With all due respect, rendering PDFs is something I consider core browser functionality.

Pocket, most certainly not, though it's nice I suppose. Hello was... I don't know what it was. Checking if the market is there at all? Not something I expected to be in a browser.

I never understood why anyone wanted their browser to render pdfs. PDF.js is slow, buggy, and can’t edit documents. Every major OS ships with a better PDF reader, and there are still better open source PDF readers available on all major platforms.

> PDF.js is slow, buggy, and can’t edit documents.

That is leagues ahead of native PDF plugins which are slow, buggy, full of vulnerabilities, closed source, require a deprecated plugin infrastructure and usually can't edit documents either.

PDF.js is one of the better things to come out of Mozilla.

Okay. But the solution is to not involve the browser at all. It's there to browse. Once you find and download the file have it open in a real application designed for reading pdf.

Ideally, yeah. But often PDFs are integrated with the web experience as another web page. Or atleast that's how I treat them. I have 2-5 research papers open in my tabs at all times, and I browse through them like I would with any other article.

Sure if it piques my interest I would download it, organize it but I wouldn't go through the pains of doing that for every pdf I lay my eyes upon.

>download it, organize it but I wouldn't go through the pains of doing that for every pdf I lay my eyes upon.

You don't have to do that. Just set your browser to open the file in a real application. It'll automagically download to some temp dir (ie, /tmp) and you won't have to care about file paths or organization at all (unless you want to).

I have tens of research papers open in my tabs at all times. But for a good experience reading I open the full text in my pdf reader.

PDF.js has had a huge amount of its own vulnerabilities.

I don't think you've spent any time with regular people. The vast majority of the world finds computers to be confusing. The less buttons they have to click the better. All they want really is another appliance that has fixed functionality.

You're being needlessly down-voted here (I can only guess at what silly HN tripwire you've activated) but you're absolutely right. Tech people like us might think it's completely obvious that a browser shouldn't render pdfs or make your toast. But end users don't care. Browser x lets them read stuff quicker without switching to another app. They want the monolith.

We can dislike it, but we need to reconcile it.

Yeah, and I know this first hand. I work at a biotech startup and we have some super intelligent people who regularly get confused with modern UIs. For e.g. Not everyone understands that the hamburger menu is actually a menu. Or that flat shaded text can actually be a UI button element.

This gets to the heart of the matter: what is the core functionality?

For some saving bookmarks isn't effective while a solution like Pocket fits them better. Hello was an experiment to see if making video communication more accessible would connect people. Consider the saying about how no one uses every part of Microsoft Word, yet everyone uses a different ~10%.

Anyway, I think it's great that Mozilla experiment and try pushing the web forward. If only they were more transparent and consistently made these things opt in.

pdf.js seems to work quite nicely.

Seems that's no different than opening with external program, or using a plugin, or using an extension. After all,those things can be updated, whereas the browser should have more core functionality that enable new document types to being opened.

pdf.js performance is horrible, and now without NPAPI it is impossible to view PDF in firefox tab.

PDF.js performs quite well for me and works fine in tabs. It has always been a better experience for me than PDFium, for example.

Building a secure, fast, usable web browser is a near-impossible task on its own. Anything else they do involves a couple of orders of magnitude less effort. Mozilla would still need the vast majority of funding they currently get to do that.

The problem isn't that they're doing too much - it's that they have nearly no business model, and any attempt to create a business model appears to be selling out to users.

No. It is actually pretty easy. Just fork chromium, strip out the spyware, and add your own layer of features on top.

There are a lot of companies doing this now and Mozilla would be far more successful if they followed this model instead of trying to copy chrome's user experience on top of their inferior browser engine.

Yet we also need variety in our engines. Otherwise it's IE6 all over again. Diversity makes us stronger.

Then we get one company deciding which features should and should not be available to a web page.

ESPECIALLY since there exists a whole "add-ons" infrastructure to handle extras.

Are those things really taking up much employee time? Frankly, they don't seem like it. And some side projects are nice - like Rust and Servo. Persona too, even if it didn't caught on.

This is not interesting to the people involved.

They hijacked Mozilla to have a private playground for fun projects paid for by the search engine integration.