Recently I was looking into some ideas for a project based around the idea of collecting body data from customers, and using it to either generate visualizations of fashion, or to make recommendations based on body-type and fit.
After a bit of research (googling), it was pretty obvious that there is a lot of technology and software around the subject of biomechanical modeling for medical and academic purposes, that might be possible to re-purpose for some retail fashion ends.
But it got me thinking about how different problems are addressed in different software domains areas. I noticed today how that some problems are really well covered by some software, and in other are still circa 1994.
An example is parsing html or plain text containing date and time strings into some structured format.
I previously looked at some software called logstash which needs to know timezones and if possible timestamps to millisecond precision for the purpose of event correlation in logfiles. So it has date string parsing down to an art. (well, its pretty good)
There is another tool called zotero, which wrestles with similar semi-unstructured text parsing issues, but is basically seriously under-powered in this area. (for zotero, getting the day-of-month and the month-of-year in the right order is usually sufficient accuracy) It makes a lot of sense, but it makes me wonder if there are some little gems of problem solving hidden away in these fairly niche projects.