Comment by scrollaway | Hacker News Reader

Comment by scrollaway | Hacker News Reader

Parent Full thread

Oh jesus christ. The entire ML thread is bad news.

Why on earth would they move away from PDF.js (which is a fantastic project) and then use that move to justify Mortar?

Also, I just learned this from the thread: Why the fuck would they kill Shumway? I'm all for killing Flash, but Shumway is an important tool not just for migration, it's also really important for archival.

God damn it Mozilla. I've talked about this before, too: Hearing Mozilla executives, each Mozfest, talk about commitment to the free web. Commitment to open technologies. What commitment, exactly?

In a handful of years, Mozilla has managed to severely damage its poster child web browser's reputation by including then subsequently dropping drm, by including then subsequently dropping Hello (a good initiative with a horrible user experience), by completely screwing up the launch of Pocket (which they could only have done by massively misunderstanding their userbase).

In that same handful of years, Mozilla took some good initiatives, such as Persona and Firefox OS, completely botched marketing and experience on both, severely underestimated the required efforts on the latter, then dropped both projects.

In that same handful of years, Mozilla dropped support for its second poster child, Thunderbird, arguably the last decent FOSS email client.

It all screams mismanagement and I haven't even talked about half of the problems. I've been upset about all this for some time but now writing it all out I'm starting to wonder of Mozilla isn't just in a free fall.

PS: I 100% understand the concerns and difficulties of allocating development time but this is a pattern I keep seeing with Mozilla: Spin up a project, get it to 80%, give up. You could build a castle with the amount of efforts wasted on projects either doomed from the beginning (FxOs), or dropped near the finish line. This is all absurd.

Anecdotally, PDF.js is slow compared to native viewers and often renders documents incorrectly. And the project is five years old now. It's not bad as a fallback, but considering how important PDF files are, looking for a better solution makes sense.

Agreed. As an OS X user, I hate it any time a browser or website uses something like PDF.js instead of the native PDF support. The native PDF support in OS X is fast and excellent.

> instead of the native PDF support

You mean instead of just downloading the file? Or does Preview support opening PDF links.

Safari will display PDFs natively without requiring you to download it to disk.

And on iOS too!

Care of QLPreviewController [1], something I wish Android had an equivalent for...

[1] https://developer.apple.com/reference/quicklook/qlpreviewcon...

And fully sandboxed on both macOS and iOS!

Preview does support opening PDF links.

While convienent, I can't see how anyone would want to use a browser-based PDF viewer over a native one.

Convenience is exactly it.

Have to admit I find it pretty damn useful myself, even though I believe PDFs to be completely out of scope of a web browser.

I really wish the format would disappear. Everybody says it's useful at being a "perfect digital representation of what things look like on paper", which is certainly super useful for printing. But I cannot understand how so much information that should simply be distributed in either HTML (forms), epub (ebooks) or jpg/png (scans and digital images) is distributed over PDF, a pain-in-the-arse, overengineered format whose readers are always extremely slow and riddled with security issues and from which it's extremely hard to extract source data.

There are a number of places where PDF has advantages that HTML, etc can't match. For forms, the PDF spec has electronic signatures, which means that you can sign legal documents without having to find a fax machine, or mail things back and forth. For print documents, it can do things like multiple column text in a sensible manner, across page breaks. As a general document format, it allows you to actually describe paper sizes, etc, so you can email someone a document that looks like it will when it's finalized. For archival, it means you can, without a bunch of hacks, create searchable documents of old, scanned in content. So things like electronics datasheets are searchable, and still retain the formatting of the originals. PDF continues to exist because there are so very many places that the alternatives are an even bigger pain in the ass, fix those, and then you can start to get rid of pdf.

All sensible uses of PDF (except your first one; electronic signatures are ridiculous, but let's not get into that).

I'm talking about where pdf is used and shouldn't be.

Would postscript classify as an "even bigger pain in the ass" with regards to pdf?

The only time I ever start a native PDF viewer is when I need to actually print it. Apart from those rare cases, to me a PDF is a web page, just worse, and I want it to live where all the web pages live on my computer, in a browser tab, with all the facilities the browser provides (history, tree style tabs, frictionless access to Google, etc etc).

I'm sure ymmv depending on what kind of PDFs you use and how often; I'm on the lower end.

I got tired of constantly updating due to vulnerabilities. I haven't had a PDF viewer installed for years, I have PDFs open in Firefox.

I got tired of having Adobe Updater constantly running in the background (why not a scheduled task run periodically like Google Updater?!) and the need to update/reboot my PC regularly when a new version of Acrobat came out.

I switched PDF viewers briefly but ended in the same update cycles I was in with Adobe.

PDF.js isn't perfect but I've yet to come across a PDF that it couldn't open for me. I use it mainly to view things like insurance bills or other random documents where pixel perfect rendering is irrelevant.

This is why package managers are so important for any OS. They shouldn't be updating independently.

Agreed. I use Firefox on OS X and I love pdf.js. I found it a huge hassle to get native PDF working on Firefox in the past, and with pdf.js I just don't have to worry about security. I've also not had much/any compatibility issues with pdfs. Bills, concert tickets, and specs all render acceptably. I will be very sad to see it go.

Nonsense. On windows I have used foxit for a long time and I never had all that shit that you are talking about. On Mac if anyone told me that they used happily a JavaScript PDF viewer instead of the built in viewer I would seriously worry for their mental health.

As a scientist, pdf.js is simply not capable enough for any serious work with pdf.

I would love to do as you say, and read pdfs in the browser only. From what I have learned in the context of this discussion, switching away from pdf.js will contribute to enabling that.

> As a scientist, pdf.js is simply not capable enough for any serious work with pdf.

As a counterpoint, I regularly read (CS) research papers in pdf.js and have never had any trouble. One of the reasons I stuck with Firefox is the superior Zotero integration.

Hmmm... pure math papers are often fine but I used to still see misformatted math relatively frequently, maybe I should give it another try.

I might be in the minority, but I really dislike reading PDFs on OS X's Preview while using the touchpad because of accidental horizontal scrolling. I do not understand why Apple decided to make the "rubberband" effect when clearly there's nothing to scroll on the left and right sides of the PDF. If anyone has a solution to prevent horizontal scrolling on OS X applications (Safari, Preview etc), please do share.

Try this: Preview -> View -> Continuous Scroll.

It did the trick for me on macOS Sierra.

YMMV

I'll give this a try when I finally upgrade to Sierra. Thanks!

PDF.js sole purpose is to be non native. I don't want to rely on native existence of a safe enough, fast enough PDF renderer. PDF.js provides just that.

Then they should make PDF.js faster! The way it uses a canvas could be significantly improved. I guess the point is that they don't want to invest the engineering resources to do that, which is a shame but I understand the reasoning.

My work machine is now completely Adobe-free. thanks to PDF.js. I hope they keep this project afloat.

Shipping Shumway has some problems due to patent claims by Adobe, as I understand. It was thought that things could be worked out, but apparently that fell through.

So the options are to either keep developing Shumway but not ship it, or stop developing it and not ship it. The latter is a bit more attractive if resources are limited.

I have previously installed shumway but it never seemed to work on the websites I wanted to load. :( So perhaps the remaining effort to get it working was too much.

Thunderbird is still my email client and it's still getting updates. Not sure where it will be in a couple of years, but for now it's still very much decent. And I liked the idea behind PDF.js very much, have been a big fan, but it's still rendering documents incorrectly and doesn't do forms.

Mozilla is an organization that needs to be careful about allocating their resources. These projects are open-source and with enough interest they can continue. If Mozilla's Persona died, that's because it never had a future in the first place and killing it was the right call ;-)

And on Hello and Pocket, I really don't understand the drama. I think the Internet is making some of us way too pissy.

They're not killing PDF.js, just taking a step in a direction that might eventually lead to it dying. This is explicitly mentioned in the ML thread.

Shumway hasn't been worked on in a while. No idea why, but I would guess that it has something to do with how large the Flash feature set is. Long tail and all that.

The reasoning here is probably like so (speculating here):

- We want to deprecate most plugin APIs

- We still want to support flash

- Shumway isn't there yet, let's use pepper flash

- Flash and PDF probably need very similar feature sets as far as the API is concerned

- Supporting the subset of PPAPI needed for Flash is probably going to let us include PDFium too, which is already supported by Google. As long as we're going down this path, let's try to include PDFium too.

- In case PDFium turns out to work, we can decide if pdf.js needs to be replaced

I personally hope that pdf.js isn't replaced; since a JS thingy is far better than a vulnerability-ridden C program. At the same time, this depends on how much the demand for a full pdf feature set is.

I'm not sure what the "including then subsequently dropping drm" thing you're talking about is...

Yeah, Shumway shouldn't have been dropped. And It is annoying. And stupid.

OTOH, this probably means more work gets done on Gecko, Servo, the *Monkeys, and Rust. Which is pretty good.

> this probably means more work gets done on Gecko, Servo, the Monkeys, and Rust

Well, does it? Are they moving devs from PDF.js to the Rust ecosystem? It seems more likely to me that they would either lay them off, or move them to the PDFium stuff. Same applies to Shumway.

What a backwards step either way. As for Servo, how's it doing these days? Is it anywhere near done? What terrifies me is Mozilla would drop Servo when it's 80% done, just like it did with all its other projects, because "well, we already have Gecko, and it works, and it's too much work to finish Servo".

Aside from Rust and some Firefox tech, I haven't seen anything concrete come out of Mozilla in a long, long time. Rust, FWIU, has its own organization and is no longer handled by Mozilla alone (thank god); everything else gets killed off like a westerosi wedding.

> As for Servo, how's it doing these days? Is it anywhere near done?

Steady progress as always! You can follow along with This Week in Servo at https://blog.servo.org/ :)

Recently we just added flexbox support and the Fetch API, just to name two of many.

Thanks! Good to see a blog. I'm very excited about Servo; like I was saying, I just really hope it won't get killed off before it reaches the finish line and replaces Gecko.

Patterns are scary.

I don't think they'll kill Servo. They need to modernize Gecko, and there's so much potential in Servo to do so that it would be a waste throw the code out.

He's a troll.

Who, me?

No, scrollaway. Sorry for the confusion.

scrollaway is not behaving like a troll. Mozilla have engaged in some very bizarre moves in recent years. Whilst I don't think Servo is in trouble, it's fairly clear the leadership of Mozilla has become erratic. Thankfully the development of Servo goes beyond just one company.

I still miss Ubiquity from Mozilla Labs. Those were the days!

Keep seeing stuff related to bluetooth which only makes sense for mobile. Anything new on apk nightlies? Rustup with androideabi is a good sign.

Not necessarily moving devs; people have different competencies. But keep in mind that—despite being a nonprofit— Mozilla does pay its developers, and most of the work is done by those paid developers. Layoffs of non-core-project devs free up budget to hire more core-project devs.

As I mentioned elsewhere, there haven't been any full time mozilla devs on PDF.js for quite awhile. I'm not sure I actually see the whole maintenance cost savings argument since PDF.js has practically cost Mozllia nothing the last few years. A lot of bug fixes have come from unpaid contributors in that time.

Initially, PDFium was pitched as a freebie if we added support for chromium's flash and then we'd also get improved PDF printing and form support. However, the amount of effort that has gone into supporting PDFium is already far beyond what it would have taken to improve PDF.js form support and help improve Firefox's printing (which would have benefited the web in general). Though, this is my very biased opinion as I was tech lead of PDF.js.

Good god. So the project was already unmaintained, and now they're dropping it in favour of a maintenance burden, all the while regressing from a JS solution to a plugin-based native one?

I was wrong, Mozilla isn't in a free-fall. They hit the ground and are now digging the grave to bury themselves in.

I'm guessing flash support tipped the balance?

Reply on news.ycombinator.com