Google has stopped selling the Pixel C, which is not too surprising, the writing has been on the wall for some time.

I've personally used Android for tablets for pretty much for as long as there has been Android for tablets, and I've personally not considered it to be any worse than iOS on tablets (read iPad). 

All this being said, I currently own a Nexus 9, which is now long in the tooth. I was not persuaded to replace the Nexus 9 with the Pixel C, now demised, due to price, more than anything.

Two things, however, have moved my primary tablet usage from Android to iOS/iPad. First off, due to a need at work for a device running a current version of iOS, I bought a refurbed iPad Mini 4 and frankly find it to be the best piece of tablet hardware available. Actually, prior to iOS 11, I would say that Android is the best tablet OS while iPad is the best tablet hardware.

This year I bought the iPad Pro 10.5 and Apple Pencil, which for me is the ultimate work tool. The Apple Pencil is the second reason why I moved from Android tablets to the iPad.

One area where Google failed with tablets, in my opinion, is in the lack of stylus support. Despite what Steve Jobs, among others, may have said, it turns out a stylus/active pen is a very useful accessory for tablets. Please note there is an important difference between a stylus as an accessory, as is the case for the Pro models of the iPad, and requiring a stylus to use a device, as was the case for all Personal Digital Assistants and smartphones prior to the iPhone.

I find it interesting that the new Pixelbook has support for an active pen, however because the Chromebook Pixel will not be used primarily in "tablet" mode, I doubt many people will buy the Pixelbook Pen for the limited functionality it provides.

Despite the predictable hand wringing that will go on about Android and tablets, I personally think the lack of a tablet from Google speaks more about tablets in general and less about Android. 

Keep in mind, iPad sales were dropping until this year when Apple dropped the price significantly. I suspect the lower priced iPads sold very well this holiday season. At the previous high prices, iPads were bought either by people with plenty of cash or who saw a real use for them that justified the expense. At lower prices, the iPad moves more to "why not" purchases, and that alone might spure the iPad of fulfilling the dream of it replacing notebook computers.

With the flagship brand having difficulty selling its own tablet, is there really a market for tablets? Most likely not, particularly when you through large screen smartphones in to the mix. 

So, at this point, I don't expect Google to ever release another tablet. Given the work they have been doing in combining Chrome OS and Android, I think any device with a 7-inch screen or larger will run Chrome OS. The question is, will Google simple stand pat with a $650 phone and a $999 laptop, or do they need another in-between priced, larger screen device? 

If Google releases another Pixel brand product, I think it most likely to be be a 2-in-1 detachable akin to the Microsoft Surface Pro. Two reasons make this the likely approach, in my opinion.

First despite how nice it is, I doubt Google is going to sell a lot of Pixelbooks, for nothing more than Google cannot overcome the stigma of ChromeOS on a expensive notebook. People are programmed into thinking ChromeOS is not a full operating system, and seem to think that a $1000 notebook must have a full operating system akin to Windows 10.

Second, a detachable 2-in-1 enables Google to kill two birds with one stone. A detachable provides a real tablet (in my opinion fold behinds are not real tablets) when you don't use a keyboard, and such a product could be priced from $600 to $800 to provide a lower price alternative to the Pixelbook. 

The tell for all this will be Google I/O 2018. If Google does not announce something to replace the Pixel C then, I doubt they will ever release a Pixel slate/tablet, which means they cede that particular niche of hardware to other companies. 

I got a Raspberry Pi Zero W as a Christmas gift and I have spent yesterday and today getting it set up, which for the most part means installing the operating system.

First off, now that I have one in hand, I am really surprised by the size of the Pi Zero, it is 2.5 inches long by 1.25 inches tall. For some reason I thought the Pi Zero was larger.

The kit that I received comes with everything you need to get going. It includes the Pi Zero W board, a case with three different tops, a mini HDMI to full HDMI dongle, a micro USB to USB A dongle, a micro SD card, and power supply.

The micro SD card is the primary storage device for the Raspberry Pi and one of the first things that you need to do is configure it with an operating system. The card is pre-configured for a New Out Of Box (NOOB) experience, meaning that it comes configured to guide you through installing a variety of different operating systems.

Raspbian is a customized distribution of Debian tailored for the Raspberry Pi, and is what is recommended for new users. There are other OS images configured to set up a Raspberry Pi as a media server, for example, while Raspbian is configured to function like a standard Linux desktop.

I am not really a new user, I first bought a Raspberry Pi 2 a couple of years ago, and therefore I knew that the version of Raspbian that came with the Pi Zero is based on Debian “Jessie” that is not the most current version, that is based on Debian “Stretch” and available for download from raspberrypi.org.

It has taken me several attempts but I have finally got Debian “Stretch” installed and current on my Pi Zero.

First, I did the default install, but when so fast through the menus that I failed to configure the Pi Zero to connect to my WiFi network, consequently the installed failed in its attempts to obtain the latest version of files resulting in being half complete.

The good thing about Raspberry Pi is that it is intended for experimentation and failures, which leads to learning, and the Raspberry Pi was developed first and foremost as a learning tool. To start over all I need to do is press the Shift key while starting up to put the Pi in Recovery mode from which I can start the install process over.

This time, I paid more attention and configured NOOB to connect to my WiFi network, and that resulted in seeing more options, including one for installing Debian Stretch, which was labeled as Recommended, and so that is what I selected.

Alas, the computer gods were not being cooperative and while downloading the install files my home network lost connection to the Internet. After fixing the Internet connection I attempted to restart the installation, but ended up in the end with kernel panics and other errors that lead me to believe I had another bad installation. Strike 2.

Next, I thought I would revert back to the Jessie version of Raspbian that is on the SD card and then do a dist-upgrade. I seemed to make more progress, succeeding in getting Jessie installed but then the attempt to do the dist-upgrade was taking too long and I grew impatient and cancelled it.

At this point I decided to download the version of Raspbian Stretch to my MacBook and flash it on to the storage card using Etcher.io. I noted earlier that the kit that I got came with a case, and I installed the Pi Zero into that case that has cutouts for the HDMI, USB, and power ports, but oddly it does not have a cutout for the SD card. To remove the SD card you have to remove the Pi Zero from the case, which at least is easy to do by gently pushing the board up from the bottom via an exposed slot.

The recommendation I would make to others is to get the Pi Zero all configured before installing it in the case, just to make life a little easier if you end up having to remove the SD card several times.

After downloading the OS image, flashing it with Etcher, inserting it in the Pi Zero and booting it up, I ended up at the Pixel desktop in which I could then customize how I want the Pi Zero to boot. In my case I changed the default password, configured it for access via SSH, and set it to boot to the command line rather than the full desktop.

After rebooting so that the SSH settings could take affect, I was ready for the final pass at upgrading (apt-get upgrade) which I did, but ended up with another kernel panic when it attempted to upgrade openssh. Another reboot later, I did another apt-get upgrade and was instructed to run dpkg –configure -a to finish the upgrade that was interrupted and finally ended up with a complete install.

For the moment I am not sure exactly how I will be using the Pi Zero, but there are many projects available on the Internet. I have a Pi 2 and Pi 3 running nearly full time in my home office.

My Pi 2, which is mounted to a seven inch display, is serving as a desk clock that displays my diary for the next three days, the local weather and four day forecast, and CNN newsfeed over pictures that rotate every couple of minutes. The display power cord is connected to a smart plug controlled by a motion sensor that turns it and a light off when there is no movement near my desk for ten minutes.

The Pi 3 is running an in-house instance of River5, which I use to follow my RSS subscriptions. I also use it to test a variety of Nodejs apps, and I have even experimented with running Mac OS 7 in emulation. 

One of my goals for the holiday break is to complete a course on Nodejs, and because I am a bit crazy, I want to do this using my iPad Pro, which I am trying to use as my primary personal computer.

Unfortunately, there is no app/container for Nodejs on iOS, so the only way I can run the code I am writing is to use another computer. In this case I am using a server running Ubuntu, which I access using an ssh client on my iPad.

To write the code I could use any text editor, but a code editor, or IDE, is more useful. Since I am accessing a server via ssh, I could simply use vim or nano on that server, but that requires a persistent connection to the server and I want to first try writing the code in an app on my iPad.

I did some research about apps for programming on an iPad and found Coda to be the most recommended, but relative to most iOS apps, the $25 price of Coda is really expensive. However, I justified the purchase to myself by noting that Coda also has an SSH client, so it is basically two apps in one making the price tag more palatable.

I am using Git/Github to move the code I am writing on the iPad to the Ubuntu server. First, I created a repository under my Github account. Next, I set up remote access to that repository using WorkingCopy, which most consider the best Git app for iOS.

First, I set up a terminal session to the Ubuntu server in Coda. Next, I created a folder in Coda to store the JavaScript files that I am writing. To test my apps, I export the JavaScript file from Coda to WorkingCopy, commit and then push the changes in WorkingCopy, and finally go back to the SSH terminal in Coda to run the script.

Obviously, while I can write scripts while disconnected, I cannot check them into Github nor run them without a connection to the Internet. However, the reality is that I am hardly ever not connected, so I think this workflow will work pretty well. 

A few weeks ago I received an invite to Micro.blog, having requested one earlier in the year. It has now opened up to everyone, so if you are interested in it, here is what I have learned about it so far.

Micro.blog, developed by Manton Reece, is a cross between Twitter and blogs. Reece started the project on Kickstarter. The idea is to provide a way for users to own short forms of web content. Overall I like the service and plan to continue to use it, although the one thing I think it is missing as a blogging platform is the idea of archives. (Or at least I haven't discovered it if it exists.)

Like Twitter, Micro.blog has a timeline, where you see items posted by other Micro.blog users that you are following. You can favorite items, see mentions, and discover other users on separate tabs. 

New posts are published to a timeline (like Twitter), a profile page, and a blog. Items that you share into Micro.blog are only published to the timeline, you have to explicitly create new blog posts via the app. You can configure RSS and JSON feeds to automatically post to the timeline, but those items are not published to your blog.

The app is available on the web, and for OS X and iOS. An API exists so that anyone can write an app for other platforms, and I believe an Android app is in development. To grow, Micro.blog needs at a minimum an Android app, in my opinion. If you use a Mac and use MarsEdit, you can use it to post items to micro.blog.

Blogs are hosted and cost $5 per month. It is a little unclear to me whether you can use the timeline (Twitter) features and not host a micro.blog. The URL for blogs are [username].micro.blog by default, which is pretty nice, but you can map a domain name you own pretty easily. 

There are six default styles for your blog, but you can create your own style via CSS. You also have the ability to mirror your blog to a Github site, as a backup. 

I find the hosted aspect of Micro.blog a bit inconsistent with my understanding of Indieweb. As far as I can tell, the only way to get your data out of micro.blog is to mirror it to Github, which is better than nothing, but to me not the same as publishing directly to a web server under one's control. 

What I like about micro.blog is the clean layout and lack of advertising, although you are paying $5 per month. The iOS app is simple to use, and makes creating new blog posts on an iPhone and iPad easier than with Wordpress. Frankly, I like having both web and tablet apps, unlike just one or the other as is the case with 1999.io that I use for this site.

Another aspect of interest to me, but which I have not looked into, is the plan for SSL of hosted sites. Given the push to SSL, the value of which for blogs is debatable, I like the idea of an easy way to set up SSL. 

My one nit with micro.blog at this time is that I think the blogging component needs to support archives. Right now the blog is one continual thread you page through, which means things you write will be harder to find over time. I would like micro.blog to provide a way to consolidate a month's worth of posts to an index page, such as you find on my site here. Most users appear to be using their micro.blog as an index to other blog sites that has archival features.

Now that micro.blog has opened up, I expect to see more people checking it out. Whether they decide to pay the money to stay around after the first month, only time will tell. The question is, will people stick to it for the network that will form around the timeline, or because it is an easy to use blogging platform? 

I found a service called Skeddly that provides an easy way to automate and schedule administrative tasks in AWS, including creating Lightsail snapshots. Skeddly has a free usage tier that provides creating 20 snapshots per month, which should cover me. Each additional snapshot beyond the 20 costs $0.15. I've signed up for the 30 day free trial.

Yesterday I did more research about backing up or exporting the data that I put in to Monica. Monica does have an Export Data function that exports to SQL. From what I can tell, it creates an extract or copy of the MySQL database. 

The problem is that Monica doesn't have an easy way for importing data back in. The Import Data function exports the vCard format to provide a way to move data from Google and Outlook to Monica. The export function is intended for moving from the hosted version of Monica to your own instance and the developers assume that if you ware going to host Monica then you will know enough about MySQL to figure out how to import the data, but I don't, yet. I've entered a feature request to provide a way to backup data.

Another option is to backup the server, and for that I can use Lightsail's instance snapshots. It is pretty easy to create a snapshot, either using the portal of the AWS CLI, but if I were to go this route and I wanted regular backups I would need to do scripting to create a snapshot on a nightly basis. 

The other concern I have with snapshots is that I can't tell how much I am going to be billed for them. Amazon says it will charge $0.05 per GB month for storage, and if you only have one snapshot then you pay for the entire disk. Subsequent snapshots are only charged for deltas. Consequently, the baseline snapshot of the 20 GB disk cost me $1.00 per month, subsequent deltas may cost pennies.

For the moment, creating Lightsail instance snapshots appears to be the simplest approach to backing up data. I have found an AWS Lambda script that automates the creation of snapshots.

I've created a server in Amazon Lightsail to host Monica. Lightsail is a user-friendly version of AWS, with easier to understand pricing. I've provisioned the lowest cost virtual server using Debian Jessie.

I initially tried to implement the instructions for hosting on Debian Stretch but the instructions didn't translate backward to Jessie perfectly, so I aborted that attempt. My next attempt using the Docker instructions was successful. When I installed Monica earlier this year on my Mintbox I used Docker. I got the instructions up and running pretty quickly. Before I go further with entering data I want to understand how to backup data given that it is being stored in a container running mysql.

This is my second go with Monica, when I first learned of it I installed the Dockerized version on a PC in my home office. I stopped using after a while, mostly because since it was hosted on a computer on my home network it wasn't accessible to me outside the home. 

My interest in personal relationship management comes from my need to store information about people with whom I work with that goes beyond what you normally store in an address book. The question at hand is, do I need structured data such as you put in a database like Monica, or do I need unstructured data like you put in Evernote?

My first instinct is unstructured, built in to my electronic notebooks that I use for capturing information. The ideal for me is something like a Wiki where I can create pages for each person with whom I interact that contains their information, and then when I write notes and enter that person's name, a link is automatically created to that information page. I know VoodooPad can do this, but it only runs on OS X and I need something that runs in Windows and iOS.

Evernote and OneNote both provide ways to link to pages, but they aren't automatic and therefore take too much time to create the network of links. The other challenge is that I am writing most of my notes in digital ink in OneNote so doing what I want would require OCR and linking, which is not something I hold out hope in finding any time soon. 

Getting some technical items cleared out today. I've set up a backup of my micro.blog and now I am playing a bit more with the viewport setting here, let's see if this helps. Changed the template to just set the initial view, removing device width. Changing to .5 went in the wrong direction. Let's double and see what happens. Now, back a little bit. And a little more. Now with a personal copy of the CSS. 

I am getting closer. Now it is rendering nice on my iPad mini, but not so on the iPad Pro. 

And boom! Success after bumping up maevice-width to 1024px. Now this site should be rendering correctly on all size tablets.

Do I need to republish all pages? Going to anyway. 

From November 27 thru December 7, 2017, I was in Bermuda for a business trip. Here are what I will remember most from this trip.

Narrow, winding roads. The taxi ride from the airport to hotel was my first introduction to why one cannot rent a car in Bermuda. Not only do the taxi cabs want to retain their business, but the roads require experience driving. Scooters are the predominant mode of transportation. Driving is on the left side of the road, steering wheel on the right.

Beautiful weather. Sunshine was prominent with some clouds in the morning, usually clearing during the day. It was noticeably hot and humid the first couple of days, but the humidity gave way to warm breezes during the rest of my stay. I spent time most evenings sitting on the patio of my hotel room, overlooking the harbor.

Walking. The offices in which I was working was a mile away from the hotel, and we walked there and back and every where else. On Saturday I walked 20,000 steps.

Hills. For some reason I was not expecting as many hills as there are in Bermuda. I should have known given the island was formed by volcanoes.

Expensive. It is easy to spend $20 on lunch and double that or more on dinner. Just about everything is imported, which contributes to the higher prices.

While the reason why I was there was business, I did spend a weekend and got to do some sightseeing. We took a ferry to the Royal Navy Dockyard, and spent about three hours walking around. November/December is “off peak” travel time to Bermuda, so there weren’t as many tourists as one might expect. The Dockyard was not crowded, which made for comfortable walking but probably dampened the energy.

The trip was very satisfying for some obvious and not so obvious reasons, and I hope to be able to go back again some day.

I am back home in Michigan after spending the last ten days in Bermuda. Yesterday my day started with sunshine and 65 degrees and ended in the dark with 26 degrees.