreuer* or John D's Optimisation guide, both of which I found useful.

> Is it to teach new users how to use MATLAB?
No, buy a book or visit a library (or download Cleve Moler's). If all that fails JFGI.

> Is it a place for people to post their code in hopes that others will improve that code?
No, that is the primary use of the newsgroup threads.

> Is it somewhere for people to post their code to have it reviewed and judged
> like some sort of dog show? Or is it something else?
It should all be quality code performing a worthwhile function. I think we would all like that. The big question is how to achieve that.
 
 
> And while the contest entries may not be very useful for you or me,
> for someone who's new to using MATLAB for data analysis examples showing
> what types of analysis people actually use on a "real" data set could be
> useful, and would probably lead them to look at functions and capabilities
> they had not used before.
I think that the blogs and video tutorials should be a starting point for people wanting to discover better uses of Matlab. The only trace of this contest that I would like to see on the FEX would be a 'tips and tricks' file giving a general overview of visualisation techniques. That one file should be enough for people to find some new capabilities and functions.

JMHO,
Darren