Clay Morton

November 11, 2026 · 7 min read · systems, networking

It's Universal! (Until You Plug It In)

Standards promise us that everything will work with everything all the time, but so often that is not true.

It feels like every couple of years we get a big announcement from all the big companies or some global organization that there is a new universal standard. Boy doesn’t that sound good‽ You start hearing things like “we’re all going to get on the wagon together and start making all our devices have one universal port” or “there will be just one protocol for all smart home devices, and it will just work.” It starts to look and feel like the fragmentation is all done and over with—the press releases said so!—but then you try to plug in that cable or that new IoT device, and suddenly, everything isn’t so universal.

I am a big fan of standards, universal compliance (in some regards), and ultimately I am a fan of ease-of-use for myself and other users. Almost all of my work in all the different areas I am in relies on some set of standards. Whether I create them (e.g. brand guidelines) or an organizational body does (e.g. WCAG). Standards aren’t the problem, but thus far, they also haven’t created a solution. I think “universal” more accurately describes the goal of a technology, not the result.

Connector ≠\neq protocol

USB-C is the best example of connectors that seem like a protocol, since most products and companies have adopted them as the main/only port on their devices. The issue is, they all look the same, but they carry dramatically different information.

A USB-C cable might be capable of moving data at USB 2.0 speeds, or could be Thunderbolt 5 speeds. It may carry 60W of power or 240W. It may carry video and audio, it might not—this is dependent on both the devices supporting DisplayPort over USB-C and the cable needs to support it as well. So, needless to say, two cables can look identical and behave entirely differently, and the only way to tell them apart is likely some printed-on label that will eventually rub off after lots of use.

The “standard” of using USB-C has made it very easy to carry one cable to charge all of our devices and has solved the issue of having lots of ports for different tasks on your device. The simplification is fantastic to look at, because our devices don’t look like electronic Swiss cheese, but it didn’t solve the question(s) that come along with these cables. You have to be very deliberate when choosing a USB-C cable so that you know it will do the task you need it to. Despite the cable being universal, the customer still needs to make sure that they’re getting the right version of the same cable (and probably labeling them too so you don’t get confused).

The “smart” home

If you’ve gone looking for a new smart device, you’ve seen those “works with” logos which are supposed to indicate the smart home system the device is compatible with. You select a device that’s on your platform and when you get home, setup should be simple. That’s the theory, anyway.

In practice, we see devices are only going to work as well as the weakest layer between them and you. A protocol like Matter was supposed to solve the “works with” issue so that all devices were on an even playing field, working with all smart home ecosystems, and making the only bottleneck your network latency. The protocol defines a set of device types/features and every platform then decides how much of that set it will support and how it displays it (i.e. in an app). This creates another hole in the design as well. A light that offers color adjustments on one smart platform may only show a power switch on another, meaning users get a different experience depending on their chosen app. This does not fit the definition of universal at all! So now you have a “certified universal” compatible device that doesn’t fully work with your ecosystem.

Your home network also adds a lot of variability to the performance of your devices. I am staying with my parents while I am in grad school, but their network was nowhere near fast enough or far-reaching enough for me in my bedroom, so I decided to give them an upgrade. Due to band steering in the new router(s) I wasn’t able to connect most of the smart home devices to the new network. This meant I needed to split off a dedicated IoT network, locked in at 2.4GHz, so that the devices could cope and alleviate the band steering issue. Luckily I know how to do that, and why things weren’t working. Most people probably would not know what was wrong, and would eventually get upset and return the device or perform percussive maintenance on it.

Passing a spec ≠\neq working

This next one is personal. Accessibility standards can be applied to documents through things like PDF/UA (an international standard for PDF accessibility). These standards define how a document needs to be structured and what an author must do to ensure the document is usable by assistive technology.

My current work at Clemson is on LaTeX\LaTeX documents, and how we can make them accessible. You can see the Clemson LaTeX\LaTeX Guide here to see the work I have done and give feedback on the guide. The clemson.sty package is built to bring LaTeX\LaTeX documents up to PDF/UA-2 compliance with as few changes to the ways you write your code as possible, but you do still have to do some minor tweaks. This is still not a universal solution though. The only true universal solution is the LaTeX\LaTeX source being accessible from the first download. Even with this package you still can’t guarantee accessibility and that drives me crazy!

A file can pass a validator and still not be accessible or give you a different experience entirely depending on the system you’re using, or even the validator. One validator may complain about tables, and another about math tags, and another about the figures. The document may still be conformant, but whether a blind reader can actually use it depends on code the author never sees, and shouldn’t have to worry about. The Clemson package is a means to an end. We’re actively working to get PRs and issues opened so that we can patch the holes in the LaTeX\LaTeX kernel and get the accessible ball rolling.

The point here is that the standard can tell you a file is correct, or a software that thinks it knows the standard could tell you, but it cannot make every screen reader or assistive device implement it correctly. That gap is the entire problem in accessibility today, because there is no work-around for the person on the other end of the pipeline.

So what’s the point in the “standards” then?

The pattern we see in all of these is the exact same:

  • Room for interpretation in the spec. Committees and organizations write broad standards, so every member company can sign on. Broad means vague. This is how we get two implementations that dramatically differ in output.
  • Vendor implementation is custom & optional. Anything that is optional is typically left just that in a final product. These companies can save time and money—and keep their ecosystem intact—if they don’t make a competitor’s device work with their products, and they don’t need to develop their own either.
  • Certified is just for paperwork, not for people. A device or file can pass every single required test that there is and still fail in the situations that actually matter. The tests only go against the spec, not against what a person may actually be trying to do.

Protocols can only be as universal as their worst implementation is, and the user doesn’t get a choice in the implementation that they get.

What actually works here?

There are some truly universal standards. Think TCP/IP connections, which connect every device from any manufacturer on the planet. HTTP serves this very webpage you’re viewing to whatever browser you have it pulled up in. Emails can go from one provider to another thanks to a protocol that you don’t even have to think about.

Each of these things has a small core that every implementation has to get right, with all the added extras built on top. Tons of interoperability testing went on, with different teams implementing different ways and sending/receiving from each other, and the protocols grew. The IETF’s motto is “rough consensus and running code,” and it works! These things are designed so that no company owns the experience, which removes the commercial incentive to break someone else’s implementation.

The lesson that I have learned as a UPIC intern in accessibility at Clemson is the thing my boss keeps written on the board in her office: “Compliance is the floor.” There is still more work to do after that! We can’t just check standards against the document defining it, we must check it against reality.

Until we figure that out though, I guess I will keep a drawer full of labeled USB-C cables.