Sneak Peek at Behind the Scenes

Posted: September 10th, 2009  |  Author: Sergey Mikhanov  |  Comment on that »

Joel Spolsky of Joel on Software has once again gave utterance to one widely spread idea in his recent colum in Inc. The idea is that for startup there’s a safe way of getting more sales: releasing a better product. It’s better than advertisement, it’s better than anything else a small company could focus its efforts on.

This idea is simple and obvious; few people will argue with that. Needless to say, here at Critical Point we strongly believe in this approach. Up to date we’ve released two versions of our Transit application and it definitely sucks less with every release.

For the transport application like Transit there are two vectors of potential improvement:

• deliver the better user experience and more features, and
• cover more cities.

While first is very well defined in terms of software engineering, the second turns into a huge task when non-streamlined. Our Ruby-fu master, Esad, was able to tackle this process very effeciently with a tool of his own called Matilda. Here is a sneak peek at behind the scenes at Critical Point.

First, the city get its transport map and we are filling our internal database with the data about the lines and stations.

“Active” flag here means that Matilda now records the bitmap coordinates of the stations. Crosshair cursor helps a lot.

When something is wrong with the station name (Wikipedia, our main source of data, messes with diacrytics sometimes), it could easily be fixed.

After the station data including names, lines relation, colors, bitmap and GPS coordinates is prepared, it’s easy to review it.

This U1 line in Vienna seem to be processed correctly.

During the build process the data is hooked from the internal database and injected into the application. All for the user to see this:

---