Yesterday morning I learned that GPSD has been awarded the first Good Code Grant by the Alliance for Code Excellence.
This is how I think about it now:
Amount of test hardware you can now afford: $500
Being able to tell people that your project is funded by the sale of indulgences…priceless!
How much is it going to take to offset the NetHack source code? :)
Congrats!
Al Gore’s carbon-offset company is funding your project? Cool.
Seriously, congratulations on the grant.
Amount of test hardware you can now afford: $500
You can barely afford an N900 with that……which I’m having a monkey’s ass of a time getting GPSD to run on, btw ;)
Congratulations to the whole project team!
Congratulations to the GPSD team; you guys have done a terrific job with what is obviously a very thorny problem.
Now for the question we all want the answer to is … what goodies are you going to buy with it? :)
>Now for the question we all want the answer to is … what goodies are you going to buy with it? :)
A couple of GPS mice, probably. They tend to run between about $80 and $200 each. We’ll look for odd types we don’t presently cover.
Does it come with a “get out of Purgatory free” card?
and how much of it is your code? I think you maintained it for a while but dunno whether you actually contributed any good chunk of code to it…
>how much of it is your code?
Probably around 80%. There’s almost none of the 1.x code by Remco Treffkorn and the original crew left; the two major rewrites I did (the first factoring it into a driver/framework architecture, the second replacing the old protocol with a JSON application) saw to that. The framework code, packet sniffing, autoconfiguration, hotplug support, AIS code, and test tools are pretty much all me.
The two most-used drivers, for SiRF and NMEA, are collaborations between Chris Kuethe and myself. Gary Miller owns the PPS support and the Garmin driver. Other peoples’ code tends to be concentrated in drivers for odd devices and the corresponding individual logic branches inside the packet-sniffer DFSA; making such contributions easy was part of the point of the framework code. On the client side, Jeff Francis wrote cgps and Amaury Jacquot wrote gpxlogger, but xgps is all mine (I completely rewrote it in Python+pygtk recently to get rid of a Motif dependency).
The history of gpsmon is representative of what tends to happen to other peoples’ GPSD code. It started out as a SiRF-specific utility called sirfmon, written by Rob Jensen. I rewrote it once in order to factor the SiRF lower level out so that it could either talk direct to a serial/USB port or a running daemon instance. When I changed its name to gpsmon, I rewrote it a second time to use a framework/driver architecture so it could direct-monitor other device types as well. There’s probably none of Rob Jensen’s original code left in it.
Ever heard of version control? You can look at the Subversion repository to see who committed what. There are also tools to analyze that information for you like Ohloh. As you can see from Ohloh’s contributor list for GPSD, the person with the most commits is, indeed, our very own esr.
Judging by the number of commits, one would have to conclude that most of the code is esr’s.
>Judging by the number of commits, one would have to conclude that most of the code is esr’s.
The commit-frequency graphs on Ohloh track my sense of who’s responsible for what pretty closely. Note the presence of my two lieutenants, Chris Kuethe and Gary Miller, in the #2 and #3 spots respectively.
No doubt. After seeing Ohloh about 3-4 years ago, I wondered why someone hadn’t done that before. It’s fairly trivial to walk a version control tree and count up commits, lines of code, frequency of commits, etc., to determine project activity, key contributors, etc. But then again, I think most hackers are as resistant to what suits call “metrics” as they are to ties. ;)
>But then again, I think most hackers are as resistant to what suits call “metrics†as they are to ties. ;)
Yes, we’re resistant to bullshit “metrics” that are used as a substitute for actual judgment or a crutch for clueless managers, but every hacker I’ve discussed it with thinks Ohloh is brilliant. It was clearly designed by open-source programmers to compute statistics that actual programmers actually want to see.
If you have never used ohcount, I highly recommend that you do so. Ohloh isn’t just about counting commits and ranking people. In fact, all of that is rather incidental to what it actually provides in the process of marketing itself, which it has done rather well.
I suspect 2/3 of all ohloh haters are just suffering from check envy.
Congradulations :)
A short time ago I was in a coffeeshop run by the same sort of people Eric thinks I am. Hint: It has an attached yoga gym and the fare is 100% vegan. If you order a cup of coffee you don’t get milk or cream: you have to purchase it at a nominal fee. The funds become a donation to some sort of organic farm charity.
That’s right. Offsets for using animal by-products. Welcome to the early 21st, folks.