Shallow Thoughts : : Jul

Remapping the Caps Lock key on Raspbian

I've remapped my CapsLock key to be another Ctrl key since time immemorial. (Actually, since the ridiculous IBM PC layout replaced the older keyboards that had Ctrl there already, to the left of the A.)

On normal current Debian distros, that's fairly easy: you can edit /etc/default/keyboard to have XKBOPTIONS="ctrl:nocaps.

You might think that would work in Raspbian, since it also has /etc/default/keyboard and raspi-config writes keyboard options to it if you set any (though of course CapsLock isn't among the choices it offers you). But it doesn't work in the PIXEL desktop: there, that key still acts as a Caps Lock.

Apparently lxde (under PIXEL's hood) overrides the keyboard options in /etc/default/keyboard without actually giving you a UI to set them. But you can add your own override by editing ~/.config/lxkeymap.cfg. Make the option line look something like this:

option = ctrl:nocaps

Then when you restart PIXEL, you should have a Control key where CapsLock used to be.

Nambé Lake Grey Jays

This week's hike was to Nambé Lake, high in the Sangre de Cristos above Santa Fe.

It's a gorgeous spot, a clear, shallow mountain lake surrounded by steep rocky slopes up to Lake Peak and Santa Fe Baldy. I assume it's a glacial cirque, though I can't seem to find any confirmation of that online.

There's a raucous local population of grey jays, fearless and curious. One of my hiking companions suggested they'd take food from my hand if I offered. I broke off a bit of my sandwich and offered it, and sure enough, a jay flew right over. Eventually we had three or four of them hanging around our lunch spot.

The rocky slopes are home to pikas, but they're shy and seldom seen. We did see a couple of marmots in the rocks, and I caught a brief glimpse of a small, squirrel-sized head that looked more grey than brown like I'd expect from a rock squirrel. Was it a pika? I'll never know.

We also saw some great flowers. Photos: Nambé Lake Grey Jays.

Translating Markdown to LibreOffice or Word

For the Raspberry Pi Zero W book I'm writing, the publisher, Maker Media, wants submissions in Word format (but stressed that LibreOffice was fine and lots of people there use it, a nice difference from Apress). That's fine ... but when I'm actually writing, I want to be able to work in emacs; I don't want to be distracted fighting with LibreOffice while trying to write.

For the GIMP book, I wrote in plaintext first, and formatted it later. But that means the formatting step took a long time and needed exceptionally thorough proofreading. This time, I decided to experiment with Markdown, so I could add emphasis, section headings, lists and images all without leaving my text editor.

Of course, it would be nice to be able to preview what the formatted version will look like, and that turned out to be easy with a markdown editor called ReText, which has a lovely preview mode, as long as you enable Edit->Use WebKit renderer (I'm not sure why that isn't the default).

Okay, a chapter is written and proofread. The big question: how to get it into the Word format the publisher wants?

First thought: ReText has a File->Export menu. Woohoo -- it offers ODT. So I should be able to export to ODT then open the resulting file in LibreOffice.

Not so much. The resulting LibreOffice document is a mess, with formatting that doesn't look much like the original, and images that are all sorts of random sizes. I started going through it, resizing all the images and fixing the formatting, then realized what a big job it was going to be and decided to investigate other options first.

ReText's Export menu also offers HTML, and the HTML it produces looks quite nice in Firefox. Surely I could open that in LibreOffice, then save it (maybe with a little minor reformatting) as DOCX?

Well, no, at least not directly. It turns out LibreOffice has no obvious way to import an HTML file into a normal text document. If you Open the HTML file, it displays okay (except the images are all tiny thumbnails and need to be resized one by one); but LibreOffice can't save it in any format besides HTML or plaintext. Those are the only formats available in the menu in the Save dialog. LibreOffice also has a Document Converter, but it only converts Office formats, not HTML; and there's no Import... in LO's File. There's a Wizards->Web Page, but it's geared to creating a new web page and saving as HTML, not importing an existing HTML-formatted document.

But eventually I discovered that if I "Create a new Text Document" in LibreOffice, I can Select All and Copy in Firefox, followed by Paste into Libre Office. It works great. All the images are the correct size, the formatting is correct needing almost no corrections, and LibreOffice can save it as DOCX, ODT or whatever I need.

Image Captions

I mentioned that the document needed almost no corrections. The exception is captions. Images in a book need captions and figure numbers, unlike images in HTML.

Markdown specifies images as

![Image description][path/to/image.jpg)

Unfortunately, the Image description part is only visible as a mouseover, which only works if you're exporting to a format intended for a web browser that runs on desktop and laptop computers. It's no help in making a visible caption for print, or for tablets or phones that don't have mouseover. And the mouseover text disappears completely when you paste the document from Firefox into LibreOffice.

I also tried making a table with the image above and the caption underneath. But I found it looked just as good in ReText, and much better in HTML, just to add a new paragraph of italics below the image:

![][path/to/image.jpg)

*Image description here*

That looks pretty nice in a browser or when pasted into LibreOffice. But before submitting a chapter, I changed them into real LibreOffice captions.

In LibreOffice, right-click on the image; Add Caption is in the context menu. It can even add numbers automatically. It initially wants to call every caption "Illustration" (e.g. "Illustration 1", "Illustration 2" and so on), and strangely, "Figure" isn't one of the available alternatives; but you can edit the category and change it to Figure, and that persists for the rest of the document, helpfully numbering all your figures in order. The caption dialog when you add each caption always says that the caption will be "Illustration 1: (whatever you typed)" even if it's the fourteenth image you've captioned; but when you dismiss the dialog it shows up correctly as Figure 14, not as a fourteenth Figure 1.

The only problem arises if you have to insert a new image in the middle of a chapter. If you do that, you end up with two Figure 6 (or whatever the number is) and it's not clear how to persuade LibreOffice to start over with its renumbering. You can fix it if you remove all the captions and start over, but ugh. I never found a better way, and web searches on LibreOffice caption numbers suggest this is a perennial source of frustration with LibreOffice.

The bright side: struggling with captions in LibreOffice convinced me that I made the right choice to do most of my work in emacs and markdown!

Writing a Book on the Raspberry Pi Zero W

It's official: I'm working on another book!

This one will be much shorter than Beginning GIMP. It's a mini-book for Make Media on the Raspberry Pi Zero W and some fun projects you can build with it.

I don't want to give too much away at this early stage, but I predict it will include light shows, temperature sensors, control of household devices, Twitter access and web scraping. And lots of code samples.

I'll be posting more about the book, and about various Raspberry Pi Zero W projects I'm exploring during the course of writing it. But for now ... if you'll excuse me, I have a chapter that's due today, and a string of addressable LEDs sitting on my desk calling out to be played with as part of the next chapter.