I just got back from Signal 2017, the Twilio conference. First off, you will probably ask “What’s Twilio?”; I get that a lot.
Twilio provides an ever expanding range of telecomm services – phone numbers, SIP phone connections, SMS, video, notifications, chat and so on. They have a developer focus, so if you just want a landline for your house, you don’t want to come here unless you are a total geek*.
I became interested in Twilio when I was designing the “Vastra” phone system, I started using their Elastic SIP trunking service.
Then I found out about their non-profit grants and got one for Geo-CEG. Now Geo-CEG has a public number and a simple IVR (“Press 1 for Brian, ” etc) at no cost to us. I am currently cooking up a project with the Laguna de Santa Rosa Foundation to help them with outreach. No idea what it will be yet. Maybe something cool, maybe nothing at all!
Anyway, back to Signal; two talks stood out for me personally. They were Lucky by Saron Yitbarek and At-Home Batphone by Lauren Leto. (Follow the links to watch them.)
A lot of what Saron said in “Lucky” resonated with me, sometimes I feel sorry for myself because of some insignificant set back and have to push back and say “hey, your whole life has been one long pleasure cruise!” I do lots of volunteer work to give back, in my own way, and to support things I believe in.
Regarding Lauren’s “At-Home…”, tech is fun, but it’s causing people to draw themselves into what I see more and more as a destructive mindspace. She seems to be realizing this, and has a goal of helping people manage the information overload. I need to find out more about Listen and specifics about what she is trying to do.
In the meantime I am doing what I always do. Trying to find a path to enlightenment and trying to solve problems with tech even though I know it is not a long term fix for anything. We need a shift in consciousness.
* Did I mention my landline for my house is a Twilio SIP trunk running into one of the 4 existing Vastra phone systems?
This image shows building and parcel layers for Solano county on top of their aerial photo. Attributes for the submerged parcels show “government owned” and “below minimum tax” but no indication of “below sea level”. There are apparently no buildings there.
I am hoping this was planned bay fill that went by the wayside when people came to their senses.
Synopsis: If you do a lot of editing in ArcMap, you could probably benefit from using a Contour ShuttleXpress (about $60) . You will have to configure it because there are no ArcMap presets, but that’s very easy to do.
I searched the Web for tips on how to set my Shuttle up for ArcMap, and found NOTHING. It is a great tool so I am writing it up here.
The Contour ShuttleXpress is a USB device that has 5 buttons, a jog control and a scroll wheel. Contour provides software for Mac and Windows. I use it with a Mac. (From now on I will just call it the “Shuttle”)
ArcMap is a popular GIS program from ESRI. If you are reading this I assume you know that already.
Contour also makes a bigger device with way too many buttons on it to be useful, the ShuttleXpress Pro.
I bought the Shuttle years ago thinking it might be useful with ArcMap. At that time the only heavy edit sessions I was doing were for City of Corvallis, where I was not allowed to installed the required software. I ended up throwing the Shuttle into a drawer until this week when I dove into a big project.
The current project required doing quality control on point files, each containing 15k-100k points. The task was to check for and correct missing, misplaced, or duplicated points against aerial photos.
Using the Shuttle helped me streamline the edit workflow. I programmed the controls to work as follows. Each control on the Shuttle can be set up to generate keystrokes and ArcMap has a wide range of keystroke shortcuts, so you can program it to do many many things in ArcMap. These were my choices.
Button 1 : Copy (Ctl-C)
Button 2: Edit (letter E)
Button 3: Toggle layer (spacebar)
Button 4: Delete (Forward delete key)
Button 5: Paste (Ctl-V)
Jog left : Previous Extent (< key)
Jog right: Next Extent (> key)
Scroll wheel: not set yet, currently same as mouse scroll wheel
For most edit operations, I keep three fingers resting on buttons 2-3-4. When replicating big blocks of points, I tap the Copy and Paste keys with my pinky and thumb. (It’s all much easier than it sounds.)
Configuration via the Contour Shuttle app was easy, I won’t go into it here except to show you a screenshot here.
After programming my Shuttle I put labels on it, because in a month or two I will have forgotten how I set it up.
What the controls do:
Cut and Paste: I think you already know what Cut and Paste do.
Edit mode (E keystroke) When in an edit session, the E key will flip between two edit modes. Normally it will flip between Create Feature and Select Feature. This makes it easy for me to switch between adding new points and deleting or moving existing points without moving my hands around. There are some wonky things with this control in ArcMap (see next section.)
Toggle layer (spacebar) Typically I set things up so that I can toggle the aerial on and off. This command is wonky too.
Del: I put a Delete key onto the Shuttle because my computer is a Mac Book Pro and it has NO delete key. Apple mislabels the backspace key as “delete”. To generate a delete (aka “forward delete”) requires holding down the “fn” key, so I have to use two hands. Getting a working Delete key was the main motivation to try the Shuttle. (I also use an external Mac keyboard sometimes, but that’s clunky because it takes lots of desk space.)
Jog wheel: I developed a work flow where I pan across the image when zoomed to 1:1250 and then zoom into an area of interest (1:500) to do some edits, and then zoom back out and continue panning. I found the mouse scroll wheel works well for zooming in and out, but that being able to go back to the same extent at 1:1250 to resume panning over the workspace was important. I needed to be able to get back to the same spot. That’s where the jog wheel comes in.
I keep my fingers on the buttons as mentioned above, but my thumb rests on the edge of the jog wheel. After editing points at 1:500 and possibly panning around at that level, I bump the jog wheel to the left with my thumb until I see the 1:1250 extent again. Then I resume panning. If I spot something as I am zooming out, I bump the jog wheel right to go forward in the extent stack so it’s handy to be able to jump either to previous or next extent with the jog wheel.
Scroll wheel: This control is set at its default, so it works like the mouse scroll wheel. You can put a finger in its detent and make continuous circles either direction. I don’t use it yet. Soon I suppose I will think up something to do with it. Suggestions? Tell me.
As mentioned above there are a couple oddities you have to adjust to when using these keystroke shortcuts. The problems exist with the Shuttle or the keyboard; they are ArcMap bugs having nothing to do with the Shuttle.
In ArcMap there are problems with the “E” and “spacebar” shortcuts that you have to live with.
E toggles, but you have to “prime” it with what it toggles, by using the mouse to flip between a create template and the Edit select arrow. Until you have flipped between two commands it does not know what two things to toggle between. Once you have primed it then tapping E will flip between Create and Select. If you pop out of ArcMap for example to read help docs in a Web browser, you will probably have to re-select the tools with the mouse to get back into the right state in ArcMap.
You will be able to tell when it is primed correctly, because the cursor will change on each E keypress.
Spacebar only works in the Table of Contents. This is probably the most annoying thing here. The workflow would be pan with mouse – spacebar toggle – pan – toggle… but every time you click in the map window to pan, the map becomes the active control and it ignores the spacebar. So each time you want the spacebar toggle to work, you have to click somewhere in the Table of Contents control to reactivate it. It is still much easier than aiming the mouse cursor at the little checkbox on the layer, but I wish spacebar worked in the map. It would be easier.
The best thing about spacebar toggle is that you can be flipping the layer on and off with your left hand while your right hand moves the mouse cursor to a spot you need to edit.
One last glitch. Once you have activated the Table of Contents by clicking in it, the E shortcut won’t work for you. If you have a layer that starts with letter E and you hit the E key (on Shuttle or keyboard) instead of doing the edit toggle, it selects the layer. You have to live with this. You have to click in the map or on a map control before it will become the active control again. I usually click on the Edit toolbar select arrow.
MacBookPro + ESRI ArcGIS ?
I find the combination works quite well. ArcMap runs only in Windows, but I use Parallels to run Windows 7. It works great; the Shuttle seamlessly passes keystrokes from the Mac to Windows. I use Parallels because it has a much better video driver than either VMWare Fusion or VirtualBox.
The only problem I encounter with using the MacBook with ArcGIS (and many Windows apps) is that it really requires using the right mouse button. On the Mac side I use “control + mouse click”. On the Windows side, this does not work, and double tapping the tablet control is too clumsy. I deal with this by keeping a small travel mouse in my laptop bag and pull it out whenever I use Windows apps. I find the mouse is a better tool for GIS than the built in tablet anyway, so it’s a small thing.
If you don’t do a lot of edits in ArcMap using the mouse and keyboard shortcuts will be fine for you. But when you are pushing hard on a big project with a deadline looming, having this gadget might make things go a little bit faster for you. So it’s well worth the $60.
I have no connection with the manufacturers of anything mentioned here. I just like their products and use them.
I am testing various web-based mapping services now. Today I am testing CartoDB.
Using CartoDB at the free level I uploaded my datasets and created a map. I published it. I found a few bugs along the way.
The biggest problem I have seen so far in all the tests is that none of the services let me zoom in to a parcel level scale. Look at my test maps and you will see what I mean. For this map to be useful, I need to be able to see the irrigation valves and the zones.
One of my polygon layers did not display, saying it has a SQL error. I had changed the name in “metadata” and apparently that change not just metadata but also the name of the table.
2015-August-03– I wrote up this Cricket article for a Maker Camp project I created for the Novato library, July 2015.
What we’re building
A “cricket”! It’s all digital, using an 8-pin microcontroller and just a few other components. A “microcontroller” is like a tiny computer, complete with processor controleld by a program, some memory, and peripherals like timers and inputs and outputs.
I wrote a program for the microcontroller to make it act like a cricket. That means it will chirp when the lights go down. When it gets really dark it chirps louder and faster. If the lights come on, it gets scared so it stops. If it’s been chirping for a few minutes and the lighting has not changed at all it will fall asleep.
Before we start: How solder joints should look.
These were done by someone (Julie, my wife) who practiced soldering for about 20 minutes. They are not perfect but they work. Your solder joints need to have a smooth shiny appearance. If they are dull grey or rough, they might not work. You can reflow the joint by heating it again and applying a little more solder.
You need ENOUGH solder but not big massive blobs.
We’re using “dead bug” style construction. We turn the PIC upside down so its legs stick up in the air like a dead bug.
Bill of materials
Microchip 10LF322 microcontroller in the 8-pin DIP package
1/4 watt carbon film resistors: 1K and 10K
CR2032 button battery and holder
Four pieces of hookup wire, about an inch long each. I used 24 gauge solid conductor.
Something to hold all the parts; I used an Altoids tin, it’s sort of traditional
PIC 10LF322 is an 8-bit microcontroller made by Microchip.
The piezo speaker bends when voltage is applied to it, bend it fast enough and it makes a sound.
Resistors are devices that limit current.
A photoresistor is a special resistor that can detect light and varies its resistance.
Each resistor has a color code to tell you what its value is in ohms. For the cricket we need two, 1K = 1000 ohms and 22K = 22,000 ohms.
For Maker Camp, I preloaded a program into the PIC microcontroller that makes everything work. If you are reading this on the web site, I would be happy to provide more information on the program and how it gets loaded, or to send you a pre-programmed PIC. The program is written entirely in the C language and it’s maybe a hundred lines of code.
We’re using “dead bug” construction. It’s called that because when you flip the PIC upside down with its legs sticking up in the air, it looks like a dead bug.
Using dead bug construction will make this project cheap and quick to build. The downside is, it’s also easier to break pins off the PIC and harder to reuse parts.
To reduce the chances of messing up, we will build parts first into 2 subassemblies, then attach those subassemblies to the PIC.
The first subassembly is the speaker.
We have to attach a wire to one lead of the speaker and the 1K (RED-BLACK-BROWN-GOLD) resistor to the other lead.
It’s easiest to solder the resistor to the speaker and then the wire to the speaker as two steps.
Step 1: attach resistor to speaker; use either pin on the speaker.
Step 2. Attach a wire to the other speaker lead.
The second subassembly is the light sensor.
Together, the photoresistor and the 22K resistor form a circuit known as a “voltage divider”.
When we’re done, the photoresistor goes to the + side of the battery and the resistor to the – side.
The wire lead goes to the input on the PIC controller. Put the three parts together like this then solder. That is, attach a wire and the 22K resistor to one of the leads on the photoresistor by crimping them with needle nose pliers, then solder the joint in one step, this is better than trying to solder one part, adding the second and soldering it, because the first joint will fall apart when you heat it the second time.
Next The sensor gets attached to the battery. DON’T SOLDER YET. There are more wires to add.
Add a wire to each terminal on the battery holder. Use a light color (I used yellow) for “+” and a dark (purple) for “-“. The standard is to use red for “+” and black for “-” but I did not have red and black.
The battery holder has a little “-” in a circle near the “-” terminal. Of course, if you get the wires backwards, nothing will work.
After connecting the resistors and wires your battery should look something like this.
If the wires don’t fit through the battery tab holes then just attach them onto the resistor leads, near the battery holder tabs.
Now solder everything. Solder should flow over the whole joint, you need more here than anywhere else in the project.
Here is the back of the battery holder. Solder flowed down over the whole tab.
Now onto the heart of the project. The controoler. The PIC controller has eight pins. Pin 1 is marked by a little dimple.
From the top… (I need to add a diagram here.)
As a dead bug… upside down
You HAVE to keep track of pin 1 once you turn things upside down else NOTHING WILL WORK!!!
The way I did this was by putting a DOT on the can with a Sharpie marker, then gluing the PIC down so that PIN 1 is NEAR the DOT. Another way is to actually color pin one with the Sharpie– we won’t be soldering anything to it anyway. Here’s how I marked my tin can.
Next you need to glue the PIC into the can. Use hot melt glue, and do this QUICKLY — the cold metal can will cause the glue to glob up and harden fast. If that happens just peel off the glue and try again. You DON’T want the PIC to fall off in the next step, so use a good size blob.
Now attach the leads for the speaker assembly to the PIC.
6 3 ———– speaker
5 4 ———– 1k resistor
After soldering, carefully bend leads to position the speaker inside the lid. It has to be far enough inside so the lid can still close.
Then glue the speaker down. This will just help keep everything together so it won’t get broken as easily.
Don’t short out the metal connections to the can.
Mount the battery in the can.
Some of these cans have paint on them but to be on the safe side, I put a piece of tape down to insulate the battery terminals from the metal. If you short the battery terminals together, the battery will go dead in about a second.
So first put down a piece of tape and then glue the battery to the tape. Pin 2 on the PIC gets the “-” wire and pin 7 gets “+” so position the battery / sensor assembly to make it
convenient to hook the wires up.
After the glue has set, carefully bend each wire around and attach them to the PIC. Solder.
Again that’s PIN 7 = YELLOW (“+”)
and PIN 2 = PURPLE (“-“)
As I said earlier, normally RED = “+” and BLACK = “-” but I did not have red and black. The electrons only care if the wires are hooked up correctly.
All the wiring is now done. The battery looks like this.
Pop in the battery with the “+” side UP (visible when in the holder).
The cricket will make a beep as soon as the battery is plugged in. Then it will start chirping. The cricket wants to be in dark places, so if it does not start chirping try closing the lid on the tin or turn out the lights in the room.
To make the battery last longer (and to be less annoying!) if the
lighting is not changing then the cricket will stop in a few
minutes. Waving a flashlight at it or shading the sensor (any significant change in lighting) will wake it up and it will start chirping again.
That’s it, we’re done!
This device is a “cricket” only because of the program loaded into it. With this hardware, it has the ability to sense light and temperature and to make simple sounds. (I did not use the temperature sensor in the final version of the cricket program.)
What else could this hardware do?
Example: it could be an alarm for an open door on a refrigerator. It would sense light when the door has been open for say 5 minutes and then start beeping. Then when you close the door and it’s dark, it would stop.
All you’d need to do to make the cricket into a refrigerator alarm would be to write and load a new program into it.
On today’s geocaching hike I carried my NanoBeam M5/16 up into the hills.
Now I am sitting on a hillside, the temp is about 74, a gentle breeze is blowing, and I am connected to the College of Marin network at 40 Mbps. Whoo whoo! A turkey vulture is circling in front of me about 40 feet away! Wow!
The NanoBeam works well for this application because the whole radio is housed in the small dome antenna you see in the picture. I used it with a power injector wired to a small 12 volt SLA battery. The mast is a piece of PVC tubing.
I bought the NanoBeam back in January to do testing when I was working with a certain local WISP, before learning 1/2 their network operates on nonstandard frequencies.
Time to break it all down though and go have lunch with Julie. Well – it would be better if Julie were HERE instead of work. <3
Range according to my QGIS map is just over 1/2 mile. The NanoBeam is connecting with very good signal strength to a Ruckus Wireless access point located inside a campus building.
I am slowly moving away from public / free services and back onto my own servers. About a week ago I installed ownCloud at home, and now instead of 5 GB of Dropbox space I have 240 GB of ownCloud space, and I can easily increase that anytime I want.
If you don’t happen to run your own servers this might not make any sense, but I do, and so far it does make sense to me.
I get space limited only by my hard drive capacity, and I feel I have more privacy.
The more I use it, the more I like it. Since I move around a lot, my Mac always has copies of the projects I am working on. My desktop at home now keeps a local copy of the same projects. I can create a folder for GIS data and sync it up between the home server, desktop, and laptop, so I no longer actually need a high speed server to deliver data across a 1G network at home. Now I can go back to running a smaller home server that uses less energy; it can spin down its hard drives nearly all the time, and I will never see that lag when they spin up to accept ownCloud updates.
The only downside so far is that my O’Reilly account is linked to Dropbox. I am thinking that I will set up an ownCloud server to be a Dropbox client, and it will serve that directory space to my ownCloud clients. This should make it transparent to my clients without duplicating any files.
I just installed CyanogenMod CM12 on my Nexus 7 GSM (aka “Tilapia”), and it works now. Last time I tried (around December), the launcher would only run about 30 seconds. Crash restart crash restart. This made the device very hard to use. 🙂
So a month or so ago Google released Lollypop for the Tilapia and I tested it. Two problems. (1) SLOW SLOW SLOW (2) Battery life is GONE.
To be fair, battery life might have to do with running a SIP phone program all the time. But it’s still too slow.
So last night I installed CM12 again, it works, and it’s snappy!
I am now in the process of installing apps WITHOUT Google Apps. This means no Play Store. I am 58 so I don’t want to go to the “Play Store” anyway. I go to F-Droid now instead.
The App of the Day is “FareBot”. It lets me read the data on my Clipper card (used for public transit in SF Bay Area, including buses and BART), and the really cool and creepy part – the card has a log of all the trips I have taken. Date, time, and fare paid.
I keep more detailed notes on my Nexus 7 here: Nexus 7
Shortly after returning to Sonoma county, I got in touch with a friend
from my IHPVA/WHPVA, Steve Delaire of Rotator Recumbents. Though no longer in the recumbent business, Steve is still active with IHPVA.
From what I can tell IHPVA now pretty much just does one thing– they run a one-week speed event outside Battle Mountain, Nevada.
Speed events are fun but I have always been more interested in practical applications of human power, like commuting by bike.
Still and all, Steve told me he had volunteered to work on communications for the event. Providing communication support is
interesting to me. So I signed on to help.
There is no cellular network out there in the middle of Nevada, so mobile phones are useless. They’ve already tried handy talkies, which do not have enough range.
There is a nearby 2m repeater, so my first idea was to enlist local
ham radio people. Turns out that approach had already been tried and they are not interested in helping.
Steve’s idea was to use WiFi and smart phones. I am not convinced this is the best approach but it is certainly interesting to me, so it’s
what we’re trying. Steve still has more optimism than I about the
limitations of WiFI in smart phones. 🙂 My experience indicates that
you need a good antenna to make wifi work over more than about a hundred feet and I’ve never seen a smart phone with a good WiFi antenna.
But I love working with all these goodies, so I dove in.
I should have written many posts by now, so much stuff is going on.
A hardware failure in June inclined me to retire my beloved 2009 13″ Mac Book Pro and to replace it with a brand-new 13″. It is better in just about every way. My beloved told me to get the top end model, so it has the fastest i7 processor, 16GB RAM, the 1GB SSD. I thought the Retina display was just hype but especially for my old eyes, but now I find every other screen looks blurry and faded.
It is smaller and lighter, in part because there is no built-in DVD drive. I got an external USB drive for $30 from Newegg.
Committed as I am to doing GIS on a Mac while living nomadically, I went through a week in July of testing VirtualBox, VMWare Fusion, and Parallels 9. Parallels 9 won the contest. I was able to directly import my existing VirtualBox VMs into Parallels and I have not looked back since. For me, it’s the video driver. Parallels has the best out of the three. Very important for graphics intensive programs like ArcGIS and Topo.
ArcGIS 10.2 and Delorme Topo 9 work great. I was disappointed by VMWare Fusion, especially since ESRI specifically mentioned it in ArcUser magazine. Don’t bother. Use Parallels.
It was not clear to me at the Parallels site, but you can install it and select “trial” mode, just download and run the installer without paying for anything. The same is true for VMWare Fusion, and you can install them both at the same time, you just can’t run them at the same time.
Parallels 10 came out in September, I upgraded, and yes, it really is better and faster. Upgrade if you have version 8 or 9 — it’s only $50. It’s worth it.