How to validate (and group) User Listings in Marketplace

Question

I am building a hobby marketplace (comic books). Users are able to list comic books they own for trading with other members. The problem I'm running into is that I would like to standardize the listings so I can display multiple options per title. Like this:

User 1 lists: "Amazing Spider-man #4" on site.com/amazing-spider-man-4

User 2 goes to list "amazing spiderman 4" -> direct him to place the listing on site.com/amazing-spider-man-4 instead.

Then this starts to get a bit complex... so in a theoretical example, there could be numerous volumes of each series. i.e. Amazing spider-man #4 (released in 1965) and then again, due to a reboot another Amazing spider-man #4 (released in 1981) or whatever.

Is the answer to this to create a database and populate it with all series titles, issue counts and volumes, and then require the user to select one of those before making a submission? Such a data-source does not exist with reliable "volume" information, so I'm not sure how I'd do it.

Answer 1

The problem is that you're allowing free text input and the user is allowed to inut whatever he/she wants to. The examples list the title ("Amazing Spiderman"), the issue ("#4" and "4"), and even the year published ("released in 1965").

The first question you should ask is: what information am I really asking the user for? Do I really want title, issue and year? Should I ask for these things separately instead of in a single field?

If you take the three values you mention ("Amazing Spider-man #4", "amazing spiderman 4" and "Amazing spider-man #4 (released in 1965)" and strip the special characters, formatting, etc you get:

amazingspiderman4
amazingspiderman4
amazingspiderman4releasedin1965

As you see the first 17 characters are the same. That means that if I start typing you can automatically match my input to what's already know and provide suggestions to me.

So I suggest two solutions:

a) break out and classify your data set better; group it into title, issue and year, and make unusual fields optional.

b) implement fuzzy string matching and present the user with the matches so he/she can pick whatever is right or type his/her own.

Option a) is probably the least effort as you're formalizing the input and not allowing as much room for error.