On Mon, Nov 04, 2024 at 06:00:03AM +0000, Andy Smith wrote: > Hi, > > On Sun, Nov 03, 2024 at 07:43:44AM -0600, Richard Owlett wrote: > > I'm working on a weird personal proof-of-concept project. > > A HTML 5 compatible browser will *NOT* be considered. > > I can see why you might THINK this is a good requirement: you're > probably thinking that by only wanting minimal features you could use > simple, robust software. However, I think it's a bad requirement that is > severely restricting your choices here. >
This - I wish Martin Wheeler were still around to give me some authoritative steer. HTML was, essentially, a subset of SGML at one point. if you write your HTML with the appropriate Document Type Definition, you limit yourself to the features supported by that DTD. You can *write* HTML 2.0 or 3.2 or 4.01 and the parser parsing it will spit out any features that are not supported. Amaya was intended as an editor and a parser by the W3C, for example. > I think it is likely that any HTML browser would be expected to render > modern HTML, and any browser that doesn't would have a very tiny user > base. > HTML itself, as distinct from CSS, is largely backwards compatible. The Wayback Machine read my website in the 1990s and it's still readable and printable in a modern browser. > Installing 20+ year old software just to make sure it can never parse > HTML5 is total lunacy. There will be no support community for such a > thing for a start, so any problem you have is going to be a showstopper. > Richard - your quest makes me think of the following analogy: "I would like to remind myself how driving was in the 1960s - so I'll build a replica 1960s car" - no, go and find an enthusiast to let you drive a Ford Anglia or whatever for an hour and remember both how good and how bad it was. Let them have the problems :) > If you want simple robust output and behaviour I think it's probably > better to get that by making sure the CONTENT is very simple HTML. Then > you get to choose from every web browser that exists today. > All absolutely true - Andy Smith is correct here. > > Any suggestions? > > Consider writing content in Markdown instead and using pandoc to turn > that into HTML and/or PDF. > Good suggestion, Andy - but that would be *another* rabbit hole for Richard to go down and for the list readers to trouble shoot and problem solve in due course. > Thanks, > Andy > All the very best to all, as ever, Andrew Cater (amaca...@debian.org) >
