Photo rating

We have added the ability for users to rate photographs. This is available from the Add/Edit menu (top navigation bar).

You will be presented with 10 photos of a mineral at a locality. Please select the two photos that you like the best. If there is only one or you don't like any of the photos you can select just one or none before you submit your vote.

Currently, only the locality/mineral pairs with greater than 10 photos will be available for voting (~1500). Right now this voting will determine which photos are displayed on the detailed locality page and the mineral/locality page. It also will affect the order the photos are shown on the mineral/locality gallery pages. It is anticipated that this will be used to determine which photos are shown in the locality and mineral galleries to keep down the total number of photos shown (see discussion of this problem in the general section of the messageboard)
[www.mindat.org]
when the gallery display program is updated. We do need your participation to help enhance your user experience at mindat.org.

i assume the extras -lower rated - photos will still be accesable, perhaps as 'supplimental images?'

Yes, all photos will either be available through the locality/mineral gallery and/or we may have mineral and locality galleries that show all photos.

A very good idea!

I've noticed that mainly common and rather pretty (-; minerals are present within this new ability.
I think it would be interesting to vote on photos of rare minerals, too (even if we could choose 1 per, let's say, 5 (-
It's often difficult to photograph well these small crystals or crude masses (-;

Thank You for this new add !

Great idea, and much needed.

Can you provide more details on how this works. How are pictures ranked based on ratings, what happens if new images are added, can we see the ratings, etc...

thanks, Jeff

The best photos are determined by manager rating, user voting, number of times viewed. If there are ties at any level, the next lower level determines ranking.

Older photos have an advantage (in that they will be accessed by search engines). Right now we have something like 2200+ photos that have been ranked by managers. Periodically we check the photo database to see if there are minerals that have not been ranked that have had a number of photos added to the database (ie see if the head photo should be changed).

You can see how various photos rank now on the detailed mineral/locality gallery (but not the raw numbers). The mineral and photo galleries do need an updating to take into account this new system (and probably some improvements in the way things are displayed). Additional expansions in this area are going to depend upon the response of the mindat.org community in using these options.

To speed up things, could "select as worst" buttons be added, to choose the 2 worst photos ? I did a few ratings, and I think it could be useful to preserve that information too.

It seems to work really well! The way the "select worst" button should work is to send the photo to the bottom of the gallery.

Maybe the rating page should have an indication to base the decisions on the full image and not on the thumbnail, as many out of focus photos look acceptable as thumbnails.

I've been watching the user photo ratings page at regular intervals. At least one person is rating so fast that I suspect the ratings can only be based on the thumbnails, not on the full picture.

The problem with quick ratings (estimate: 300 in less than 2 hours) is that beeing based on the thumbnails alone, they well be of lesser quality than well pondered ratings based on the photos themselves. But they will inevitably outnumber the careful ratings. I don't know if it is just the novelty, but there seems to be potential for addiction in this rating system. Maybe it would be necessary to weight the vote with a non-linear function of the time spent between presentation of the picture set and casting of the votes, or whether pictures have been opened or not. Otherwise we'll be drowned by the eager beavers.

Agreed, for a mineral lover this is akin to a video game. I for one have made sure to view the full image out of a short list of thumbnails which look good - it's not uncommon for the full pic to be only a little larger than the thumbnail, and these are quickly disposed of. As Dominik says, often the full pic reveals a fuzzy image. At least half of the time, I've selected "none", thank you.

Perhaps a set of recommended criteria (each will rank these according to their inclinations) could be posted, and in some cases, such as for rare species, criteria would vary:

-specimen quality
-shows representative features of the locale/species, such as associations, crystal form & color, etc.
-image quality - size matters, as this shows details; also color accuracy, etc. etc.
-technical quality - good lighting, appropriate background (NOT THE FAKE HALOES!! YUCK!!)

Baed on these, I have found myself selecting a highly disproportionate number of Lavinsky/iRocks photos, and a significant number from Fabre, plus of course the usual brilliant micromount artists. NO HALOES, PLEASE!!!

And despite having looked at about 400-500 sets, I've yet to be presented with one of my own images.....that might be a coincidence, or it might be that the system automatically leaves out images from the logged user. If so, it's not a bad thing, and I don't mind....but I'd be curious to know if that's the case.

In any event, the system devised is a very good one. If the "best" photo is subject to constant updating based on new voting, then new images will eventually make there way to their appropriate level.

Tony

Yes, you don't get to vote on your own photos. I thought that this would be fairer and would head off a possible problem of someone writing a robot to go through and pick their own photos to raise their profile.

I was rating the pictures last night & I too found that you can discard a few of the thumbnails at a glance & then open up the remainder of which most were represented by 'irocks' seeing as they are the highest contributor its not surprising! but I did get put off on voting all the time for this same contributor

I also rated 3 in a set hope that counted ok ?

Overall the pictures are very good quality & I enjoyed them immensely.

What happens with the child photos (which aren't visible in the galleries). Does the vote go to the parent ?
Also, it seems one can vote multiple times on the same set by going back with the browser. I don't know if the votes get counted, but at least the user vote counter is incremented.

Actually only the first two counted.

Hopefully we will have some other contenders in the future.

Updated the program to not pick child photos. Thanks for the heads up on that. Will have to look at the back button problem.

I'm using explorer & find the back button gives a new set of pictures !

Is it possible to have a next or skip button ?



Edited 1 time(s). Last edit at 02/16/2008 02:42PM by Debbie Woolf.

On the final thank you page (after you have selected to vote) there is "Click here to continue to rate photos" which, if you click it, will bring up another set of photos to rate. (If you don't like any on the current page, just vote with no photos selected.)



Edited 1 time(s). Last edit at 02/16/2008 02:45PM by David Von Bargen.

Yes I see that its just I'm going through the sets looking at the regions I know first (just my preference) so skipping a lot & thought these buttons could make the process quicker/easier.

The back button goes back to the same page only with the opera browser (they are really aggressive with the cacheing). It isn't a problem with IE or Firefox.

Right now we have about 3000 photos that have been voted for with about 350 being picked more than once.

If you don't want to vote on a particular locality/mineral, you can also hit your browser's refresh button to bring up a new selection.



Edited 2 time(s). Last edit at 02/16/2008 04:49PM by David Von Bargen.

Hmmmmm....I think it's a mistake to reject child photos. I have gone through some sets in previous days that were composed in part of photos of one specimen. These are from different angles/closeups, etc. I appreciated being able to judge which view was best.

I usually put up images with at least one child photo and I often debate with myself which to make parent/child. But opinions differ; one person may think the overall view should be the parent, another that the best closeup should be parent. May I suggest that the solution is, not to make it impossible to judge child photos, but to allow only one of a family to be in the "best" list. Then if a user chooses to look at that "best" photo, whether parent or not, the others will be available.

Sadly, this will require more programming by someone :-)

tdp

Aren't we getting back to the old question of 'what are we rating' here? Should we choose for the best photo, the best specimen, or the best representive specimen for the location?...
No easy answer.

Of course I adore just looking at pretty pictures, but the main use of mindat, for me then, is to have a specimen from a certain locality and looking to ID minerals, so I would like representive specimens come up first, not the most rare or pretty one...

Greetings from Antwerp,

Eddy

Just a caveat, on the high number of votes (as I seem to be at the top of the list), the rating numbers includes those for which no photos were selected/chosen. As they come up randomly, I chose to vote only on photos of species/localities which I felt familiar enough with to make a sound judgement, so that higher number also represents those which were passed by without any selection being made.

I did indeed look at the full sized image of all which I voted on, unless a photo came up that I was already quite familiar with, so no need to get your panties in a twist, the system is not being abused (at least not by me).

I do welcome this effort, as I have for a long time had problems surfing through all the photos of species/locales which are the most popularly posted (i.e. the spanish fluorites etc.).

Mark Heintzelman

Eddy, I think there are two independent aspects.
I think there is broad agreement that large, sharp, well lit and composed photos are "better" than small, out of focus photos.

But given the variety of specimens, we will get a mix of the tastes of the voters: representative, bizarre or flashy, whatever...

But maybe we could agree that the photo quality criterion should take precedence over the specimen criterion. If one has to choose between a poor quality photo of a museum piece and high quality photo of a representative specimen, one should temporarily set aside one's preference for jaw-droppers and simply vote for the better photo.

Mark,
It wasn't you I had in mind, but you are right, that could well be the explanation. I also found myself not voting rather often, and they get counted.
Dominik

Can I also suggest that people do not rate highly photos which have overzealous dealer descriptions!

Jolyon

It isn't too likely that we will be able to display a child photo rather than the parent photo in a gallery. Not only would that take some programming, it would also put a fairly decent hit on server performance. It will probably be simpler to get the photographers to change the parent photo.

There is a twofold aim to this process, to pick the "best" photo of a species from a locality (for display on the detailed locality page) and also to cut down on the number of similar photos displayed on a mineral species gallery or a locality gallery. For the gallery displays, we would like to get the widest range of colors and habits for a mineral at a particular locality.

For a description I don't neccesarily think a flowery description is much worse than a description of "2x3cm".

Just a technical question on the rating system: do the photos get an absolute number of points when voted for, or do they get a number of points relative to the set in which they were presented ? I think the later would be closer to the reality of the assessement, which is to determine the best photos in the presented set. Although the possibility not to vote when no presented photo stands out is more of an absolute decision. I'm asking this because I wonder how new entries will be treated. Will they have to accumulate points to match the point level accumulated by older photos, or could they bubble to the top quickly by dominating a given random set, and moving up multiple places at once ?
In a way, the rating system could be seen as a sorting algorithm with an outsourced human comparison operator. A modified version would be to present two photos, and ask the user which one he thinks is better.
kind regards,
Dominik

Right now it is a simple one vote, one increment of the "best" counter. If this works out, a more sophisticated system could be implemented.

We have had 8600 votes for mineral photos so far. 4800 photos have received 1 vote, 1000 have 2 votes and 350 have 3 or more votes.

12K photos have received votes (from ~30K total photos), 120 have received 6 or more votes, 850 have 4+, 202 3+ and 4800 2+ votes.

The locality and mineral galleries now have just the top 10 photos displayed for a mineral/locality combination. You can display all of the photos with that combination by clicking on the "Additional photos" link (in green after the mineral names on the 10th photo).

Thanks to everyone that has participated so far on rating the photos.

Hi,

One thing just came to my mind....
I see loads of photographs being posted each day, the new photos appear to show at the bottom of the localities photo page i.e. they didn't recieve any votes when the locality was the subject of voting for the best photos.
This can be quite a drawback as some very good new photographs should definitely be listed on the top of the page.
I realize that there can be no continuous voting, so I think that Dominik's suggestion to rather vote for the worst photographs would work better imho....

Cheers

Harjo