Flutterby™! : Eric on his dream job 2011-05-18 17:57:50.247982+00

Eric on his dream job

2011-05-18 17:57:50.247982+00 by Dan Lyke 4 comments

As If Anyone Cares: My Dream Job is worth a read, especially in the context of Eric's comment on "So where are the GI Scientists".

I think all of these things are part of a general fragmenting of knowledge. The traditional publication mechanisms are now largely irrelevant to anyone working in the field. I've been amazed in the stuff I'm looking at that technologies that I thought everyone knew and were well established, like agent based flocking algorithms and image processing to read signs, are imagined as deep future by people who are completely up to date on the general engineering in their fields.

In that So Where are the GI Scientists thread, Ed finishes up with:

However I think there is an underlying issue as I noted, in that GIS as an academic subject is in danger of becoming a purely vocational discipline, solely concerned with the application of existing tools ..

Who is going to develop the new software tools and analytical methods of the future ?

Here's another random thought: I'm currently playing with JavaScript. JavaScript documentation falls into three camps:

You don't need anything else. (ie: Dori & Tom's book).
My über framework is the one true way, you should use my funky notation to find your objects rather than document.getElementById
"I included this plugin and it worked".

I'd really like something in the middle, something that lets me build on tools that understand when the page has actually fully loaded, and give me some <canvas>-able widgets, but don't try to hide the basics under yet another level of silly indirection.

It seems to me that this problem of finding my level of abstraction may be the same thing that GIS is going through, where the social structures around the technologies are still too complex to allow for a path through them, and if you want to contribute to the field, rather than use the tools, it's easier to just reinvent the wheel.

(Totally ADD digression: Managing this is part of what's made Perl great.)

#Comment Re: made: 2011-05-18 19:28:28.45853+00 by: ebradway [edit history]

I just posted a response on Ed's blog. Basically, there has been a dramatic increase in the demand for people who are able to apply GIS in their jobs. But this doesn't mean there aren't GIScientists who research how to advance the field. If anything, the increased vocational demand creates more day-jobs for GIScience researchers.

As for the middle-path problem... It's easy to define the extremes - bareback (or bare book?) at one end and uber framework at the other. But the error is assuming the space between the two extremes is one-dimensional - that there is some middle point that balances the two. I'd argue the space between has infinite dimensions. There are middle points appropriate for every application. If you are lucky, your middle point will be close enough to someone else's middle point and you can reuse their code. Or, as is typical for freelance developers, you only take jobs that cluster around your middle point allowing you to leverage your existing code.

FYI: That's basically the application of Tobler's First Law of Geography to a software development problem...

I'm installing a Juice system in the Honduras. This morning our customer and Nancy and I knocked off a few hours to go snorkeling from a small purple boat right after the "utility cook" fixed us breakfast.

One of the reasons you are often "..wedged between the hardcore geeks and the pointy-haired managers" is that you bring a set of skills that includes the ability to articulate complex concepts to less technical people. Having an open mind and pretty easy going disposition helps. You also touched on a key issue in technology: You are a self-learner and can synthesize new knowledge from previous knowledge and experience. As I travel and meet more and more "educated" people, that ability is proving to me to be increasingly rare.

I see a lot of the practical, day-to-day side of implementing and using GIS tools where I work - in fact the matter is considered complex and important enough that we have an in-house GIS specialist.

I think GIS has an extra-bad case of the problem where the people who are building the tools are not ultimately going to be the big users of those tools, and the people who use the tools know what they want but are not technically capable of building them. Since my co-worker the specialist is the rare case sitting in the intersection of those sets, I hear about this a lot.

There's also not a lot of standardization, I gather. I myself, in the limited amount I have played around with GIS, was surprised to see how little agreement there was on rookie things like data-file formats. That was a couple of years back, so things may have improved since then.

Everybody wants to develop their own middleware, because they think Their Way Is Best (as you say, this is a problem with JavaScript as well). I hate to wish for a dominant app to force everyone else to do it that way, because I distrust monocultures, but in some way it would make life simpler. (Google Earth is the closest thing to being a consumer-level nine-hundred-pound-gorilla in this case; fortunately for all of us, the tech they bought from Keyhole and have barely touched since then is pretty lucid.)

By the by, I think there's kind of a contradiction inherent in saying both "I'd like something that lets me build on tools and provides useful widgets" and saying "don't want another level of silly indirection." How do you get useful widgets/chunks that work on top of [raw language] WITHOUT some degree of indirection? Or am I misunderstanding you?

[I use the Yahoo! UI Toolkit, which may be the only thoroughly sound thing to ever come from Yahoo since they began their long slow decline. It has a sane DOM and its tools, while not the easiest to comprehend in terms of sorting through all the boilerplate, are also thin enough that you can see exactly how they're manipulating the underlying page code. But I also don't write much on-page, AJAXy stuff these days; my job has shifted almost entirely to writing invisible link-middleware to shuffle data between various systems/databases.]