Adam Christian

Every year I like to make myself a road map of how I will be spending my time during OSCON. As there are so many interesting possible talks, gatherings and social events it’s tough to get to all the things you care about.

At this point in my career my focus is on Web Development, Test Automation (specifically for the web & browsers), and social networking. Obviously on a moment by moment basis your interests are pulled in varying directions, but that sums up the bulk of my attention.

If you are interested in the full schedule grid, it can be seen here: Oscon 08 Schedule Grid.

Wednesday

• 8:45 AM: Welcome
• 9:30 AM: Keynote
• 10:45 AM: “An Introduction to Ruby Web Frameworks” (It’s going to be tough to convince me to move away from Django)“Changing Education… Open Content, Open Hardware, Open Curricula” looks more interesting today.
• 11:35 AM: This one is tough, either “Web Graphics and Animations without Flash”, “Beautiful Concurrency with Erlang”, or “Beyond REST? Building Data Services with XMPP PubSub”, “CouchDB from 10,000 ft” apparently thats the thing see, or “What Has Ruby Done for You Lately?
• 12:20 PM: Really important, LUNCH!
• 1:45 PM: Probably “Thunderbird 3″, maybe “The Open-Source Identity Revolution”
• 2:35 PM: “Caching and Performance Lessons from Facebook”, never know when this one might come in handy working for Slide inc.
• 4:30 PM: “Open Source Community Antipatterns”, I’m really looking forward to hearing Ted Leung explain how to NOT run an Open Source Project…
• 5:30 PM: Probably “Give your Site a Boost with memcached”, or “Shell Scripting Craftmanship”

This is my final business review from the high school era, however this one is especially important because it forced me to get my hands dirty with some serious database work and made me write more php boiler plate than I had ever dreamed up until this point. FYI, the person driving this business was a teacher at my high school (his last name goes in the graphic above).. that I never took a class from. He had spent a lot of time working in the appraisal part of the real estate market, and as with any repetitive process — people start to wonder how it could be automated and simplified.

Idea

The booming real estate markets of the late 90’s and early 00’s inspired many (especially those who had been involved in the industry) to start seeing dollar signs. As more people were buying and building homes, more appraisals and inspections were ordered. In case you haven’t been around anyone who does appraisal work, you should know that the research and comparison pieces of the report consume large chunks of time. 

There was a point in time where to get information about lots, land and peoples homes, you would have to physically go to the county assessors office and look through the stock piles of records, plat maps etc. to find you comparable properties in order to base your valuation. All of this information is publicly available and one just needs to go ask to see it. 

It didn’t take long for a few companies to spring up with the idea that they would aggregate all this data, and they did it well enough to make a pretty solid business out of it. However, the distribution of this data via companies like the MLS at that time were on CD’s which you received regularly and had to load onto your computer (I’m sure they still have this as an option) but Mr. Teacher had the idea that it would be much more convenient if people could just access all this data via the Internet. 

Stack

Incase you were wondering about the technology stack we were using to build this, it was as follows:
Apache Web Server, PHP3, MYSQL. Your standard LAMP stack, but before it was your “standard LAMP stack”.

Pitfalls

I must admit, that when I accepted this gig I really had no idea what I was getting into. I made promises that I wasn’t completely confident about, ultimately my lack of experience didn’t turn out to be the killer.

• For a site like this to succeed we would need many counties worth of data
• Data needs to be kept up to date (picking up CD’s all over the state every other day is unreasonable)
• Provided data was not in a reliable format
• CD’s full of 100 meg comma delimited files are difficult to work with
• Building a web based competitor to the MLS by yourself when you are 16 is rather daunting

To expand a bit on the above, even after I had a site designed, user logins working, profiles working, and the first round of data for each county searchable I still hadn’t even reached the bulk of the work. At this point my method was to create a PHP script for each file’s particular format and parse through it doing DB inserts. Since the format of each file (even new files for counties I already supported) had changing formats, I was continually updating the scripts trying to make the exploded entries in the arrays match up to the DB columns etc.

Killer

When you are looking to jump into any market, you first need to take a look at the competition. What is going to keep them from squashing you like a bug. Think about it, they have resources, money, people and hopefully some insight into the market. It is much easier for them to create and deploy than it is for you, and they will, and they did.

Not too long after our 4th or 5th iteration of data and some testing, MLS announced their web based service. Around that same time, many smaller (already existing) companies in the real estate market announced that they would be doing the exact same thing.

We could have forged ahead, we had a working rough beta and with some serious persistence we could have built up a small user base by offering lower pricing… but that wasn’t my top concern. I believe that after my involvement tapered down, Mr. Teacher continued forging forward. A moment ago I checked the domain where the beta was available, and it’s no longer even registered. 

Lessons

• Do your market research
• If time is an issue, hire a reasonable size team
• Always get signed contracts (I’m pretty sure he still owes me money)