The Web is No Longer Open and It's Getting a Lot Worse Again
THE other day we saw this old article from 1996. It was written by Jahn Rentmeister and it covered many issues that we still see today. It's still relevant to the World Wide Web and now it'll be in Gemini too, cautioning people against optimising for particular clients/browsers. There are horrendous long-term ramifications.
Over the past few days we've tested our stylesheets for compatibility and graceful fallbacks with Web browsers as rudimentary and rare as Netsurf. Our goal is to ensure the site remains fully accessible even to very lightweight or old browsers. The "old" site worked OK with browsers as old as 20 years. That was by intention.
We're not sure if people noticed, but the wiki theme we've used since 2009 basically got "broken" by newer versions of Webkit and whatever Chromium's rendering engine calls itself now (there's some fragmentation, historically). Basically, the navigation menu fell below page content and there was no easy fix. Credit to Mozilla: in Gecko it never broke! Newer versions of browsers did the same to schestowitz.com's WordPress blog (the navigation menu fell below page content around 2007 even in Opera, well before Chrome existed). Over time it became a problem in almost every browser, but not in Gecko-based browsers. So the lesson of it all is, things were OK until or except when some new rendering engines came along with a different "mood" or "interpretation". Back in 2008 many still used Gecko (Firefox was very widespread) and IE6/7 worked OK with the existing pages. However, in "modern" browsers, there's risk of sites breaking due to a new version of the browser or new browsers becoming available or new releases of a rendering engine. A new release may break things at any time; they won't fix their "bugs", you need to fix your now-"buggy" site to facilitate whatever change they spew out... until the next version comes out (then things may break again).
This sort of mindset has led to considerable waste, such as Web developers having to test many versions of many different browsers. It never ends. Insisting that you use a particular version of a particular client is unfair and messy. It invalidates the premise that Web means "cross-platform" and "no need to install anything"....
The best case scenario: works with WWW (in general, nothing specific to any Web browser)
OK scenario: works with all major browsers
A bad scenario: works best with [browser]
The worst case scenario: works best with [browser] [version] (not tested with anything else)
"That is a return to the browser wars of 1996," a reader noted, "apparently on purpose, with all the baggage included."
We're going to write more about this subject as we continue to migrate old site pages; it's a good decade to talk about the open Web because we've quickly lost what's left of it in recent years. We must speak about the issue repeatedly, demanding standards and freedom because we are losing both; "EME, WEI, WASM etc are sliding into place and sometime soon the switch will be flipped; the question is how far down into the system it will reach and whether (how soon) it will connect through to TPM at the other end," a reader noted. Then there are centralised CAs, outsourced (e.g. to Google) fonts, lots of hotlinked "content" (even fonts) and so on and so forth.
"Regarding Jahn's article," the reader added, "Java in the Web page has been displaced by JavaScript and the shortcomings with both, especially the former, can be said about WASM as well." █