Services Are Not Forever, Files Might Be
The existence and long-term availability of any single online service is predicated on a simple balance: are you -- the customer -- making the business more money than you are costing it?
If we are dealing with a paid service, this customer-business relationship is more direct, and it's easier to figure out how much money each customer makes the business, and how fixed business costs are amortized across the userbase.
On a “free-to-the-user” service, on the other hand, this computation is not as straightforward. The free offering might be a “loss-leader”, hopefully attracting customers to interact with other parts of the business with higher profit margins. Or the customer might not be the user, but instead the parties that buy the user's behavioral data or user's attention through ads.
Either way, services only exist as long as it makes financial sense for them to continue to be operated. If you are aiming for something to last and be available and useful for decades, don't bet it on services. Services are temporarily useful tools.
On the other side of the coin, there are files, an ancient (in computer timelines) paradigm the consumer-facing technology industry seems to be moving away from.
Files are finite sequences of bytes, which can be thought of as numbers from 0 to 255. When a file is on non-volatile storage (a hard-drive, flash memory, SSD), you own it. Nobody can take it away from you. It will be available as long as the physical media is able to serve back the bytes that were written into it.
The caveat being that, strictly speaking, the file is just a part of it, you also need a program to make sense of the data and to convert that sequence of numbers into a intelligible experience, such as a text document, photograph or video.
The good news is that there are file formats which are open standards. They have existed for decades and the information of how to decode the bytes into information is well-specified and publicly known. Here are some examples:
- UTF-8-encoded plain-text (TXT) files.
- For images: JPEG, PNG.
- For text: Markdown, HTML, PDF (previously proprietary).
- For audio: MP3 (the patent expired), OGG.
- For tabular data: CSV.
- For compression: 7-Zip.
- For variant data: JSON.
There are plenty of proprietary formats that can be used instead, but their advantage is marginal and, in my opinion, not worth it if the goal is long-term archival and longevity.
Own your data. Do not surrender it to the faceless services and proprietary formats for convenience.