Sat 4 Oct 2008
profiling and searching
Posted by hambrushman under Senior Project
NOTE: Throughout this post I keep using the terms “profilers” and “searchers” this is to define the roles of users when they create/edit their profiles and when they search others profiles. I realize that all users are both, but I feel they put on different hats in each section and have diffrent expectations.
I have been doing some research on advanced searching methods on popular sites like Amazon, CNet Reviews, Facebook, LinkedIn, Monster, CareerBuilder, UArtsCareers, and yes, Toggin. I have also been looking at the way they profile users. I found a lot of good inspirations mostly from Facebook, LinkedIn and UArtsCareers. I also have always enjoyed Amazon and CNet Reviews filtering methods, where you essentially just select a general department, and just filter and filter until you have a select group.
All of this research, however, got me thinking that maybe none of these paradigms are the best fit for this project. Knowing how people categorize themselves and the fields of study they are passionate about, is a complicated procedure. I figure I can:
- Do a LOT of research and force people to chose between lists of categories and options of my choosing. Although this would make searching easier, it would probably feel confining and inaccurate to the profilers.
- Leave open-ended text fields for the profilers something along the lines of: “List your skills seperated by commas”. This would be harder on searchers who don’t know what they are looking for and who may have to search for “Actionscript” as well as “Action Script” as well as “AS3″ as well as “AS 3.0″.
- Do something a little experimental…
My experiment I am considering is basically to allow users to create their own categories. The best way to explain it is to give you a case study:
- I am a multimedia student who is creating/editing my profile. I see that no one else in multimedia, or the administrator himself, has created a “programming languages” category. I create it and list the programming languages I know: “Actionscript 2.0, Actionscript 3.0, C++, PHP”. I also list some languages I know about but don’t know myself: “PERL, ASP, etc”. I also add a note describing “programming languages” for the sake of the searchers. I check off for MY profile only the languages that I know and log out.
- I am a different multimedia student who is creating/editing my profile. I know HTML, and I see that someone in multimedia has created a “programming languages” category. I’m not sure if I consider HTML a “programming language”, I consider it more of a “markup language” but I will add it to that list anyway for the sake of the community. It doesn’t benefit the searchers to have too many categories if they are not familiar with this field of study (at least that’s what the administrator’s note says in the “create a category” section of the page). I also see someone has created a “graphics software” category, and while they included Photoshop and Illustrator, the neglected Fireworks and MS Paint. I add both to the category, and for my own profile I include Photoshop, Illustrator, and MS Paint.
- I am an acting major who has been struggling with a personal website, and I want to pay someone to create a flash video player for me. I search under multimedia because the administrator’s notes says multimedia majors include graphic artists, web designers, etc. There is a list of categories to search/filter from including “graphics software” and “programming languages”. I can see from the description that I probably want to search “programming languages” so even though I can select multiple categories, I only select that one. Then I can see from the list of programming languages, I want to select a search/filter of “Actionscript 2.0″, “Actionscript 3.0″, and “HTML”. I also select the search/filter to include graphic design majors because they have a category for “web design” and a list item for “Flash”. My search results return both multimedia majors and graphic designers.
This obviously would require a lot of programing and moderation on the back end, but I believe, if it was used properly, it would be a great benefit to the end users. My major concerns are:
- is it too complex for profilers and searchers who are or are not web savvy
- would users abuse it (i.e. create a “porn skills” category, and list “yo momma”)
- would profilers not be willing to generalize their skills (i.e. putting HTML in a “markup languages” category), making it more confusing on the searchers
- would necessary moderation appear too authoritarian by users (i.e. editing their category descriptions, combining categories, etc)
- would the research and time involved in implementing this be worth it
I keep trying to think about what 37signals would say. On one hand it greatly benefits my epicenter, the profiles. On the other it is in some ways expanding the project rather than focusing it.
7 Responses to “ profiling and searching ”
Comments:
Leave a Reply
Trackbacks & Pingbacks:
-
Pingback from Project Management according to Douglas Adams » jonathanTheDavis
October 20th, 2008 at 6:59 pm[...] started a module called “tagharmony” to accomplish this. Contrary to my idea in the profiling and searching post, I am not segregating any categories or skills for diffrent majors. There is no reason why [...]
October 4th, 2008 at 8:28 pm
i, for one, think this is a swell idea. i completely understand your concerns, especially the abuse one and the one about how much time you’re putting in / what you’re getting out. since you’re basically working with the content procedures of a wiki, is there a framework you can use that’s available on the web somewhere?
that said, i think you can probably not be as concerned about the administrative aspects and the reactions to authority. i would give the users the ability to flag content as inappropriate and let it go for now. how you handle flagged content is one thing, but if users themselves are the ones doing the “administrating”, you’re off the hook.
as far as users who aren’t as web savvy, i think a lot of the solution there will lie in your composition and presentation. what you’re asking them to do is not technologically complex, but it is mental exercise. we’ve all had to develop résumés before, and this is much the same idea. focus on making it simple, logical and concise, and the worst that can happen is that a reluctant or confused user comes back and adds new skills after they’ve been “created” by others.
one more, then i’ll stop yapping
i feel like a lot of the issues present in the web-savvy question are present in the skill-generalization question, particularly “where do i list which skill?”. i think you’ll find that people will be pretty quick to catch on, especially if you start them off with the basic idea [e.g. if you feel like HTML is going to be often miscategorized, put it under "markup languages" yourself. people will figure it out, and they'll start adding more and managing their own content [maybe you could also allow flagging for "this would fit better in this category"]. making things easy for searchers often means making searching as universal as possible, since you don’t always know what you’re looking for. cnet allows you to drill down using filters until you reach practically one device, review or program, but you can also search their whole database from one text field.
i’ll make it a point to sit and talk with you some more on tuesday, and we can bounce some more ideas around if you want. keep up the good work!
October 5th, 2008 at 2:53 pm
I think this is a really good idea, but you’ve definitely identified a lot of difficulties you face with it. I think, if your main point is to create a database of human resources with minimal amounts of confusion, you should strive to make this as user-friendly and across-the-board standard as possible.
What I mean is that in creating a searchable database of people and their skills, you leave a lot of room open for the profilers. You touched on this. To avoid “AS3″ vs “AS 3.0″ -type scenarios, I would suggest a list of skills (general and specific) that people would choose from when describing themselves. Draw on department specialties for the skills at first, and perhaps allow for an “other” type-in. As your resource is used, patterns of what else should be on the list will show up, and that’ll make it easier to create a specific list. This will also simplify the search process, as all the skills are right there: you just have to pick.
It probably won’t work right for a bit, but this seems like the kind of project that’ll pull itself together soon enough to become a really great resource.
I hope at least a little bit of that is at all useful.
L.
October 5th, 2008 at 11:18 pm
It seems you’re circling around a purely folksonomic model (ie. user-generated tags as in delicious and flickr), so why not adopt it entirely? A collaborative tagging approach would, over time, eliminate redundancy and semantic discrepancies as more and more users gravitate towards standardized tags. Eg. I know Action Script 3.0 and I’m not sure how to tag that skill but I see that others have used the tag “actionscript3″ so the decision is made by a kind of consensus. The user who is looking for an action script programmer searches or browses a tag cloud and filters “actionscript” or “actionscript3″ or “programming” and they’re off. User-generated tags also provide an accurate reflection of the actual sets of skills represented on the site. Tagging is widely understood, I think, and should be quite easy to grasp and use for most.
If you must assert some degree of top-down, hierarchical categorization, then be minimal and broad. Constrain yourself to top-level categories: programming, web design, graphic design, illustration, photography, etc. That said, though, dictating such categories may just trip up the folksonomy, so maybe they’re actually not necessary.
October 6th, 2008 at 2:09 pm
I for one am proficient in yo momma.
Anyway, I agree with Jeremy. My only thought is given the scale of the project once it’s started, moderation might initially be necessary to condense and standardize tags, until it “gets on its feet”, as it were.
October 7th, 2008 at 12:47 am
Jonathan,
I agree with Jeremy with regards to embracing a folksonomy approach, with Jimmy on the likely necessity of initial moderation, and with Tom in that you should be sure to keep ‘cognitive load’ in this tagging process to a minimum, i.e. ‘Don’t make me think!’
Delicious has a nice functionality in that it suggests pre-existing tags. I will post some links that you know who suggested to me on folksonomy…
October 7th, 2008 at 2:00 pm
Ahoy, Jonathan,
I will have to agree with Jeremy on the whole folksonomy idea. However, I do think that what Tom brought up with flagging is a good idea as well.
Anyways, best of luck with your project and I will see you in class.