The Linux Last Mile Problem

Fri 21 July 2023 by R.L. Dane

I was watching The Linux Experiment's latest video entitled "Sometimes, I get tired of Linux," and I realized how similar the issues that he mentioned are to the classic "Last Mile Problem."

First of all, mad props to folks like Nick (a.k.a. "The Linux Experiment"), Veronica Explains, The Linux Cast, Art Reeder and so many others that not only create a lot of really great linux/FOSS-related videos, but upload them to Peertube as well as that other video platform. They do so out of love for the community, as there's absolutely no business model for Peertube as of yet — it's all one big passion project. I have more PeerTube info and recommendations available here for your perusal.

A couple of Nick's points I'd like to consider are:

1) 3:07 - Linux distros' and major projects' updates (such as DEs) have become very piecemeal and less impressive/interesting compared to years ago
2) 8:02 - "The 99% There Status" - So much of the Linux desktops and FOSS programs are very useful and feature-rich, but there are still many features that are incomplete or buggy after so many years.

So, how does this relate to telecom? Well, the big problem with the last mile in telecom is that while it's only one mile, it's one mile multiplied by millions of subscribers. So even through telcos have spent millions of dollars upgrading the network backbone and trunk lines, the last mile is always going to be the bottleneck, because those millions of last miles will always be far more difficult to upgrade than the thousands of miles of backbone and trunk lines.

Regarding Linux, much of the work needed to make Linux usable as a daily-driver desktop has been accomplished, and that which remains is not relevant to all users, nor interesting to all devs. I'm sure there are many users that care deeply about having HDR support (@8:31), for example, but most likely, and equal number (if not more) don't care or don't even know what HDR video is, and most likely, there aren't as many devs that find HDR as attractive a target for their efforts as something like a major UI bug or improvement (or more likely, a pet feature they want to see implemented). That's not to say that HDR isn't important, only that it isn't important to as many people, and therefore remains in the domain of the Last Mile Stagnation (at least, for now).

Another major issue (and this ties into the first point (@3:07) is that companies like RedHat and Canonical which spurred on so much rapid growth and innovation in the 2000s just don't see perfecting the Linux Desktop as a worthy endeavor anymore, or they merely consider what is available today as good enough for their real goals: enterprise license sales and cloud service contracts. You don't need a super fluid and colorful desktop for managing vms. You don't need HDR or fractional scaling, either.

I don't have any concrete figures, but I surmise that much of the development effort going into desktop interfaces today is done by large independent development communities (like KDE), rather than directed by big corporations with deep pockets. With fewer resources (both human and financial) to pour into new features (and more developer-cats to herd into a single direction/objective), it takes much longer to implement new features.