 |
 |
On Feb 15, 2013, at 12:27 PM, Kyle Banerjee wrote: > On Fri, Feb 15, 2013 at 6:45 AM, Diane Hillmann <[log in to unmask]>wrote: > >> I'm all for people learning to code if they want to and think it will help >> them. But it isn't >> the only thing library people need to know, and in fact, the other >> key skill needed is far rarer: knowledge of library data... >> >> ...More useful, I think, is for each side of that skills divide to value >> the skills >> of other other, and learn to work together.... > > > Well put. No amount of technical skill substitutes for understanding what > people are actually doing -- it's very easy to write apps that nail any set > of specifications and then some but are still totally useless. > > Even if you never intend to do any programming, it's still useful to know > how to code because it will help you know what is feasible, what questions > to ask, what is feasible, and how to interpret responses. > > That doesn't mean you need to know any particular language. It does mean > you need to grok the fundamental methodologies and constraints. And the vocabulary (which Alison also mentioned, but for those who read Stranger in a Strange Land know that 'grok' was also associated with understanding the language to be able to explain what something was.) I've had *way* too many incidents where the problem was simply mis-communication because one group was using a term that had a specific meaning to the other group with some other intended meaning. I even gave a talk last year on the problem: http://www.igniteshow.com/videos/polysemous-terms-did-everyone-understand-your-message And one of the presenters earlier that night touched on the issue, for scientists talking to politicians and the public: http://www.igniteshow.com/videos/return-jedis-so-what-making-your-science-matter It takes more than just people skills to coordinate between the customers & the software people.* Being able to translate between the problem domain's jargon and the programmers (possibly via some requirements language, like UML), or even just normalizing metadata between the sub-communities is probably 25-50% of my work. As a quick example, there's 'data' ... it means something completely different if you're dealing with scientists, programmers, or information scientists. For the scientists, metadata vs. data is a legitimate distinction as not all of what programmers would consider 'data' is considered to be 'scientific data'. -Joe * http://www.youtube.com/watch?v=mGS2tKQhdhY