| Location | USA (Austin, TX) |
Not sure how it is that substantial portions of my Zotero library disappeared, but I'm not thrilled about it, and I'm not looking forward to importing from backup and then manually deduping 1000+ items.
Total items as of 2024-06-12: 1496
Total items as of 2025-07-05: 1424
A publishing setup that doesn't let you keep existing references online isn't doing its job.
Suppose your family ran a corner shop and you needed someone to ring up items at the register, take payment, etc. Or maybe it's a pizza parlor and you need someone to work with you to make the pizzas. Would you hire someone to do these things and keep employing them if they showed up and refused to take payments at the register or make pizza?
It's way too common for complaints about unresolvable references to get brushed off because e.g. the page has been "edited".
nit: it hasn't—it's been replaced with a *different* page that, granted, may be substantially similar to an earlier one, but is nonetheless distinct—and therefore should have its own distinct identity (read: URL).
Or the tech stack (CMS/blogging setup) has changed.
Or someone decided to reorganize the folder structure in some uncool attempt to tidy up…
This is a problem that has been, well, *a problem* since before recent politically motivated changes.
The Web is almost 35 years old. It's time to stop behaving like volatility should be the norm. It certainly doesn't *need* to be this way, despite whatever strange notions people get into their heads (and how they tend to adopt a totally unrelated posture about a subject) when computers are suddenly involved.
I offer an observation for all those expressing concerns about the disappearing of pages, snippets, blurbs, &c from gov websites:
In what is broadly the field of "information", which includes everyone from librarians to programmers to authors and editors, professionals have failed to inculcate (i.e. among colleagues—and promulgate to the public) adequate appreciation/respect for the long-term availability of information resources (e.g. website "content") and integrity of references to it.
If you're really courageous, you might even be interested in "Sun Microsystems":
You can talk about a certain software package's performance—
"The interpreter distributed with software Z is very efficient."
"The code generator from the Y toolchain outputs binaries that execute speedily (on the subset of hardware that we're interested in)"
But these are properties *of _THAT_ software package*, not a property of the notation that a given program is expressed with.
Programming languages are notation.
A programming language cannot be "fast" or "slow".*
It's a LANGUAGE. It's how you communicate by laying down, in writing, the way that a program treats and defines its inputs and outputs.
* wrt how a program encoded in that language runs, that is—it is certainly the case that some forms of notation are faster/slower to write than others
Simple litmus test for Web accessibility: if I can't get the Wayback Machine to take a snapshot of it and then redisplay it correctly, that's too inaccessible.
This is mostly meant to be the rubric that overly dynamic, JS-heavy websites should be graded against, but it also applies to stuff that's behind a form submission for no good reason, too.