Here’s that CGI application I’ve been working on; a work-in-progress, basic HTML frontend to the Wayback Machine, for browsing the old web straight from an old browser, including the Wayback Machine’s site search 💻🌐📅

It’s not especially efficient right now and doesn’t cache any of the data it fetches from the Wayback Machine, so *please* be gentle, but it’s working great for some retro browsing!

and yes every page *really is* completely valid HTML 4.01

(okay, *almost* every page, just found one that isn’t 😅)

I have now fixed compatibility with NCSA Mosaic, just in case you were itching for that

I updated this with what feels like a novel way to account for Japanese 🌀Dreamcast browsers interpreting everything as Shift JIS; I’m sending a hidden field in the form with a Unicode value and seeing what it encodes it as, then outputting Shift JIS values where needed

~technically~ this means in those cases the UTF-8 header is a lie, but the browser is straight up ignoring it anyway so nbd, instead it means you can actually read the descriptions on Japanese websites!

someone suggested supporting web archive aggregators; I‘d love to, but this only runs usably fast because the CDX API can be filtered so effectively; the aggregators use the Memento protocol which you can’t filter, so the list of snapshots often takes minutes or times out

I’m very aggressively avoiding asking for too much data at a time; filtering down to a single snapshot *per month* to build the main index, and then drilling down to request only a single month of snapshots to optimise for that higher precision

if Memento were extended to better support these sorts of operations, namely an index of dates, and a specific time period’s snapshots, I’d move to an aggregator in a heartbeat, but it’s unclear who owns Memento or how to give feedback, so I post it into the void instead 😂

now that this respectfully caches data from the Wayback Machine, I’ve given it its own domain; say hello to

newly added to site maps! 🗺📍

follow a URL with an asterisk to get an index of URLs the Wayback Machine has which match, then filter further via URL and MIME type matches, no JS required

it’s also had a bit of a lick of paint (serif fonts [where available], and a background tweaked for improved contrast), and gained a help page

the site map feature was something I could only really add with caching - I’d feel really bad to be hammering the Wayback Machine API with requests just for filtering it down further server-side, so I’m glad it’s all in place now! 😃

(that should be sans-serif, sorry, I’m tired)

@ticky no no it's really cool! i'm not trying to diminish yours!

@ticky This is all so amazing! Can I make a UX suggestion?

@ticky On lower res devices, like 640x480 and under, the search bar is a big scroll down past the advanced search description - wonder if it might be possible to have search above the advanced stuff!

But I realise that's more edge case.

@LaCrecerelle well, when you’re on the search help page I figure you’re looking for that help first and foremost, so I put it at the end so you can use it after reading. The home page, though, has search front and centre!

Sign in to participate in the conversation

cybrespace: the social hub of the information superhighway jack in to the mastodon fediverse today and surf the dataflow through our cybrepunk, slightly glitchy web portal support us on patreon or liberapay!