Or how to browse the modern internet on your retro computer.
If you love prehistoric computers and handhelds like Psion, Palm Pilot etc, and just like me you like to connect these to the internet, you’ve probably found out that the old browsers that are available for these devices don’t work very well with modern internet. First of all, they more or less all lack the ability to connect to secure (https) websites, which is a requirement for most sites nowadays. Also, they don’t render modern scripting and modern HTML very well. The result is that pages don’t load, or that if they load, you get a bunch of garbage. To fix this, there are some options available. By far the most known one is frogfind.com I think. There are also some options where you host a proxy in a docker container or on another (modern) computer. What all of these solutions have in common is that they serve as a translation layer between the modern internet and your very old computer. They strip all modern stuff from the websites, and ditch the https requirement.
For a website like frogfind, no additional setup is needed. Just open frogfind, enter your search or url, and it will serve you a stripped down version of the website. The self hosted proxy servers are a bit more complex, and require you to set a proxy server in your old browser.
While frogfind is much more accessible, for me it lacks some important features. Frogfind doesn’t show images. In stead, it shows a list of links on the page to all images that are found. That means links for every button-image, logo and the content images, but you never know which one you get before you click.
I wanted to make my take on this and make a version that shows either the images as links, but then inline, so on the actual place where they should be, or the images themselves, also inline, as a scaled down version. All this, while trying to filter irrelevant images at the same time.
Now I can code fine, but AI can do it better nowadays, so I took it on me to vibecode this project. The result can be found here:
Features:
- Compatible with HTML 2.0 browsers
- Content extraction should be better in general. This means that the actual content of the site is shown and not all noise that is surrounding it.
- Images are rendered as inline links OR images (select in top bar)
- Images are resized and converted to small gif files for compatibility and to use only little bandwidth.
- The search box also accepts URLs and opens them directly.
If you open a website like https://nos.nl through both RetroWEB and FrogFind you should get a way better result with the actual first headlines in place when using RetroWEB.
It’s not perfect, but I think this is as good as it gets since modern websites are extremely complex and it is very hard to interpret and re-render them correctly.
Currently I’m leaving some HTML 3.2 stuff in the proxied pages, hoping HTML 2.0 browsers will simply ignore this. If this gives any issue, please let me know.
I’m curious to hear your experience
Leave a Reply