Every decision is a trade-off. Proprietary software with support from a vendor can greatly speed up your product viability and market profitability.
For most companies, and especially startups, that is far more important than the very unlikely risk that a vendor completely disappears overnight, and the even more unlikely risk that their software also stops working completely before you can migrate to something else.
The vendor disappearing overnight is only one of many risks from using proprietary software. Others include the vendor discontinuing the product, taking the product in a radically different direction, the company being acquired, or simply changing the licensing model (see Adobe Creative Cloud) to dramatically increase the costs of using the software.
Look, right now I am working for a company that is in the midst of attempting to transition out of a very old source code and release management system and they're having a hell of a time doing it. That system happens to be proprietary and the support plus licensing fees are astronomical while the actual tech support is abysmal.
Yes, the risk may seem like an easy tradeoff when you're starting out and you need to ship and you don't have any market share to worry about. It's a whole different story when you're dealing with a very clunky, yet very profitable legacy system that you're not allowed to fix because it's proprietary and yet your business depends on it.
> The vendor disappearing overnight is only one of many risks from using proprietary software. Others include...
> the vendor discontinuing the product
This is not a zero risk proposition with open source software either; your costs go up significantly if you have to start maintaining a legacy codebase.
> taking the product in a radically different direction
See above; if you just want to run the old version, proprietary software lets you do this as well.
> the company being acquired
This is definitely the biggest risk with Datomic; if Cognitect decides to EOL Datomic, there is a very high chance that they open source it (see the various free software they develop already), but if they are acquired by Oracle that chance becomes zero.
> or simply changing the licensing model (see Adobe Creative Cloud) to dramatically increase the costs of using the software.
Datomic licenses are perpetual I believe, so not a risk with Datomic.
> Datomic licenses are perpetual I believe, so not a risk with Datomic.
On-prem is perpetual with a year of maintenance (which can be extended), while Datomic Cloud is integrated with AWS and charged and licensed like other AWS services: month-by-month.
You can't (at least not with some great effort), cloud depends completely on various aws services, that's the worst part of datomic IMO, cloud and on-prem are two different incompatible databases.
Sure, but that's my point: weigh the actual risks.
How big is the company you're working for? Could they have gotten that big in the first place without using these tools? Companies change as they scale and solutions that worked when they were young will almost always need to change as they grow, so I don't see that as a particularly bad situation. It's a cycle of constant change management and risk mitigation.
Usually the bigger company has the resources to make changes while a startup trying to plan for 100x future size usually ends up limiting its own growth.
For most companies, and especially startups, that is far more important than the very unlikely risk that a vendor completely disappears overnight, and the even more unlikely risk that their software also stops working completely before you can migrate to something else.