Jason Antman’s Blog

In my daily reading of Tom Limoncelli’s blog, I happened by a post with a link to an article, “Geek behaviors present during conversations” by Phillip Guo. The article is quite interesting, and describes quite a few phenomena that can be noticed when conversing with a geek. One of the parts that immediately stuck out in my mind, though, was:

Using technical terms without checking for understanding

When geeks try to teach non-geeks about technical matters, they often
make heavy use of technical terminology because it’s the most concise
and accurate way to convey their thoughts. Unfortunately, many
non-geeks cannot understand such jargon and thus grow confused or
frustrated.

Non-geeks might perceive geeks as being snooty and elitist by talking to
them using cryptic ‘geek-speak’, but in fact, I suspect that their
intentions are more innocent: Geeks often don’t remember what it was
like not to have an understanding of technical concepts, especially
related to their area of expertise such as computers or electronic
gadgets. Thus, when communicating with non-geeks, they often throw
around technical terms without first tactfully checking to see whether
listeners know what these terms mean. In the geek’s mind, technical
jargon (e.g., router, dongle) and especially acronyms (e.g., DVI,
TCP/IP, SQL) sound as natural and colloquial as regular English
words. “How could anyone NOT know what these words mean?”

While I’m definitely guilty of throwing jargon into explanations when I’m at least half-conscious that the other party doesn’t understand it, I’d say that it’s usually (for me and many who I know) for a different reason. Or, more accurately, a combination of reasons that I’ll call “the expectation of effort” and “the Google effect”. Firstly, among most geeks (and definitely those of us in the open source world) there’s an expectation that people are interested, are curious, and will research something that they don’t know about when they hear it. The second part is what I call “the Google effect”; most of us geeks do most of our communication (especially about technical things) at a computer, whether via email, IM, forums, or chat. Therefore, there’s usually an expectation that if someone doesn’t understand a term, the answer is as easy as Ctrl+T and typing the term into Google.

I also must take exception to one other item in the article:

Favoring complexity and detail over simplicity in descriptions

When a geek is describing his latest pet project, he often tries to make
his descriptions sound complicated and detailed, in order to convey
the sophistication of his technical prowess.

I’ve described plenty of my projects to plenty of people, and yes, my descriptions are generally very detailed. But it’s not out of any desire to convey sophistication or prowess – my description may be the same whether I’m telling it to someone who can’t operate a computer or someone of much greater skill than my own. It’s simply the burning need to be accurate and complete. Put as simply and generally as possible, if I’m describing a project, I don’t want whoever I’m talking to to ask questions when I’m done describing it. It’s like documentation – if I were writing a man page, whoever reads it shouldn’t be left with major questions to ask when they finish.

Miscellaneous Geek Stuff geek, interaction, jargon, social

So, here’s an idea that I’ve been toying around with for a while.

I’ve been dong some development on a pair of Soekris embedded Linix boards I bought to use as a firewall/router and WAP. So, naturally, my interest in embedded systems is piqued. I know that there are some smaller embedded Linux devices out there. So, the idea is as follows:

Get an embedded Linux box that is either PoE or uses a hardwired connection with a wall-wart, or better yet, uses both, so it can be designed to have PoE as the primary power, with a 9v battery backup (ok, it won’t run too long, but can handle minor power losses if the PoE switch isn’t on a UPS).

Attach to it a clock-style LED display, or a matrix LED display, and an additional really bright red LED, both controlled by software (serial or GPIO).

The system has one RJ-45 connector for Ethernet (10-BaseT) and a serial console. The console is a captive login to a setup program to define IP, etc. but by default uses DHCP.

Plug it in, and it queries your DHCP server for an NTP server (or in the absence of one, uses a default Stratum 2). We re-update the time every day or so. Now, you have an accurate-to-the-second clock.

We add a speaker to the box, and have a few default built-in MP3s or WAVs for alarm tones, and add a GPIO “snooze” and “off” button.

The serial console allows setting a daily alarm or two, as well as selecting one of the default tones.

The Geekier Part:
The alarm clock is controlled by a script called by a Cron job. Now, we add the miniserv web server and a PHP admin gateway. Login via a password (user-selectable) and we can upload new audio files for alarms, as well as adding alarms via the web as cron jobs – now I can program in my class schedule, and automatically have an alarm at the right time each day of the week!

The Geekiest Part:
Remember the matrix display and that additional really bright LED? We setup simple communication between machines via SNMP or a script on the clock (called with an Expect script on a remote machine) and we can activate the super-bright LED, a custom sound, and even a scrolling message on the LED display (alternates between message and time, of course) for critical events. Now, when the servers go down, I have another device (in addition to the SMS alerts on my cell) to tell me to wake up and deal with it. If we want to get even more fancy, we can enable a feature through the serial console or web interface that emits a loud tone for a few seconds and blinks the LED if PoE is lost, or we can’t ping a user-selectable address for more than, say, one minute.

I guess I have too much free time on my hands..

Ideas and Rants alarm clock, embedded, geek, PoE