Irreal

Over at Meta Redux, Bozhidar Batsov, who will need no introduction to Emacs users, has a nice post on how to make a screencast from within Emacs. The result is actually an animated GIF but the results in his demo video look very good. The nice thing about his method is that it takes a screen capture only when you do something so the resulting files are smaller and there are no pauses when you’re not doing something.

The screencasts are made with the emacs-gif-screencast package that’s available on Melpa. You’ll need a couple of extra packages as well: Imagemagick and Gifsicle both of which are available on HomeBrew for MacOS users and are probably already installed on Linux—if not prebuilt binaries or source are available at the above links.

There are a few weaknesses, which Batsov discusses but nothing serious and he concludes that it’s a “pretty cool tool” so if you’re looking for any easy way to make a screencast from within Emacs you should definitely give it a look. You won’t be able to make the type of video that, say, Mike Zamansky produces but if you need a quick screencast to demo some process, emacs-gif-screencast is ideal.

Eli Zaretskii writes on the Emacs Development list that he’s created the Emacs 27 release branch. Next up is the Emacs 27.1 pretest, which he hopes to release soon. He’s also set the version number of Emacs on the Master branch to 28.0.50.

Emacs 27 is finally in sight. In the meantime, we can start obsessing about Emacs 28. If you haven’t already, take a look at John Wiegley’s video on what’s coming in Emacs 27.

Regular readers know that I’m fascinated by the ways that non-technical folks make use of Emacs. Despite our prejudiced assumptions, they are often sophisticated and subtle in the ways they leverage Emacs. Recently, Eric Fraga tweeted a pointer to a long article by Kieran Healy on using plain text and Emacs for writing in the social sciences.

The article is mainly addressed to the beginning graduate school student and makes the case that if you are going to do any quantitative work in your research then you’re better off using plain text and a text editor rather than a word processor for writing your papers and books. Most Irreal readers won’t find this the slightest bit controversial and will need no convincing but social scientists are different. Their culture and training has led them to believe that writing papers is something done with Word. Their professional journals agree. With that in mind, it’s easy to see why Haley’s position can be a hard sell.

Healy is mostly concerned with workflow and although he is an Emacs user, he doesn’t insist that it’s the only good solution. He suggests several alternatives but does note that Emacs is a powerful tool that can make writing easier. He assumes that “quantitative” means “statistical” and that his readers will be using R. His workflow involves writing in Markdown and using knitr to evaluate the R code blocks and insert the results in much the same way that Org and Babel work. Org and Babel would provide a more seamless and slightly easier solution so I’m not sure why he doesn’t use them. Healy is a Full Professor at Duke so Org probably wasn’t available when he started and he’s just stuck with his “good enough” solution. The article discusses his entire tool chain with particular attention to revision control and backups so it’s about more than just what editor and markup language to use.

As I said, the article is long—a PDF version runs to 49 pages (although with generous margins)—but it’s interesting and makes many good points even if you don’t agree with all his choices of tools. I recommend reading the PDF version because it has the footnotes and is a single document rather than a series of pages.

In another of the videos from Emacs Conference 2019, John Wiegley talks about what’s coming in Emacs 27. Wiegley is still the official maintainer but Eli Zaretskii stepped up to do the heavy lifting when Wiegley’s job intervened. Wiegley says that Zaretskii is now the de facto maintainer but that he was not available to give the talk so Wiegley made the presentation.

There are several interesting changes coming. Lispers will be happy to learn that Elisp will finally be getting BIGINT support. It still doesn’t have a full numerical tower but BIGINT support is probably the most important missing piece.

On the display side, Emacs 27 is getting support for Harfbuzz, Cairo, and better ligature support. Another display-like feature is support for tabs. This is not just tabs in the browser sense—because who needs that in Emacs—but a mechanism for maintaining window configurations and switching between them.

Some other interesting enhancements are the portable dumper (a long time coming but we’re finally free of depending on glibc hacks to build our core image), initial support for the XDG directory configuration standard, so-long-mode to help deal with very long lines slowing down Emacs, and gettext integration to support internationalization of error messages and the like.

You can get more details by watching Wiegley’s talk but, unfortunately, the audio dropped out several times during his presentation. Still, every serious Emacser will want to watch the video to see what’s coming in Emacs 27.

I’m finally finding time to watch more of the videos from the Emacs Conference 2019. Howard Abrams gave an interesting talk on using Emacs as his shell and scripting language. As he says, it’s just him trying out some ideas meant for his personal use.

He starts out with a typical shell pipeline and wonders how the same thing might be done in Emacs. Why would you want to do that? The problem is that a pipeline’s intermediate results are pretty much opaque. You see the input and you see the output but not what’s happening in the middle. By using Emacs, you can see what the data looks like at every step. Abrams also says that in general Emacs tools are superior to the shell so you gain there too. It turns out not to be very difficult to use Emacs interactively to do a task that you might otherwise use a pipeline for. Abrams demonstrates this by using Emacs to do the same thing as his original pipeline.

In the second half of the talk, Abrams considers how you might use Elisp as a scripting language. He puts together some macros and functions that lets him write vaguely shell-like scripts but within Emacs. The advantage to this approach is that he has the full power of Emacs available whenever he needs it.

I can imagine circumstances where I might use Emacs interactively instead of a pipeline but I don’t think I’d bother with the scripting. I’m happy to use a shell script if that’s what the job requires. Still, Abrams’ approach was interesting and definitely worth taking a look at. You might even decided to adopt his approach.

Sadly, there were technical problems and the video feed was lost in the middle of the talk. That’s not as serious as it could have been because it happened near the end of the talk proper and mostly affected the Q&A. Secondly, Abrams has provided a sort of transcript of the talk complete with the slides. That makes it pretty easy to follow along in the part without video. As usual with Abrams’ videos, you’ll probably learns some new things and it is definitely worth your while.