Looks like they've reinvented identd, with all the same features and flaws.

SMF instead of systemd?
ZFS instead of BTRFS, meaning you can boot from ZFS a mirrored pool along with the other features?
A platform that is outside the Red Hat/Ubuntu corporate circles?
Multiarch support (i86 and x86_64) that just kinda works instead of the current Linux solutions?

How about you try this as "what could this become?" rather than "why does this look like everything else I've tried?"?
How about looking at what Illumos and friends could be ported to?
Your mileage, and consequently exposure to new ideas, may vary

Have them select an episode of Mythbusters and then analyze their scientific method. Have them state the hypothesis then compare their approach to that taken on the show. Have the student analyze the results, talk about variables and risk management. Then have them discuss some next steps for the experiment, and discuss practical applications of the findings. Mythbusters is awesome laying out a question and applying the scientific method, but due to time, they don't always pursue a question far as possible. I think they provide a good stepping stone to scientific research, and encourage critical thinking.

Although an incredibly dry beginning, reaching into information theory may be the most rewarding path. A neat thing about information theory is that it isn't all discrete sets and number crunching. You could start with an oscilloscope and just show what data looks like moving around. Such as a serial port--very easy to visualize and a stepping stone to any fancier stream. Also with a scope you can show AM and FM with different phase/amplitude modulations. Which leads right up to a ham radio license. I suspect she'll get just as caught up in error correcting/detecting codes, such as Hamming codes and Reed-Solomon codes, where some redundancy can make all the difference in receiving data correctly. And how much noise can a transmission tolerate? (Clue: CRC) Ethernet 10-base-T might be a nice case study of all of the above, and how carrier sense and collision detection was necessary. On top of error correcting codes you can introduce her to compression, covering both dictionary (LZA) and frequency (Huffman) approaches. Anyone that has dial-up (v.42bis and friends) is employing all of the above. After compression, introduce cryptography--being in some ways the opposite of error correction and compression. And there are a wealth of examples for cryptography and information theory, both full mechanical and computer programs! I think once you know where to look, this is a great opportunity!

Running Man by Richard Bachman (Steven King). Not like the movie at all--no mass market appeal. But the development of the main character and ending left me dazed for days.

New laws aren't really necessary, as the existing ones can be used to force additional improvements in this area. The manufacturers of the tubes are already held to FDA 820 for design and development, 21 CFR 820, of which section 820.30 (g) calls out risk analysis during validation. All the FDA has to do is indicate that the risk analysis includes foreseeable misuse which includes incorrect tubing connections, and all manufacturers will have to be able to demonstrate that their device is incompatible with a (justifiable) subset of existing similar tubes.

The FDA is already aware of the adverse events (21 CFR 820.198), and may even see trends. Some of these cases are the result of overstressed individuals, for which better staffing is the best answer. But for some of these cases I'll borrow from Forrest Gump, "How do you fix stupid?"

Moreover, would you want the government backing a specific industry standard for tubing? I can foresee many ways in which that system could be abused. The FDA is hoping someone else steps up and makes it easy for the different manufacturers to cooperate on this.

I use OpenSolaris on two systems at home, my desktop and a server (print and squid).

OpenSolaris provides a choice of userland (utilities) by setting the order of your path. So for those concerned about "incompatibilities" or "taint", you can choose your own userland, associated trappings, and level of standards conformance.

OS upgrades, or other significant package installs, are a no risk endeavor since the zfs root clones mean that you can boot to or rollback to a previous (working) instance if things go wrong. Snapshots from package installs and time-slider also provide a sense of security that you can "get back to that thing you just lost".

The package system deliberately has a concept of OS release, and ties the packages you install to that version. This reduces some of the confusion and problems that come from choosing packages implicitly based on a URL. Moreover, it also has means for a third-party vendor to host their own package repo and integrate it into the overall package scheme with user accounts, as appropriate.

The version you can download has all of the same features as that used in server rooms; so you can see how Sun met their customer requests for enterprise features compared to other distributions. Not just iSCSI and zones, but also MPIO among others are there for review.

The introduction of features to Open/Solaris feels more mature and predictable. Linux leads the bleeding edge with ideas and issues, whereas Solaris is a few steps behind, with solutions that feel more thought out.

If you'd like to meet OpenSolaris half-way, try Nexenta: http://nexenta.org/ which is the OpenSolaris kernel with a GNU (Debian) userland.