glass formation
Making a glass is a race. On one side is the melt, which given enough time always wants to crystallize into its lowest-energy ordered form. On the other side is the clock: crystals need atoms to migrate and assemble, and that takes time. If you cool the melt fast enough that the atoms freeze into place before crystals can nucleate and grow, you win the race and trap the liquid structure as a glass. Cool too slowly and the crystals win. Glass formation is simply beating crystallization to the finish line.
How fast is fast enough depends on the material. The competition is often drawn as a time-temperature-transformation curve, a nose-shaped boundary marking where crystals first appear; you must steer your cooling path to the left of the nose to miss it. Silica is a lazy crystallizer because its network is so viscous that atoms can barely move, so it needs a critical cooling rate of only a fraction of a degree per second and forms glass almost effortlessly. A pure metal is the opposite: its atoms are nimble and it crystallizes so eagerly that you need to quench it at roughly a million degrees per second to make a metallic glass. Ordinary window glass sits comfortably in between.
The honest headline is that almost any liquid can be made into a glass if you can cool it fast enough, and conversely even a great glass-former will crystallize if held too long in the danger zone. Glass formation is therefore a kinetic accident we have learned to engineer, not a special property that only a few magic materials possess. The oxides that form glass easily, the good network-formers, are the ones whose melts are already stiff and slow.
Obsidian is nature's glass: rhyolitic lava that oozed onto cool air and chilled fast enough to skip crystallization, leaving black volcanic glass sharp enough for a scalpel. The very same magma cooling slowly deep underground instead grows into coarse crystalline granite. One melt, two rocks, decided entirely by cooling rate.
Obsidian versus granite from the same melt shows glass formation is a race between quenching and crystallization.
Good glass-forming ability is not magic; it just means the melt is viscous enough near its freezing range that atoms cannot rearrange in time. Add a network-modifier and you make the melt runnier, which actually raises the risk of devitrification.