Working Towards a Do Less, Automate More Workflow ⭐

The other day, my wife emailed me and asked if I could send her a full-sized version of a map of a walk we had created for our book Eternal Spring. I create these maps in Affinity, and because I want them to look as much like Google Maps as possible–meaning, exactly like Google Maps minus all the extraneous locations and text–I use the same fonts that Google uses. (They’re available for free on the web if you’re curious.) I keep high-resolution original versions of these map files and other images we use, and the versions that go into the book are exported to be smaller. But Stephanie posts to our website and wanted to put the walk on there but with a higher-quality version of the map.

This request triggered two “oh, right” moments. The first was tied to the Internet connectivity problem we had this past month in Mexico, where I briefly wondered whether I had the original map files anywhere easily accessible (I did). And the second was related to a mini-project I’d been thinking about anyway but started actively working on because of that problem. Which is difficult to describe in an obvious word or phrase. I guess I think of it as a sort of go-bag for getting up and running on a new PC. Which, as a reviewer, I do all too often.

In short, I’m always looking to make my workflows more efficient, and the process I undergo when setting up a new PC (or a newly reset PC) is a key workflow. I can do other things while this is happening, and often do, but it’s still a largely manual set of steps, and it takes some time and a bit of attention. I would like to optimize as much of this as possible, but I would also just like to finish it as quickly as possible. Because once this is done, I can simply move between several PCs, as I do all the time, and never worry that the laptop I pick up next is missing something.

Except, of course, this is me. And I always miss something. So even before the weird recent NAS issue, I had been thinking about simplifying this process as much as possible and eliminating the chance that I miss something as much as possible. It’s fair to say that the NAS issue has only added a sense of urgency.

OK, so what does this look like?

The workflow

I have data that I sync to every PC. This is documents and other files that I am either working on actively or may need in the course of a typical day, and it’s separate from archival data I also store–in multiple places–but will never sync. This is what I think of as my “current” work, and as I write this, it’s a set of data across three or four folders that takes up about 15 GB on disk. This can be in any cloud storage service, though I’d been using Synology Drive until the issue noted above. Right now, I’m using Google Drive, but I could also use OneDrive, which, granted, has some appeal because it’s built right into Windows 11.

I also have a set of apps that I install on every PC, and I automate that process with the Windows Package Manager (winget), a command-line interface (CLI) that I’ve adapted over the years and changed yet again in the past few months in the course of writing De-Enshittify Windows 11 and the article Switcher 2026: Minimizing the Microsoft in Windows 11 ⭐. In both cases, there is a new emphasis on using a local account instead of a Microsoft account (MSA). And while I do that with roughly half the PCs I use, understanding how Windows 11 behaves differently in both scenarios is very useful to me. It’s played a big role with the big Windows 11 Field Guide rewrite I’ve been working on this summer, for example.

There is also a set of configuration changes I make to Windows 11 and to the apps I use, and this is the grayest of areas right now. I am investigating how I can automate all that configuration as it is the most manual and tedious of the work I need to do when bringing up a new computer, but this is the area in which I am the furthest from a suitable solution. For now, I just do a lot of this manually.

There are probably many ways to improve this process, meaning to make it more efficient. But my simple brain has latched onto two general ideas for now:

• Do less. If I could install and configure fewer apps and/or sync less data, it would happen more quickly.
• Automate more. The only part of this process I really automate now is the winget-based bulk installation of all the apps I need. Configuring most of those apps is a manual process–there are a few minor exceptions–as is configuring Windows 11 itself.

To get more specific, here’s how I do things. This is something I’ve documented before, probably many times, though it has evolved a lot over the years. So this is the most up-to-date explanation. For now.

There’s this PC. It’s either new, as with a review unit from a PC maker, or it was just reset. Either way, the process of bringing this up goes like so:

• I step through the Windows Setup Out-of-Box Experience (OOBE) and I either sign in to my Microsoft account (MSA) or just create a local account. This is mostly immaterial with exceptions noted in Switcher 2026: Minimizing the Microsoft in Windows 11 ⭐.
• I install all the updates in Windows Update, the Microsoft Store, and the Windows Package Manager (winget), rebooting as often as necessary. And I let OneDrive do that initial sync while preventing Folder backup from happening when I’m given the option to do so.
• I bulk install the apps I need using winget, but per that above linked article, I now exclusively use the winget (web) repository versions of each app, and not the Microsoft Store versions. For years, I’ve maintained a script or text file that bulk installs everything, and this is now simpler and easier to read.
• Once the apps are all installed, I have to configure each. This is the part that’s not automated and I wish it was. Every app is different. Some just require a sign-in, and some of that will be a Google account sign-in, while others will be a normal username/password combo or a passkey that I need my password manager (Proton Pass) and/or 2FA (Proton Authenticator) app for. So there’s an order to these things. Bringing up a web browser immediately makes sense since I can sign into the Proton Pass extension, which streamlines later app configurations. (Many apps these days will launch a web page so you can sign in.) Getting my storage service installed (if needed) and configured early makes sense too since I need to make certain folders available offline, and they need time to sync.
• This whole process takes a few hours, during which OneDrive syncs (just the stubs), my storage service (Synology Drive for the past year, Google Drive at the moment) syncs, and I bring up the installed apps one by one and configure them accordingly. And so at the end of the process, I’m good to go. Except, of course, that I’m not. Not entirely.

As noted, one problem is that I always forget something. Or, more realistically, many somethings. And I discover this, across multiple PCs, at different times and as I use each of the PCs. I mentioned my wife’s request for a high-resolution image of a map I had created. The low-resolution version is in my Books folder, which syncs everywhere. The high-resolution original is in my (badly named) To-do folder, which also syncs everywhere. Simple.

But this is how things always go down when it comes to these map files: I open File Explorer and navigate to where the file is, and find it. I open it in Affinity and … it doesn’t look right. And the reason it doesn’t look right is that I used a Google font for all the text, and that Google font is not installed on this PC, which I likely just brought up. (Or not, it doesn’t matter.) Unless I’m actively working on a map image, I don’t think about the font(s) I need, and they’re not included in Windows 11. So then I have to think about where those files are located. They’re in a different place, and once I find them, I have to install them, go back to that map file, and then reopen it in Affinity, where now it looks correct. So I output it to a higher-resolution version and email it to my wife in this case.

This doesn’t take a lot of time per se, but it’s just one of dozens of examples of this kind of thing. And so I started rethinking how I store data. Again.

Improving the workflow for apps

To goal here is what I wrote above. Do less. Automate more. But the little wrinkle is that I can also semi-automate parts of it, too, and make certain things easier and/or faster. I can put all these pieces in a single place so that I can step through them each in turn as part of the configuration process and gain some minor semblance of automation. Getting there is a moving target. But that’s what I’m working on now.

A few specific examples.

I have an Apps folder in OneDrive, Google Drive, and Synology Drive. Until recently, it looked like this.

There was a lot of cruft in there. A lot of things I will never need or use, out-of-date apps, whatever. There are also little automations in there, like scripts, Registry files for certain settings (like disabling the Caps Lock key on the keyboard), product keys, and whatever else. So I moved most of that crap out of there, leaving only a small subset of what used to be there. It takes up just 16 MB on disk, so I can just sync the entire thing to the PC. (And I should probably rename this folder, since it’s not really “apps” anymore, but rather “things I do up front on a newly configured PC.” But muscle memory is muscle memory. We’ll see.)

This is what’s in there:

• Fonts to install. I just open the folder, select all the files, right-click, and choose “Install.”
• Typora. This is the folder described below, tied to the app I use to write.
• A Terminal configuration file (JSON) so that app looks/works exactly how I want it to.
• A Bonjourr configuration file (JSON) so my preferred New tab experience is available if it doesn’t auto-sync with whatever browser(s) I’m using.
• My winget installation script, which is much simpler now and described below.
• Registry files that disable Caps Lock, remove duplicate removable storage drives from the File Explorer navigation pane, remove Gallery from the File Explorer navigation pane, and remove OneDrive from the File Explorer navigation pane.

And that’s it. It will grow in time, but it’s simple.

Some of the next steps are convoluted. For example, one of my key apps, Typora, is configured with specific fonts (here we go again), themes, and other settings. Typora doesn’t have a way to import or export these things, so rather than re-do all that manually, I have a Typora folder with saved copies of its themes and conf (configuration) folders, and a simple script that copies the right files from my saved folder versions into the installed app’s folders. (This folder used to have the fonts Typora needs, but they’re in the root of Apps now with other fonts to install.)

This is what I have to do to get up and running with Typora.

• I have to make the Apps/Typora folder with all that data noted above available offline in whatever storage service I’m using. That might mean I need to install that storage service’s client first and sync it, or I can just use the version in OneDrive, which also needs to sync.
• I install the app with winget along with whatever else. It’s one of the first apps I install.
• I launch Typora, sign in with my email address and product key (saved in that Typora folder in Apps), which activates it, and then I close the app.
• I used to install the fonts in Typora\fonts, but this now happens first with those other fonts.
• I run the script I created to copy the Typora themes and configuration from the versions saved in Apps\Typora.
• Then, I can run Typora and it comes up configured the way I want. I just switch to the theme I use (which mimics the iA Writer look and feel), and I’m good.

This doesn’t take all that long, though the ability to sync that Apps\Typora folder early on is key. It’s a lot of moving parts, though, and many steps to take. And this is just one app.

Speaking of apps, the bulk app installation script is likewise much simpler than before. What I did was bring up a new PC a month or so ago, look at the apps I had installed on another PC and then used winget search to find the names of the winget (web) repository version of each, which I copied to a text file. It started out something like this:

Google.GoogleDrive
ImputNet.Helium
Brave.Brave
appmakes.Typora
Notion.Notion

And so on, just the names. When I had listed out each app I needed, I then added the winget commands needed to install each app. So it looked something like this:

winget install –silent Google.GoogleDrive;
winget install –silent ImputNet.Helium;
winget install –silent Brave.Brave;
winget install –silent appmakes.Typora;
winget install –silent Notion.Notion;

It’s a little more involved than that because there are many more apps, and I don’t actually need all of them on every PC. But I put the must-haves at the top, broke the apps into groups, and I can open this file anytime I need to install something I either forgot or just happen to need on a given PC.

(Complicating this, there are apps that are not available in winget or not available from the web. For example, on Copilot+ PCs, the first time you open Paint, an app I use every day, it prompts you to install a local AI model. So I now just install that as part of this script. It’s the one item in there that’s from the Store because that’s the only place to get it.)

This will evolve. We just flew to Mexico, and as I was setting up here, it occurred to me that I will need to install OBS Studio for the next Hands-on Windows podcast recording. That’s an app I didn’t think to add to this script. Yet. But that’s how this works. As things occur, I will update accordingly. I would like to somehow automate the configurations, and am looking into that.

Improving the workflow for data

And then there’s the data. The general idea here is a separation of the data that must always sync from that which will not. I have personal and work archives we can ignore and current work documents and files that need to sync, and it doesn’t matter which cloud storage service I use. The total size varies. In my Work folder, I have the following subfolders (and a few others not worth bothering with):

• Book. This contains my current book files, each of which is in GitHub. It’s about 1.2 GB in size.
• Code. This contains my coding projects. I’m taking a break from this at the moment to focus on other things, so I’m not syncing it now.
• To-do. This is the big one, and is misnamed. It contains files related to the things I’m actively working on. More on this below, but it’s about 14 GB at the moment and should perhaps be culled down.
• 2026-07. This a year/month-named folder that I use when I’m done with something. For example, if I am writing a review, the in-progress document is kept in To-do, which syncs everywhere. But when I publish the review, this document is moved to the current year/month folder. When there’s a new month, I add that folder (2026-08 will be next, obviously) and then I archive the old year/month folder. I don’t sync this.

So the amount of data I’m syncing to each PC right now is roughly 15 GB. This feels a bit high, but it used to be worse. I’m trying to simplify things now, but this isn’t the first time. And the way I work–or maybe it’s just the type of work I do–requires a bit of planning unless I want to start syncing a lot more data everywhere.

Consider a book like the Windows 11 Field Guide. There is a GitHub repository that I could sync locally at any time, and to any location. It’s about 1.12 GB in size. (The other three books in the Book folder are much smaller, just 420 MB combined.) But it’s here in Work\Book on whatever storage service I’m using. As I work on this book, as I’ve been doing a lot for the past month or so, I’m working on Markdown and image files in there, and I use Git on the command line to sync them back up to GitHub as I go. (This is necessary to preview or publish full or partial versions of the book through Leanpub.)

But there are two related storage concerns here.

I have current and older versions of some files, the worst being images, that are taking up space in there. The Markdown files that make up the book are in a manuscript folder inside the windows11fieldguide book folder. And there’s an image folder inside that with subfolders for each version–22H2, 23H2, 24H2, and now 2026–so there are numerous duplicates or older images I’m no longer using. This is a good place to cut, of course, but the contents of the 22H2, 23H2, and 24H2 image folders are just 240 MB, so the space savings would be minimal. Plus, a random few of them are still being used in the book. I need to figure that out.

There are also files related to the book that are not part of the book. This isn’t an issue for the Field Guide, as I don’t need a lot in the way of reference. But it’s a very big deal for the Eternal Spring book. In addition to the book folder itself–with its manuscript folder, Markdown files, and images in sub-folders–I have a lot of data to maintain. Like book cover designs, photos of places and other images of all kinds, fonts, maps, and other assets, some of which are for the website too. This data is about 9 GB in size, so it’s pretty big. And it doesn’t make sense to keep all that with book files, meaning the items in GitHub that are used to create and update the actual book.

What I came up with was simple enough, though I may change this soon to further minimize the amount of data I sync to every PC. I have that To-do folder, as noted, and I’ve long had sub-folders in there for Editorials, Reviews, Series, Thurrott.com, and more, plus some in-progress articles in the root of that folder. So I just added to that. I have subfolders for Podcasts now, and Books. And so the reference materials or whatever other files I need for each book are in that folder. That’s a good place for books I may write too, like the Markdown book I plan to write more of soon.

There’s more to this, like other books that are no longer being updated and are now archived for posterity and never need to be synced anywhere. But in trying to minimize the overall size here, I’m looking at this again.

More to come

If you made it this far, apologies. A lot of this is very me-specific. I know there will be at least some “why don’t you just …”-type commentary, and I’m open to it. But this is decades of habits evolved. There are reasons. Sometimes.

I know, or at least believe, that nothing I’m changing here would have prevented the data overage issue I experienced previously. But I’ll know more on that count soon, as I will be bringing up the NAS here in Mexico this week to see if I can tell whether it was at fault–meaning a configuration change I made was at fault–and if so, exactly what it was. Regardless of that, there are further simplifications and efficiencies to come that will impact the two NASes and how I use them. But I’m still sorting through my options there.