Guidelines For Making A Cool Web Service

Over and over I hear suggestions from people regarding some idea for a website that will do x thing with y other site that will just be the neatest damn thing ever. These are usually suggestions by decent coders who have the itch to write something cool and have a million ideas floating in their heads about how to do it; they are usually people like me. Unfortunately, many of these plans never come to fruition because they haven't followed a few guidelines.

You Aren't The First Is there already a service out there is that doing this? Nine times out of ten your cool idea is already half-done somewhere else and you would be creating way more work for yourself by stumbling through all the pitfalls again. See what they did, copy as many of their ideas as you can. If you are interested in making a photo sharing site, sorry pal, Flickr's got your number, your best bet is to put them on your team and use their interfaces. I don't know how many times I've watched somebody dive headlong into a project that somebody else already did without realizing there was already something else out there that they could use or copy.

Coding Takes Time I don't care how much experience you have building web apps, you aren't going to be able to just "set something up" from scratch that has an ounce of the usefulness building your site on top of Drupal, Ruby on Rails or Twisted (if you already have the know-how to use the thing, fuck does it have a high learning curve) will give you. Honestly, for 90% of sites, there is no reason that a sub-five-minute Drupal install won't accomplish almost all of the work involved in starting a website immediately, and it has the nicest (and one of the best documented) plugin systems I have ever worked with. You aren't going to come up with something better than these frameworks and still have time for what you were really trying to do, a cool web service.

Steal This Data So you have this awesome idea for a service that will rate apartments for rent based on their proximity to free wifi hotspots, but you only know the locations of the hotspots near your house, how will this be useful to people in Mizuruh? Well, it won't until you find some site that lists the hotspots in Mizuruh and steal all their data. Good sites provide open APIs to their data, bad ones make you scrape their html, but if the site has a lot of the data you need, find a way to get that data. Flickr, Amazon, eBay and Audioscrobbler are all sites that provide easy access to their data, and sites like Starbucks, Google Maps, and Orkut have already had their interfaces reverse engineered by other talented people for you. If it is popular data, chances have it somebody already has a way to get it.

Share and Share Alike Face it, almost all of these project fizzle out before they actually amount to anything so make sure you release what you are doing along the way. Even if the project tanks, maybe that Drupal module you wrote to display Ticketmaster concert listings or the spider to scrape the 2600 meetings list will come in handy to somebody else. Failed projects is probably where you got all the code you are using anyway, you should give back. Who knows, if you get attention for releasing your Gucci flash-disassembler code maybe other people will want to help you make your fashion site.