Another Emacs blog post. If this gets me thinking about blogging and tech I am good with it and I hope you are.

I used Elfeed for a long time. What I didn't realise is Emacs has been shipping with one for a long time. Enter NewsTicker.

It seems OK. Configuration is fairly easy:

(setq newsticker-url-list-defaults nil)

;; Define your custom news feeds
(setq newsticker-url-list
      '(("Spider flock" "https://spiderflock.com/feed.xml")
        ("Null program" "http://prog21.dadgum.com/atom.xml" )
        ("Sacha emacs"  "https://sachachua.com/blog/feed/")
        ("positech"     "http://positech.co.uk/cliffsblog/?feed=rss2")))

Keyboard shortcuts are quite reasonable. It doesn't feel amazing but it does seem usable enough to just use it and forget about alternatives. It not been heavily tests but I like the idea of having a hassle free rss reader.

A curated set of websites appeals to me. Humans doing human things and writing about them - warts and all. It would be wonderful to kick back and read some great blogs again. I am sure they still exist and as they don't hit the mark when it comes to making money people can just be themselves. Including their typos, spelling mistakes and rambling exposition.

Changing subjects but staying with Emacs. I previously mentioned I had two Emacs packages installed. One is a Tron theme and the other is gptel. Gptel is a package for using LLM, both locally and cloud hosted one.

I played with ollama to have one running locally but it made my machine crawl so moved to the free tier of the cloud one. I pretty much use it as alternative to googling. It lets me stay inside Emacs rather than jumping browser and provides answer quite quickly.

I would like to see self hosted models dominate this area but that's not something that will happen soon, if at all.

Yes I know me talking about wanting to find great human written blogs and then saying I am making use of LLM's may seem a little odd. I guess it is what it is.

I decided I wanted to work out how to create a new window in Emacs and send some text to it. After that I wanted to have it toggle in and out when I pressed a button. Nothing to exciting and I am not exactly sure what I am going to do with it other than put it on this post.

(setq log-window nil)

(defun sf-toggle-window ()
  (interactive)
  (if (not log-window)
      (progn
        (let ((buf (get-buffer-create "log"))
              (win (split-window nil nil 'right)))
          (set-window-buffer win buf)
          (setq log-window win)
          ;;  (select-window win)  ; Optional: moves focus to the new window
          )
        )
    (progn
      (delete-window log-window)
      (setq log-window nil)
      )
    )
  )

(sf-toggle-window)

(keymap-global-set  "<f2>" 'sf-toggle-window)

(defun sf-write-to-log (str)
  (interactive)
  (let ((buf (get-buffer-create "log")))
    (when (not log-window)
      (sf-toggle-window))
    (with-current-buffer buf
     (save-excursion
      (goto-char (point-max))
      (insert str))))

  )

(sf-write-to-log "hello sailor\n")

One thing I have not done is make it use special mode. A mode for read only output. This is meant for output from external tools or things like the Messages buffer.

It has a set key bindings including pressing q to close the window, but not the buffer. I think I will save that little area of Emacs for another adventure.

After on and off using Emacs for more years than I care to admit I think I finally understand it! Having approached Emacs as just another editor it's slowly becoming clear there may be a better approach. Until now my Emacs usage was get vim bindings set up with Evil and attempt to maximize my productivity in Emacs from the point of view of editing.

My level of abstraction was limited to editing and writing. I used Emacs to solve problems at that level, limiting my understanding of Emacs. It was like I was stuck in a local maxima. Instead of understanding Emacs I was attempting to configure it as quickly as possible and think about other problems.

For past couple of weeks I have been attempting to use Emacs in a way that lets me learn about it. I want to maximize the time I spend in Emacs and treat Emacs as the default application when using a computer.

How does one approach Emacs in this way. For me it started with realizing that installing external packages meant I was learning to configure rather than build and understand. Sure they made things easy but took away my opportunity to learn. Often, taking the easy path now is not the best path over a longer period.

I have not totally closed the door to them but I am trying to keep them below five. At this moment the count stands at two with one of them being the Tron theme Yes I am still configuring things but I find myself closer to the core of Emacs and that matters.

I have to admit it helps that Emacs is at version 30 and seems to have reached a point where a lot of features are shipped with it rather than needing external packages.

Emacs lisp is now a given. Gradually building a strong foundation feels like the best way to shape Emacs to fit me.

It's not entirely clear whether I will keep at this or if it is even a good idea. I may be better off building an environment around the shell and small tools like vim/lazygit and so on. But it's fun so over the coming months I am happy to see where it takes me.

Yes I know there are hundreds of YouTube videos and blog posts that tell you to take this approach. I just never did but once you do...

My very unexpected journey into Emacs continues...

As I mentioned previously I have been a user of Emacs for decades but never taken the time to truly learn it. I would just use a hacked together init file pulling in packages without much thought. Occasionally I would follow a blog post and it would improve. Ever few years I would try something like Doom emacs but find a reason to not like it.

Complicated development I usually do with an IDE, these days Clion.

After playing with Omarchy and realising it was providing an awesome shell orientated approach to using a computer it seemed sensible to build my computing life around the shell using TUI style apps and things like neovim.

I didn't bother installing my usual emacs config and just left it bare bones. Occasionally I would use it to edit a file but nothing more. One evening I copied over my basic config to make the toolbar disappear but didn't download any packages. My plan was really to just use it to write blog posts. Like this one.

Unexpectedly I found myself really enjoying using Emacs. Enough that I started to read the doc. This lead to me wondering if I could write a function to increase the size of the margins on a window to create a "focused" writing mode.

Turns out that was not that hard, from there creating a minor mode and configuring that was also easy.

I don't think it is perfect but my current goal is to keep downloaded packages to a minimum so this serves as an alternative to doing that.

I think I finally get what they mean by you shape emacs into the editor you want.

Will this continue I am not sure.

There are much better packages available that do this type of thing so actually publishing for all to use isn't worth it. It does serve as to how small a minor mode can be. Here's the code.

(make-variable-buffer-local
 (defvar focus-mode nil))

(make-variable-buffer-local
 (defvar prev-left-margin-width nil))

(make-variable-buffer-local
 (defvar prev-right-margin-width nil))

(defun store-margins ()
  (setq prev-left-margin-width left-margin-width)
  (setq prev-right-margin-width right-margin-width)
  ) 

(defun set-margins-to (left-width right-width)
  (setq left-margin-width left-width)
  (setq right-margin-width right-width)
  (set-window-buffer nil (current-buffer))
  )

(defun restore-margins ()
  (interactive)
  (when  prev-right-margin-width
    (progn
      (set-margins-to prev-left-margin-width prev-right-margin-width)
      (setq prev-left-margin-width nil)
      (setq prev-right-margin-width nil)      
      )))
  

(defun toggle-focus ()
;  (message "State (%s) (%s) (%s)" focus-mode
;           prev-left-margin-width
;           prev-right-margin-width)
  (if (not focus-mode) 
      (let ((margin-width (/ (- (window-body-width) (current-fill-column)) 2)))
        (when (> margin-width 1) (progn
          (store-margins)
          
          (setq left-margin-width margin-width)
          (setq right-margin-width margin-width)
          (set-window-buffer nil (current-buffer))
          (setq focus-mode t)
          )))
                                        ; else
    (progn
      (restore-margins)
      (setq focus-mode nil))
    ))

;;; ###autoload
(define-minor-mode sf-focus-mode
  "Lets get focussed."
  :lighter " focus"
  (toggle-focus)
  )

(provide `sf-focus-mode)

It's fair to say my time is still very limited. I say that every time I post. Which is fine as these posts are really for me as they don't really educate of help others. Unless the tribulations of a random guy on the internet provides some escape for others.

Lets talk some random tech:

I have moved over to Omarchy... I know the politics but I am trying to avoid that in my life and Omarchy is, well, nice to use. Perhaps I should say if seems to fit my brain. I could achieve the same using different tech but this is easy. Did I mention my limited free time.

From there I messed around with ollama and installed a few local LLM's. It is nice to play with them but they don't quite feel at the state I would want to use them that much. I am still exploring though, as a self hosted reasonable LLM would be ideal.

Omarchy is a shell orientated distribution, that means I am slowly discovering the joy of TUI applications. LazyGit is and oblivious good starter. I have always been a bit of a command line person so this is a joy.

At first I thought having this well thought out system would make me remove Emacs from my life. Emacs can consume all your time just exploring it and configuring it. It hasn't turned out that way.

I have completely restarted my config file. Leaving only the bare minimum in there to make it usable. I have only added one package on top of the base install. A Tron theme. For the first time in over a decade I am using Emacs without evil installed. It feel nice, almost liberating.

It helps that Omarchy has a great default font.

I am going to try to use the base install as much as possible and just see where it takes me. I have neovim installed if I really miss the vim bindings and Clion if I need the power of an IDE.

I almost wrote about future plans but experience has told me I rarely end up where I expect to go.

I am not heralding my return to blogging. This post is a collection of small bits of tech I thought would be more difficult than they actually are. A little life update is tagged onto the end. Who can resist one of those.

Git Submodules:

I avoided these for such a long time, instead depending on various pack managers.

Usually my dependencies are quite simple. SDL3 and a couple of other things which make them ideal for git submodules. There is perhaps 3 commands in git you need to learn and be aware that there is a config file.

And that's it. Sure you have to figure out integration with your build system. Assuming you are using C++ that is. If you are using rust just use cargo if you can.

Overall very painless and now I am wondering why, for so long, I avoided taking the 10 minutes to learn about them.

Different ssh keys for different sites:

Again there is a simple to use config file which makes this painless. I don't know why I assumed this would be tricky as it isn't

CMake:

I'm not a huge fan of cmake but I have decided to go with the flow on this one, use my energy elsewhere. LLM's are really good at guiding me through what I need to do for my simple use cases. There is something about reading the documentation that makes my brain switch off.

Do I have other blind spots that I should identify?

In terms of life things are trundling along at the moment. I seem to be finding a little more time for technical hobbies but I am not entirely sure where that will take me.

For now I have resubscribed to CLion for the year, mostly as a way of avoiding going down the rabbit hole of getting that perfect Emacs set up! It's also a nice IDE even if sometimes I hit performance problems when using it on Linux. My aging hardware is more than likely the fault here.

I am unsure on what direction to take from here or if the time I seem to have at the moment will persist. I am drawn between working though some technical books or starting a longer term project. Both would be fascinating and fun which it the requirement for this year.

I am lot more settled than any point in my life. While I have some stress day to day I am quite calm. Calmness is a prerequisite for good decisions.

Recently a discussion with a few friends about what characters from films have influenced me left me struggling to think of many. On reflection I tend to get inspired by characters that help/inspire me for some period of time before I grow past what they speak to. That's not to say I out grow them, more I have internalised what a should from them and continue on my way.

This led me to thinking about books/essays both fiction and non fiction. It's too wide a topic that I struggled to make coherent progress towards a post. Clarity on this would have taken to long to attain. Time that is in short abundance. I am sure another log style post will cover that.

Instead I decided to start a list of things to blog about and put "Books/films/essays that have inspired me" on it. That way I can nibble at that topic for new next decade or so.

My first small bite leads me to this quote:

Think of your past as a servant that shaped the present you but it may not be the best friend to take with you into the future.

It's from the 5am Club that I talked about a few times in this blog, most notably in this review.

I deliberately have not reread the original post. It was influential as at the time I was very time constrained and finding it very hard to feel like I was making progress on things that were important to me personally, outside of work and family.

I was fortunate to get ill over the Christmas period just after reading the book which cause me to go to bed very early and sleep in. As I recovered, in time for Christmas, I found myself waking up earlier than usual. Being careful not to stay up late my wake up time shifted to the elusive 5am without too much trouble.

Getting up at 5am and then doing the prescribed exercise/mental self care/learning in the first hour really boosted my day. Having 6-7am clear for personal projects before having to think about the day was awesome. Weekends where even better as I add in 45 minutes playing The Witcher 3 and getting a lot of other things done before the family start to wake.

The two youngest kids started to join me in getting up at 5 am at the weekends. Often cooking cookies and then heading of together to do some Roblox development. They once gave the reason for getting up early as they were left alone to do what they want before anyone else gets up. I was careful not to wreck that for them.

Making sure I was in bed by 9:30 most nights was a reasonable trade. At first it was hard to do late social events but eventually that seemed to just work and the occasional 1am bed time didn't seem perturb me.

Eventually though I slipped and it became more a wake up at 6 am. Yes my productivity slipped but it was a deliberate trade and one I chose and still feels like the right choice.

After splitting up with my partner I found myself spending more time with the two children not at university. The climbing gym was their favourite destination (and mine). We would often find ourselves there quite late. Consistently arriving home late enough to mean getting up a 5 would have left me sleep deprived.

My morning ritual stopped after about a year of feeling really productive.

Why would I still consider this influential? I could argue it is something that I want to return to and I suspect I will as the house empties. That's not the reason though.

According to the 5am club the first hour should be yours but structured into 3 twenty minute sections. First some exercise, this can be a run, weights or perhaps some yoga - your choice. The second section is for mental health, think mediation or journalling of some sort. The final 20 minutes should be devoted to learning, perhaps reading or something more active. Within the first hour you achieved a lot and set the momentum for the day. For me it was very effective. Even writing this makes me want to get back into it.

The middle 20 minutes was where I picked up some great mental health habits. Meditation is awesome. Journalling was a way for me to work though problems and identify improvements. Later it has become a motivator, do things so I have something to write or helps me keep track on things I am learning or tinkering with. Increasingly it is also acting as a memory bank for some great experiences. After not bothering with either mediation or journalling for most of my life even though I aware of their benefits it was the 5am Club book that got me into them.

For that I am both grateful to the author for getting me to try some new tools but am also filled with regret that I didn't try them earlier in my life.

A tertiary effect was it really demonstrated that turn up every day really drives progress.

It's been a while. That reflects my slow progress on what this blog focussed on. I am not about to do another "I'm back" post as these are always left hanging for months with no update.

Instead lets just talks about where I am at. The past few years have been fairly hard on the personal front. Not terrible just a lot to sort and attempting to steer the proverbial ship to calm waters. That is make my life boring so I can select the interesting and fun things that I actually want to do. That sounds like the start of some motivation YouTube video. I think I am there, most problems I now face are external ones as they pop up rather than a back log of problems to sort. Oh god none of that makes sense. Let's go with "Hey guys I am back"!

Tech update:

I have been reading GoldenEye 007 Boss fights book. It is pretty good and partially motivating me to want play with gaming technology again. I am also thinking about kicking off another attempt at making a small programming language. Will I do these I am not sure just thoughts at the moment.

In reality I messing around with python and C making small example programs. I was playing with Rust again but for now I seem to have settled back into C. I am trying to not think too much about actual programming languages I use. Just select the right tool and enjoy getting on with things.

Oh and I have moved over to using Omarchy as my main OS. While it's early days I find it fun to use. And that's not something I get to say about operating systems very often. I am surprised no one has thought of doing this before, or perhaps they did and I never heard about it. Obviously it is aimed at developers. Well worth a try.

PiAmiga is an Amiga emulator comes with nearly every Amiga game ever made. You have to supply the Amiga Roms but otherwise you are good (re)leve the late 80's very early 90's gaming era. I haven't installed it recently but seeing it's now on x64 platforms and I currently have a spare Beelink device I added it to the todo list.

I use the Linux operating system at home, only occasionally using Windows for gaming or, rarely these days, things that just don't work on Linux.

Developing in C or C++ is quite nice on Linux. Mostly the tooling just works. The one exception is Debugging.

Gdb is OK but really I struggle with it, typing gets annoying real quick. It is incredibly powerful but just in my hands. I want something that just work. VS Code has a competent debugger but then so does Clion.

After installing month long trial of Clion and I was debugging my problem in minutes. Eventually I ended up creating a CMake file to help out Clion and using it to edit files.

It's a nice IDE a few quirky differences but an IDE I can use. I still maintain the shell build system and am using lazyGit. Clion has just replaced Zed as an editor and comes with a debugger that is competent from a user experience point of view.

Currently I am just coding up some simple SDL3 experiments in C, nothing earth breaking. Clion trial license allows you to use their equivalent to CodePilot. I haven't used CodePilot but the Clion help is actually quite useful. Sometimes guessing wrong but often getting it close enough. I am writing stuff that has been written many times before by many people so perhaps not the best comparison.

In a week or two I will have to decide whether or to puchase Clion, if I do I am thinking I will add on the AI code helper. I estimate gives me a 10-20% productivity increase when writing code. Nice but I suspect that will drop a bit as I write more complicated code.

What is does get write though make coding a lot more pleasant. Taking some of the boiler plate out of coding. This encourages me to code more which makes it worthwhile.

Clion does seem to remove a lot of friction to coding and now has a Zig plugin which seems well supported even if I have not experimented with it. I keep pondering about moving to Zig.

The Christmas and New Year holiday is nearly over for me. Two of my children are university students so we have another week or two before the house goes quiet and I can start to focus again.

There are many small plans afoot but as is life many will not come to fruition. I feel like I am in a good place at the moment and looking forward to attempting some of the many tasks I am thinking about.

Kaizen is still my theme and I am steadily improving my environment both physical and virtual.

Earlier in the year I decided to try setting goals over a 90 day period. The idea being they are a large enough time to achieve a lot but are shorter than a year so you get the coming up to a deadline productivity boost more often.

It hasn't really worked out. For me that is, but let me explain as it might work out for you.

The amount of time I can devote to my goals is fairly variable. I obviously have a mixture of goals. I am speaking to the more creative/technical side rather than family, health, financial or other personal ones.

This variable amount of time has a time period of about one week. Some weeks I seem to have a lot of time and some my time is a scarce resource. The problem is obviously here.

For me this leads to I will do it next week when I have more time and as the deadline is many weeks away I don't worry and a week passes without progress and then another.

The comment, it appears these goals are not a priority is true. The lack of priority and urgency combined with not distant deadlines means I do not form habits. Some time is consumed with other goals and they seem to be progressing well so it is not like my life is a disaster.

The lack of habit does mean procrastination rather progress on technical projects. I am thinking of moving back to monthly goals, I track others on a monthly cadence and that seems to be working.

Whys have goals? Why not just do what I enjoy as it's not top priority? Goals imply planning and focus on getting what has been added to the goal done.The deadline effect combined with clarity are something I like.

I am also asking myself the Why questions, find your why and everything else falls into place. I'm not very far along but I know I have not found it if I still scrolling through shorts.

Finally, I am asking the question, why should would anyone want to read about productivity tips coming from someone who is openly admitting that I am not feeling productive in my tech hobbies. Am I spending way too much time thinking and writing about the evolution of my systems rather than actually using them?

I do like Kaizen style posts I wrote as they try to focus on this is working but the broader goal setting/process/systems I think I will try to avoid until I have something that is producing results.

Kaizen Update 2

Winter Arcs seem to be a thing at the moment, the idea is to put you head down and focus on an important task or personal improvement over winter and be a better person come spring. Really it is just a marketing term to make you think it is a new idea and drive clicks as you realize it just rehashing existing ideas.

I do like the idea, it appeals. My life doesn't allow that it at the moment but gradual improvement, Kaizen, does seem to be paying off though and it's fun.

I have read most of a book on the Ruby programming languages. Where is this taking me? Well, I am not sure perhaps towards automating lots of things. Ruby-koans is now a work in progres.

I expected Zellij to be a great addition but I was not enjoying neovim as a shell editor. To combat this I took a look at the Zed editor. It's nice and very fast. Of course it breaks the use shell for everything but I am OK with that. The point of Kaizen is to follow the path of improvement and not be hold on to things that are holding you back.

Cutting back coffee has worked quite well. Drinking more than a few cups no longer feels nice. I feel calmer.

If you have read the above it sounds like someone who want to get things done. Really though it you look at it properly you might realize I am a bit directionless. I enjoy learning it is my procrastination technique, perhaps not as bad as YouTube or Netflix but directionless learning is not an efficient use of time.

Why do it? It is to help reform habits, making sure I make time for my self. Any direction is better than stationary as it can at least be nudged in the direction I want to go once I figure out that part.