|
|
|
|
|
by enobrev
1595 days ago
|
|
I was thinking of this generally. I've built hundreds of websites and while many had plenty of similarities, there would still be a plethora of databases and languages to support. And that's just my own history. There are quite a few developers, languages, data formats, services, etc. And so the system that can run any "boxed" website - as in a working archive - would need to support all of these things, as they are today, in some form or would need to emulate / translate these things. But instead of normalizing and archiving, my off-the-cuff suggestion is basically to freeze it in time. OS, languages, databases, services, data, code, everything is exactly as it was last built. Ideally, provided the storage medium held all its bits and the power source were compatible, it should boot up in 200 years without issue. In a sense this is what we've accomplished with VMs and containers. But those still require an underlying system. If the system is part of the medium, you only need power and a console. |
|
I don't think you absorbed the point I was going for there.
It sounds like you're prioritizing the website as an entity unto itself above the content. You're not thinking of the content as individual entities, and the website merely being a place where these entities are collected—like books on a bookshelf. That first false step in the conceptualization of this stuff is I believe to be _the_ reason that websites are "hard" to make manageable for a non-technical audience. All the implementations of "Information Management: A Proposal" have been executed poorly because the thing that TBL was outlining was obviously going to require computer systems, and when computer systems are involved you get the kinds of people who work on computer systems and their (largely self-serving) tendencies—which a lot of the time involves making everything in that system another system.
Do you need to support a plethora of databases and languages to maintain the other books on your bookshelf—that is, the dead-tree ones? Even for digital artifacts... the music you listen to, do you encode those as programs with unbounded power that you execute every time you want to hear a song? Or do you restrict yourself to relying on "dead" formats for what is sufficiently "dead" content, and then offload the need for computation to a separate machine that operates across a well-defined interface and that exists outside the content itself—a music player app? When it comes to backups, do you exclusively use self-extracting archives instead of something reliable like a tarball or a ZIP? Why not?
At the end of the day, websites are really just about pieces of content and the names we assign to them. It's the entanglement of all the unnecessary violations of the Principle of Least Power in the form of live systems and running programs that has made this stuff hard to manage (often not just for non-technical people, but programmers and sysadmins, too).