I have been a member of deviantART for over 5 years now, on a career spanning three accounts. Will somebody get a me a damn senior status? If I get any more senior I'll be decaying in my tomb here...
All jokes aside, in the next few journals, I'm posting some reflections on DA, and thoughts I've had over the course of my time using the site.
First and foremost, please understand that I think deviantART is a great site. If this wasn't the case, I wouldn't have stuck around for all these years. I think they have done an immensely good job, and I'm sure the team is made up of some incredible people. I've had the pleasure of corresponding with some of them personally, and I promise you they truly do care about the community and have its best interests at heart, and work hard day in and out to help ensure its continued growth.
I have however, a few observations about the site that I will touch on briefly, one issue in each journal. These are minor problems or observations on a massive system that is without a doubt successful, so I won't be surprised if they are given zero attention and are not thought of as important. These are just things that as a community member, designer, developer, and dA creative wannabe, that I would like to see revisited.
Let me preface with the following: If I have overlooked something obvious, or am pointing out a feature to be added that already exists, forgive my ignorance, and please post in the comments.
So now that we have the above out of the way, let's get on to my inaugural reflection journal, 'deviantART Reflections: Search'.
It can be tough to find what you are looking for on dA.
The problem: This site is HUGE. I don't know the exact number, but I know its over 13 million registered users. There are over 100 million deviations. Therefor, in my opinion, the search tool should be incredibly powerful, with many advanced options, and it's own interface section or page itself. The current search system, in my opinion, is seriously lacking.
How I would approach the issue:I have two alternate approaches to how I would tackle this. I'll describe each.
1) Advanced Search Section / Page
deviantART has some incredibly talented and bright people. But sometimes, looking for inspiration can still be helpful. Why not take a page out of the stock-photography website handbook. Say what you will about stock photo sites, but it is their entire business and livelihood to make it
EASY to find what you're looking for in an image, and do it
fast. The quicker and easier it is for a buyer to find an image on their site, the quicker they get paid, and the more likely the user is to stick around.
So if I was designing a deviantART advanced search system and page, I would do the following:
Integrate a color tag system for images.This could be done one of two ways, either by using the existing tag system to pull out known words for color (ie. red, blue, green) and store them into a deviation color database, or, conversely, add a new feature for color tagging. Allow the user to upload an image, display a color picker, and allow the user to pick a color with an eyedropper. Fancy ajax color pickers are fun, and it would provide a great way to precisely document the color of a deviation. Then, users could search by RBG value or color name to find the deviation that has been tagged with that color. You could even take the third approach and pull the image data, do a little bit of javascript, and create a palette of RBG values and automatically store them, so the user doesn't have to tag anything.
I'd go with the first system, simply because of programming limitations. Although I'm a designer, I am also a developer, and thinking from a programming and usability standpoint, while an ajax color picker tag system would be fun to use, it would likely add to page load time, and
could also require too much database modification or addition to be done without slowing down the query significantly. However, if and when it could be done and done properly, I think it would be a great addition.
If I were to implement this type of solution for color tagging, I'd of course make it an optional feature.
Arguments AgainstThe advanced search page might be more than deviantART is willing to do, for a number of reasons.
It could be that they are trying to keep ease-of-use as a primary design element, or it could be programming limitations.
I think having a color tagging system as an aspect of an 'Advanced Search' system and page would work incredibly, but there are of course limitations. I'm know for a fact that Ryan is 'picky' to say the least when it comes to design. I'm betting the other team members are the same. For this reason, advanced search may not be something that fits into the overall design strategy and architecture for the site. It simply might just be deemed an excessive, unnecessary, or over-the-top feature, which cuts at the heart of any great design. I haven't been into branding for much longer than 2 years, but as I've learned from reading extensively, (books, journals, articles, and the reports on why/when my latest logo was denied entry to Goodlogo); simplicity is elegance. Content may be king, but simplicity is queen. Although, if the only reason for not implementing this is for sake of design simplicity or programming simplicity alone, I would make an appeal to what some have called Einstein's Razor:
Make everything as simple as possible, but not simpler.
Points to anyone who comments and knows what Einstein's Razor has been used to mean, optionally including a reference to Occam's or explanation for those who don't, since I'm trying to keep this as simple as possible and not confuse my massive collection of 1-3 readers.
Aside from the 'Advanced Search System' and page with color tagging, there is another solution I believe would work just as well. This alternative would keep in tact the ease-of-use and instant access to search as it is currently provided.
2) 'Search Filtering'
Keep the UI, Extend the FunctionalityIn this approach, minimal modifications would need to be made to the UI, however extensive functionality could still be provided.
My approach for implementing a search filtering system would be as follows:
Why not keep the search inputs and buttons the way they are now, but after searching, instead of just seeing the 'search term' followed by an X icon to cancel it, do a breadcrumb system similar to that of pages.
Show the first search term, with the same X button and functionality, and then continue to display the search input/button. When the user hovers over the submit button for the search box, there is a simple javascript dropdown. The first option remains 'Search', while underneath this the 'Filter' or 'Next' search button fades/slides in. By clicking 'filter' or 'next', it would then search the already retrieved results for the next search term. When the page loads it could show a breadcrumb of search terms, with the X icons next to each, allowing a user to remove one or navigate backwards/forwards in their search.
Arguments AgainstThis method should require little UI modification, however again, greater minds than my own at dA have probably considered this and gone against it. I would however like to know why, hence my post. Is it due to worry about users straining the db or backend? I'm fairly certain dA is a lamp environment, although I could be incorrect. If its not related to preserving search query time and limited database overhead, I'd still love to hear why this type of search wasn't chosen.
Conclusion
Finding things on dA can be difficult. Luckily, we have amazingly easy to use features available to use like Collections / Favorites, devWatch, and a host of others to ensure we will always have access to a special image, so long as we remember to mark it as special when we find it.
The two approaches I examined and discussed for providing a more user-centric and effective feature-rich search were as follows:
a.) an 'advanced search system' with color tagging, either by existing tags, an ajax color picker, or automated script', and
b.) the 'search filtering' system, using the same layout and ui but with search terms appearing as breadcrumbs, making filtering a breeze and easily navigated, all while offering extended functionality.I also discussed possible reasons against these implementations, further considerations, and made an appeal to Einstein's Razor because I think I'm smart.
If you've made it this far, thanks for your time and patience in reading this journal.
Feel free to add in below with your thoughts, comments, critique, judgement, and whatever else you can think of.
Part 2 Coming Soon.
Best,
Alex aka logiqdesign
`ChewedKandi has certainly gone out of her way to keep the vector community on the right path. Always making sure that her talents are infinitely scalable, Sharon has put her bezier curves to excellent use, and firmly anchored herself as an inspirational leader. We're absolutely delighted to bestow the Deviousness Award for June 2013 to `ChewedKandi. Congratulations, Sharon! Read More
Sta.sh
You can rotate the shapes!
You can not alter the shape.
each ) on this giveaway.
It's easy to participate 
This journal.
Yea... and if you don't read the rules, you won't get anything out of this.
What you have to do
Advertise; you don't have to but is looked up upon
3 x a 1 Month Premium-Membership
At dA we like to keep a balance between letting the community know what we are working on, but also having a little bit of showmanship and not revealing too much before we are ready to dazzle you. I especially don't want to comment too much about search, since it is not the project I am working on (and I want to let the people working on it have the pleasure of showing off).
We do realize that search is lacking right now, so we have a couple of our best engineers working on fixing it. It is actually even more important to us than you might realize, because uncountable parts of the site use search technology underneath the covers. Being able to immediately retrieve a stream of images that fit a certain set of characteristics is very useful to us, and it is very useful to our users (after all, browsing through artwork is one of the core purposes of deviantART).
That said, search is a very hard problem to deal with. There are countless "facets" that can be used to describe a piece of art, and most of them are hard for a computer to "see". You mentioned color, but there is also subject, mood, media, art movement or style, . . . the list could go on. Even within color, there are many ways to interpret the issue. In Van Gogh's famous painting of lilies there are a dozen purple flowers and one white one. To a computer the picture is predominantly purple and green, but human eyes are drawn to the one white flower so white is possibly the most important color in the painting.
You might have noticed that about a year ago the prints shop and groups browsing got a major overhaul. That was because $mccann and I overhauled the search engine that keeps an eye on the shop and groups. This year $mccann, $20after4, and $shkvarchuk have done similar things (but also much much more) to the main deviation search. We just started deploying some servers with the new search engine on it, so some of the changes that come with that should be filtering up into the user interface soon.
By the way, if you know of any engineers who specialize in search technology, we have a lot more ideas in the works, and we are looking for an engineer who can help us out.
I appreciate your time, and your response.
I am very anxious to see what the team comes up with as far as search technology goes, in the continued path to dA world domination. You brought up an interesting point I thought was worth mentioning that I left out of my initial analysis for brevity's sake.
"You mentioned color, but there is also subject, mood, media, art movement or style, . . . the list could go on."
This is so true. How we categorize art is a science in itself. How do you classify an artistic piece, do you do so by relating similar elements? How do you give a category to a piece that connects with the viewer, on a deep level, and give a name to such a thing? How is it possible to group together pieces that are each a world apart, yet, somehow, similar. The answer thus far has been through subject matter and/or medium, in addition to time period.
But this is indeed an incredibly difficult facet of designing a search tool for art, and is something I think would require a ton of thought and testing.
Also, on Van Gogh's painting of lillies, please forgive me as I know far too little about classical and important art, but I am unfamiliar with this. I thought perhaps you were referring to Monet at first, but am now after a quick search of my own I'm leaning towards perhaps Van Gogh's Irises work. Regardless, it got me thinking, and is a great example. Computers and technology have evolved so far, yet they are still restricted by our understanding and ability to communicate a set of rules and guidelines for them to follow. It would be so difficult for an automated color system to accurately depict importance of color, however, I do think it could provide a general estimation. The real question is whether a user would remember to search for the dominant color, or would they appeal to memory/emotion and search for the most meaningful color.
Lastly, "if you know of any engineers who specialize in search technology, we have a lot more ideas in the works, and we are looking for an engineer who can help us out."
I again am excited and cannot wait to see the 'more ideas' the team has brewing. I don't know a ton of engineers with the level of expertise in lamp environments most likely necessary, nor strictly search engine technology, however, should the time come when you guys need yet another creative thinker with a fresh perspective, in any situation, please let me know.
Thanks again for you time, response, and continued dedication to making this community so great.
Best,
Alex
If the user knew that they were looking for that particular work they might search for the secondary color ("van gogh purple flower" might be their search term), but most people who search for art on deviantART do not know what they are looking for. They are saying, "I want a print to go above my couch, and my couch is a slightly green off-white, and I like turtles". So, we need to be able to find a picture of a turtle where the emotional part of the image matches their particular color that is sold as a print of a certain size. If we deliver an image that is a bluish off-white, then it will clash with their couch and we haven't helped them at all.
"This is indeed an incredibly difficult facet of designing a search tool for art, and is something I think would require a ton of thought and testing". Yes, it is incredibly difficult, but you would be amazed at some of the prototype we have built. I really can't wait until all the pieces come together and we can launch it.
- Mike
"but most people who search for art on deviantART do not know what they are looking for."
This is perhaps the most important and simple realization that I never had. A great observation, which I somehow overlooked, that makes a huge impact on the issue at hand.
I am so excited to see whats coming. I wait the launch with anticipation, and know it will continue to blow away my expectations and surprise me with the level of innovation and detail.
I have some projects of my own, related to dA in a sense, that will be seeing the light of day in the near future. I have very big plans, and believe one of them you may find particular interesting. I'll let you know as the time draws near. It is a new technology in the form of code, and should have a measurable impact on its target audience, especially concerning the way we view art on certain websites.
Thanks again for your time. I really can't say how much it means to know you guys still read user comments and weigh in with opinions. Perhaps the personal touches like this are the reason dA still feels like home for me after 5 years on the site.
Best,
Alex