Ask HN: Has anyone tried turning Web Content into static documents, like PDFs? | Hacker News Reader

Ask HN: Has anyone tried turning Web Content into static documents, like PDFs?

6 pointskeepamovin7 comments

I'm aware of the good projects SingleFile, SingleFileZ, etc to archive a webpage into a file, but I'm thinking more like an actual dedicated, usable, sharable, hackable document format for sending Web Content around, but not over HTTP? I know MHTML, but I'm thinking something that more maintains visual fidelity and usability (up to what has been archives). Seems using Web Content an an authoring medium and sending these documents around could be a nice alternatives to PDFs but in a similar market? Where HTTP serving such Web Content, simply becomes another peer distribution medium for content that exists and is separate to the medium of exchange. These docs could be sent over email, messaging, all kinds of ways. Seems it would possibly offer a richer, more interactive, and potentially easier-to-hack/author alternative to PDFs, and other prop formats. I tried looking for this but didn't find anything. Is this idea fundamentally broken in ways that aren't obvious?

On https://mitta.us, I snapshot the whole page into a PNG using a headless browser. That PNG can then be sent into a pipeline to have the text extracted by Google's Document AI. That text is sent off to OpenAI to embed it, and stored in Solr, where it can be used later to search for it. The main issue is that I can't pick up links with it, but most of the time that really doesn't matter as long as I can see the page like it was rendered and have the text searchable.

There's Print to PDF printer choice - so far Chrome seems to do the best job of it - if the HTML/CSS is not too responsive it can be not good.

I don't get it. What's more portable and hackable than HTML? Why not just make a website and zip it up?

That’s fair enough it is a bit hard to articulate, I suppose.

I think there’s something about the medium of a document, a thing with a reader/viewer and an authoring tool, that you can send to people and they can send you ( fully specified rather than just a URL) that is distinct from what the web is.

It may seem so obvious that it is almost invisible, but there’s something very powerful about organizing a form of content of a medium into a document format. Shareable, distributable.

It may be an imprecise analogy or example from a non-expert but humans had writing for a long time before we really created documents like account books, printed books, legal contracts, all these genres of documents, somehow lifts whatever it is that writing is, into a new form and power by virtue of them being encapsulated in a document. In a similar way that the creation of video movie and song documents (cassette tapes records, VHS, film emulsion, etc.) lifted visual and audio information into a new, medium and power, because these documents became first class currencies of exchange of that.

I just think the same as possible to lift web writing, so Web technologies into a document format as well.

A small, but non-trivial part of this is that zipping a website and sending it it’s not really a hackable document in the same was DOCX is, and it won’t display in the same way as a regular website if you serve it from files.

I don't quite understand. Can't anyone edit the HTML once they unzip it? How is it different if you zip it up?

Is it the WYSIWYG you want? Why not just send an actual document then, like Word or LibreOffice?

Maybe I'm misunderstanding what you mean by "hackable", but I'd have thought PDFs were the least hackable. They're notoriously hard to edit.

I guess Hackable is just one part of it and maybe it’s not the best word. There’s plenty of other things that make documents important.

While it’s true PDFs are hard to edit, I mean anyone can author a PDF…and probably in some circles far more people know how to create a PDF and they do create a website.

It’s not really the W7 it’s more the richness and ubiquity of web content, has so far eluded encapsulation in a document format.

I’m sorry the idea so out there! Haha :)

I’ve used athenapdf before for related activities w good results. They’re on github

Reply on news.ycombinator.com