[flagged]
I do disagree about the web being poor for content though. Properly done it is way better than PDFs. Provided it's simple and clean.
It's definitely not better for content. It's impossible to reference easily, everything is transient and ephemeral, layouts differ all over the place, there is no concept of location other than anchors which are poorly defined, typography is "variable" and goes somewhere from hooky to crack smoking, still to this day half of the images are rasters, it's near impossible to save bits of it locally if you want to keep it, it's difficult to print and importantly the moment you add something dynamic to it you multiply the damage by 10x for all of the above.
I hate modern sites where you need JavaScript to load text and images.
My issue with pdf is it's a shit show behind the scenes.
And somehow UWP apps manage to be even more sluggish and resource-hogging than webapps despite being "native". Native Android SDK apps aren't faster than webapps either if you exactly match the layout (tried it myself in one case and was shocked by how slow native Android layouting really is).
I think the way forward is to improve and standardize local webapps. But not sure how it could work out when every (established) platform has incentives against it...
I'm not a fan of PDFs for content because they're very "static" in that they're really meant to be printed than read on screen.
If i had to pick a favorite format/container that'd be CHM - which is essentially a self-contained offline "web site" with table of contents, search, index and only whatever web tech was available in IE in 1998 or so (which is still more than enough to hang yourself with, but due to the focus on actually being all about a text format, 99.9% of CHM files restrained themselves to basic HTML), which allows for screen-based reading and dynamic layouts (mainly resizing the font, window, etc) and pixel-perfect images.
Though note that i'm using CHM as a "category", e.g. i'd also put Qt's help system, JavaHelp and even GNU info (not to be confused with Texinfo) in the same category (though info is more annoying with its hardwrapped lines, monospace text, use of titles as page names that must be unique and hacked-in support for images that most viewers do not even support).
I've got no idea how bad or good the format is, but it seems to work quite well for what I've used it, better than PDF in allowing reflow/dynamic oage sizes, but still being a kind of self-contained static web site with search, ToC and index.
Projects like Gemini [1] try to tackle this by reinventing a simpler, document/content focused protocol, but let's be honest it has no chance of really succeeding.
However, I really don't think PDF is the right way. PDF is a horrible format, only good for what it was designed to do: printing to paper. On screen, PDF is incapable of producing free-flowing resizable text for example (and I'll add that nowadays, even for printing PDF is not necessarily the best choice, you can do very advanced clean layout using web tech with Paged.js [2] for example). Plain old simple web formatting like is used in the ePub format is generally more than enough for what would be a PDF.
Bar the GitHub markdown and inevitable confluence in tech, I don’t see this anywhere else. It’s Word, people writing fixed size pages on their iPads and consuming vast amounts of physical documents. In fact when it comes to technical publishing, dynamic layouts almost always fall to pieces.
PDF, or at least the less Adobe subset that falls out of typesetting software, while not ideal is the least bad interchange format we have for this.
The only close thing I see is ebooks but I hear regular complaints appear from people who lose the concept of “page” when using those.
Gemini I am interested in and will research further.
have you tried Gemini?