4 ms·
I posted this as I think it’s fun to look back at how some “early” web stuff was done and a bit of a trip down memory lane on a Saturday morning. When I was ex
by samwillis 5y ago
I posted this as I think it’s fun to look back at how some “early” web stuff was done and a bit of a trip down memory lane on a Saturday morning.
When I was experimenting making websites as a teenager 20 (ouch!) years ago I would love to investigate how other sites worked. The BBC had this little “Text Only” link on all pages of the site that redirected to a version stripped of all images and tables (tables were used for layout, much like HN) with high contrast and large text for the visually impaired and screen readers. Brilliantly ahead of its time thinking about accessibility, which I suppose is the BBCs mandate.
The url to these pages were prefixed with something like /cgi-bin/betsie.pl?url= and the bottom of the pages liked to this page I have submitted (sadly now only in the Internet Archive), it’s affectively a proxy written in Perl that transforms each page on the fly. I remember setting Betsie up on my own site, which no one visited because who would visit the site of a 14yo.
I believe Betsy was still in use till about 2010. You can read more about it here:
https://www.bbc.co.uk/blogs/bbcinternet/2012/02/accessibility_betsie_text.html https://www.bbc.co.uk/blogs/bbcinternet/2012/02/accessibilit...
Source code:
https://web.archive.org/web/20000824120750/http://www.bbc.co.uk/education/betsie/parser.pl.txt https://web.archive.org/web/20000824120750/http://www.bbc.co...
- unfocussed_mike 5y agoGreat flashback. I used Betsie on several projects in the early 2000s. It was quite useful when you were stuck with baroque HTML generation code you were unable to modify. The BBC produce a bunch of interesting tools for web development, another being the layout comparison tool, Wraith: https://bbc.github.io/wraith/ https://bbc.github.io/wraith/
- mmaunder 5y agoI worked for the BBC in 2002. I created OneMusic which showcased unsigned bands. You can see an archived version here: https://web.archive.org/web/20050416110703/http://www.bbc.co.uk/radio1/onemusic?OMSID=50964355468749110565185546874 https://web.archive.org/web/20050416110703/http://www.bbc.co... Within 2 weeks we managed to get our first band signed. A Swedish band called Maas. I worked with a producer Andrew Barron, who takes a lot of credit for seeing the project past the finish line. Andrew sat with me over a weekend and deep into a Sunday night to get it finally launched. The project was written in mod_perl and was designed to be extrememly high performance because Radio 1 at the time (and perhaps still is) the most popular radio station in the UK. And we expected a surge of traffic every time the DJ mentioned OneMusic, which is exactly what we got. I was based in Oxford Circus and I worked with the ops team which - at the time, and I'm not sure if this is still the case - was based out in some mansion in the country somewhere. They had a kind of Bletchley Park vibe about them - this team of absolute gurus who you only got to speak with over the phone, and who would somehow make your application handle the massive amounts of traffic that would be thrown at it. They did a great job hosting that application on Apache/mod_perl in the dark days before Nginx was invented. One of the absolute worst design decisions I made was to design the application so that it would support sessions for users without cookies enabled. Which is why that Archive.org link has the OMSID query string parameter in it. The app would dynamically modify all URLs on the page to embed that assigned session ID so it stuck, even if you didn't have cookies enabled. It would follow you around the site as a query string param. I stupidly suggested the idea to Andrew and he was like "Hell yes, lets do that!" and so I did. But it had big performance implications as the URLs had to be dynamically generated for every pageview, and it added significantly to the time it to for me to create the project. But hey, we did it, and got it launched, and it actually worked and provided sessions for users who had cookies disabled. The project taught me a lot about high performance websites, delivering a critical project on-time to a big audience, and I went on to create WorkZoo (Time Mag top 50 website of 2005) which I sold, Feedjit (live traffic feed on over a million sites) and my current cybersecurity business Wordfence which employs 40 and is the best available security product for WordPress. I think creating all those projects set me up to create a WordPress plugin that is high performance, widely compatible, very stable, and delivers on user satisfaction. When I left the BBC, the team bought me a bottle of Gran Centenario Reposado tequila out of their own money. I'll never forget that and I'll never forget them. Such a great group of people.
- cnorthwood 5y agoI'm guessing the mansion was Kingswood Warren, where BBC Research was based before cuts relocated them to London: https://www.bbc.co.uk/rd/ https://www.bbc.co.uk/rd/
- mmaunder 5y agoSounds right. https://www.bbc.com/historyofthebbc/buildings/kingswood-warren/ https://www.bbc.com/historyofthebbc/buildings/kingswood-warr...
- unfocussed_mike 5y ago> One of the absolute worst design decisions I made was to design the application so that it would support sessions for users without cookies enabled. I think we all did this at least once back then and all regret it for the same reasons...
- hackernewds 5y agowhy don't we do it anymore
- samwillis 5y agoIf you include the session ID in the url it can leak to other uses if they send a urls to someone, it also exposes the session id in intermediate proxy/router log files (if using plain http, not tls). So it’s basically a security risk. Before cookies had widespread support it was basically the only way to implement sessions.