Hacker News new | past | comments | ask | show | jobs | submit login
Microsoft Is Killing ePub Support in Edge Classic (thurrott.com)
198 points by ingve 51 days ago | hide | past | web | favorite | 132 comments

Oh what? That is really bad. Epub support in edge was the one thing I used and was impressed by. The text looked great and it supported very good read-aloud that tracked the words in the text. I was impressed how well it read tricky things like numbers and sub-clauses. If you switched to one of the British accents, crank the speed to 2x and it was a great way to power through some books that might have become a slog. When you begin to tire, switch the gender of the voice and it suddenly becomes more digestible again.

I just don't know why MS spends so much time shooting itself in it's feet. So much talent and promising product gets burned for indecipherable reasons.

It's obviously not a configuration suitable for everybody, but I read epubs in a similar manner with Emacs, using the read-aloud and nov packages. It highlights each sentence as it's read, using the system TTS (MacOS's `say` in my case.)

    (use-package read-aloud :ensure t)
    (setq read-aloud-engine "say")
    (use-package nov :ensure t)
    (add-to-list 'auto-mode-alist '("\\.epub\\'" . nov-mode))

    (bind-key* "M-SPC" 'read-aloud-buf)

Killing decent products is the Google way. Now that Microsoft is using chromium maybe the Google way is seeping in.

Invasive telemetry and profiting from ads instead of selling software is definitely the "Google way" and it's been taking over MS since at least the beginnings of Windows 10 (remember the forced "free upgrade"?)

I mostly view this as the turn from SDET to Quality. SDET used to do both telemetry and quality testing, coordinating with devs frequently. Now Quality mostly coordinates with PMs. Satya as Azure head had success via this method, so it's copied throughout the entire company

Do you support telemetry if it makes the product better, and not used for ad targeting?

Not parent but it's something that's been bugging me for a while (especially since I work in a field where you often have to debug based on field logs from three weeks ago from a device halfway across the world). My two cents, I think telemetry is OK if:

* It has a retention policy that ensures a data breach five years down the road won't compromise five years' worth of relevant data. 30 days of full data + maybe an year of relevant analytics (but no detailed data) are more than enough to improve software. Five years' worth of data, even if it's not strictly speaking personal data, can make identity theft trivial.

* The processes and criteria for who gets to access that data and in what manner are clear, transparent, enforceable and can be subjected to appeal by any user. Can I be reasonably sure (i.e. based on public, legally-binding statements from Microsoft) that, for instance, a Microsoft employee who has a beef with me won't be able to stalk me based on that data?

* For a large system or a piece of hardware, if I can have a reasonable assurance that, at least in the short-term (1-2 years), the data won't end up being used for ad targeting. An OS or a piece of hardware is something that you don't switch that often. If I make a purchase, I'd like to be sure that the reasons why I made it remain valid for a while.

"Makes the product better" is a great reason and I'd wholeheartedly support that, as long as I had an assurance that my data is handled responsibly.

I'm just sayin' -- if I break the terms of doing business with Microsoft (e.g. by using a pirated copy), there's a good chance that some DMCA organization comes knocking at my door and the best I can hope for is an out-of-court settlement that ruins me. I sure as hell expect that, if Microsoft breach the terms of doing business, someone can go knocking at their door and the best they can hope for is an out-of-court settlement that ruins them. That's why I insist on "legally binding". A blog post that says we totally don't spy on you is something that you can breach without any real consequences.

Until that's the case (or, you know, until there's no data being siphoned...) I can do my work on Linux just fine, I don't need no damn WSL :-).

I think "make the product better" is really an excuse, considering what has actually happened with the changes in Win10. The most useful changes have been suggestions that came from the users themselves (of which they ignore most of them anyway), not by watching them. "The telemetry told us to do it" is a convenient excuse for making unwanted changes, because everyone else doesn't have access to the data and even if they did, it could still have been massaged to fit their narrative.

Not sure why you're being downvoted. Telemtery can be a invaluable tool for a whole bunch of reasons. It should, of course, be optional.

I think only 1 type of telemetry shouldn't ask for permission, a quick, anonymous ping to say it's been installed.

a quick, anonymous ping

Sure, but for it to be anonymous it should not contain the sender's IP address. That probably means an UDP message with nulled source ip, which runs foul of many ISP's egress filters (and rightly so).

So, realistically speaking, there is no such thing as an anonymous ping. Or am I missing an obvious solution here?

You could have an aggregation service, run by a trusted independent non-profit. It could have a strong privacy policy and external auditing.

How would we know, the is no documented case of this ever happening?

No. Because keylogging and capturing other UI events can be used to reconstruct what you're working on. It thus fails basic obligations the law puts onto you when working with customer's documents, data, or code. Using telemetry, and implementing telemetry in your app is just asking for getting sued big time.

Only if it's optional, and can be turned off completely.

And if its presence is made clear upfront, rather than buried deep in a Terms of Use contract.

That should be up to the user to decide.

If millions want to provide telemetry that's enough to make the product better. Why have other millions who don't want suffer it?

IF it is disabled by default and getting rid of it is not an enterprise feature... sure.

It used to be that Microsoft was famous for vaporware--product announcements that never materialized, to take the wind out of competitors' sails.

Not sure which I like better.

Microsoft is also good in building features that don't work or are unusable in any normal use case. Like Knockout support in Visual Studio: I have never seen it working. I believe SQL Server also has a few features that were until the current versions not really usable.

Uh huh. You should visit the Building 17 courtyard sometime, it's basically a graveyard for defunct MS products.

> Killing decent products is the Google way.

If you were around since the 1990s, you'd notice that Google learned this from Microsoft.

Microsoft is the king of backwards compatibility. They might propose a new framework every now and then, but they still support the old decades on...

> I just don't know why MS spends so much time shooting itself in it's feet.

Removing a (great) feature that almost no one uses, in a product that is being sunset is shooting oneself in the foot?

The biggest thing that was exciting to me was EPUB having native OS support, effectively, via Edge. You could assume an EPUB was readable in Windows without downloading something, just like you can assume for JPGs and TXT files. This feels like a huge step back for the format to lose this.

I just realized that EPUB is the CHM of the modern world.

Exactly. The original html based document format.

Which was not a bad format. Even as thorough Linux user with no Windows in sight, I have often used .chm files for handy, offline documentation. PHP docs come to mind, back when PHP was something I occasionally did.

Text, markup, arborescence, embedded medias, all in a single file, supported natively by the os and comparable between apps. Before the internet everywhere there is an entire generation of windows software, especially developer tools, where chm was the norm for high quality documentation.

Yeah i really appreciated the chm format too. They killed it or is it still alive?

Mac OS guides are also a very good editorial/digital work. (but no videos?). The guides on MacOS look very simple and clean. I appreciate when the software I buy has a guide like that (pixelmator) and not just a link to a website (sketch). I know the required effort so I value them a lot.

It still works just fine but the tools haven’t been updated for a while, almost no modern software supports exporting to it and it’s not “cool” anymore. So now we use often inferior stuff instead, or merely get a link to some online documentation, until someone “reinvents” it five years from now probably using some kind of electron variant making it cool again.

Hm, almost every desktop program i have on Windows that bothers with help uses CHM and pretty much every documentation generator can generate CHM files, why do you think it isn't used?

Everyone Googles for the problem or uses some monolithic tool.

Even notepad’s documentation has gotten worse over the years. In Windows 8.1, notepad doesn’t even mention the “.LOG” automatic date/time add feature.

Try to find node, go, vscode etc ... documentation in chm. They have third parties versions sure, but not official. And for those that still uses it, it’s rarely up to date.

Of those only VSCode is a desktop application and VSCode is already a web browser. Also considering it is meant to be modular, its documentation system would also need to be modular and i'm not sure if CHM is good with that.

For many modern languages, the language ecosystem often comes with a tool to generate docs that also work offline.

It had massive security problems, so they mostly killed it.

eg https://blog.checkpoint.com/2015/05/12/the-microsoft-help-fi...

FWIW Lazarus[1] has its own CHM generator and CHM viewer which uses its own HTML engine written on top of its LCL framework, making it cross platform. Though personally i prefer the native viewer on Windows as LHelp (Lazarus' viewer) feels a bit clunky.

(the generator actually comes from Free Pascal, but the viewer is only on Lazarus)

[1] https://www.lazarus-ide.org/

> “Download an .epub app to keep reading,” a notification in Edge classic reads when you load an EPUB document. “Microsoft Edge will no longer be supporting [sic] e-books that use the .epub file extension. Visit the Microsoft Store to see our recommended .epub apps.”

What’s wrong with this? It looks grammatically correct to me…

It seems fine to me, too, if not the most direct wording. "Microsoft Edge will no longer support e-books..." is shorter, but the existing wording isn't wrong.

Yeah it’s grammatically correct, but semantically simple future would be a better choice to convey the state of future support.

I read the statement and to me, the future progressive seemed to be the most apt way to phrase it. When I tried to find examples to illustrate why, it occurred to me that the future progressive predominates when delivering bad news. For example:

"Unfortunately, we won't be renewing our contract with you"


"We regret to inform you that we will be going with another candidate"

Changing those to the simple future makes them sound much harsher to my ears, so I would venture that the progressive is a mechanism for softening the impact.

> What’s wrong with this?

This is answered in the next sentence of the article: it’s “support” not “be supporting,” Microsoft.

The is using the future continuous when the simple future is more appropriate to the message.

"will no longer be supporting" is completely fine, though.

It's also more consistent with "no longer", which implies a progressive action. An alternative is "will not support", which is no longer indicating (or does not indicate) that it had been supporting epub for a while.

That said, "will no longer support" is more succinct and clear enough.

> This is answered in the next sentence of the article: it’s “support” not “be supporting,” Microsoft.

That's not an answer to "What's wrong with this?", that's the author's preferred wording.

Both the future and simple versions are appropriate, whereas the author makes it a point to highlight their preference.

And of course here we are, poking it further.

“Microsoft Edge will no longer be support e-books that use the .epub file extension” sounds very wrong.

“support”, as the author of the article suggests, should replace “be supporting”, not just “supporting”. The simple future is called for, not the future continuous.

You misread the suggestion.

Will no longer support

[sic] is Latin for self-satisfied smirk

I wonder if "no longer will be supporting" is more technically correct, to keep the verbs together.

No, it would be less correct. Adverbs in future continuous (and other composite tenses) go after the first verb.

To use [sic] the quoted text has to be clearly wrong or outdated grammar, not slightly less correct.

> To use [sic] the quoted text has to be clearly wrong

(1) You are responding to a quote about a proposed rewrite of the sentence which would be clearly wrong independent of whether the quoted text was (the “less correct” was chosen because the proposal asked if it would be “more correct”), and

(2) No, there's no such requirement for the use of “[sic]”; use for things which are not strictly incorrect but not comport to a stylistic preference of the author presenting the quote (or one that the author expects the audience to have) is common. For instance, Google's dictionary's usage example for it is “a story must hold a child's interest and “enrich his [sic] life.” What the “[sic]” flags is grammatically correct, though recently usually stylistically not-preferred.

It’s probably short for sic transit gloria mundi.

By no means. See https://en.wikipedia.org/wiki/Sic for usage of the word sic.

Would have been better:

“e-books that use the .epub file extension, Microsoft Edge will no longer support”

This sort of construction is reminiscent of topic-prominent languages, e.g. Japanese, but English is a subject-prominent language and phrasing it that way would sound awkward to a native speaker.


The author is probably an English major. He was waiting all his life to finally say 'you see, my degree is not completely useless'. He would, of course, be completely and utterly wrong - both in justifying his degree, and his analysis of relnotes grammar.

Hey, there’s no need to be mean.

This makes no sense to me. EPUB is basically HTML in a ZIP container with mandatory included files. If you already have a full-blown web browser, and a ZIP library at hand, naively rendering EPUB pages by unzipping to temp files and tweaking the URI resolver is an easy addition. If you already have a full-blown web browser that already renders EPUB pages better than a lot of EPUB-specific reader programs, removing that support is doing work to reduce functionality.

If it isn't used a lot, I'd prefer the budget to go somewhere more useful instead.

They don’t have to maintain it, they just don’t need to completely get rid of it. That part is baffling.

> They don’t have to maintain it

Yes they do. Rarely used, unmaintained code is a great way to accrue security vulnerabilities and other bugs.

They would still have to maintain it, but they wouldn't need to add features. If they coded it well enough in the first place, that could be one person putting in 8 hours a month on the highest-priority issues in the backlog.

EPUB is a container format. It's basically an encapsulated multi-page website with some metadata. The marginal effort for a web browser is unpacking the files with a ZIP library and deciding how going from chapter001.html to chapter002.html works inside the existing UI. If you don't want to think much about the latter, render each page exactly as if it were downloaded from an online website, and use the forward and back buttons to step along the built in reading order.

I actually wonder why all web browsers don't support it natively.

Security vulnerabilities can only occur when the software bridges two different levels of privilege ("crossing the airtight hatchway", as Raymond Chen puts it). An EPUB reader has no business doing that, it can run completely inside the same restricted environment as any other potentially-malicious website.

Other bugs, sure, but since when did presence of bugs prevent anyone - much less Microsoft - from shipping anything? Especially since the whole browser is going away anyway, so there's no long term burden.

Didn’t Microsoft kind of do that with IE 6?

a. How well did that work out? and b. How does the ‘tech community’ view Microsoft for that choice?

An ePub reader is a little different than an entire browser. They’re already getting rid of Edge Classic at some point anyway; why remove a useful feature before then?

I'm very surprised how bad most ePub readers are. You would think that by now picking an ePub reader for the PC would be a no brainer but it seems there are very few good ones. Too bad, the ePub reader is very good. It's the only reason I use Edge.

Same with physical ebook readers, really.

How is it bad? What is the problem? The only thing that doesn't work properly is displaying PDF because many PDFs are setup bad. And sometimes - in epub - images are too small to display properly. That is something that really can be improved - image display and handling.

Not responsive enough when changing pages. Too much tech crammed in to ramp up the price of something essentially worth $15. Navigation is unintuitive, it's hard to get stuff on the device. Not all formats are supported by all readers and PDF support isn't what it should be. I can't turn pages with either my left or my right hand, no easy switching beween books and if you've got more books on there than the average person, like maybe 70, you're already screwed to find what you want to switch to reading at the moment quickly.

PDF display is another pain point. We are over 10 years into the Kindel era but it's still a problem to read PDFs on all ebook readers. Why? Why are there no display guidelines so you can read them with ease or at least a way to reformat them so they can be read in an eReader?

This was literally the only thing Edge was actually good at...

Dammit Microsoft!

It's a reasonably good PDF reader too. Because of it Windows got nearly as good as Linux on that front.

Still worth it to replace it with Sumatra right after install though in my opinion.

No kidding. I was just thinking about that last night!

They should spin off the old Edge as a reincarnation of Reader from Windows 8.

It was a great app and I hoped it would become as versatile as Preview on Mac.

It was annoying when I double clicked a EPUB (or PDF) it would open the browser and sometimes all my tabs from last session would open and in most cases just the EPUB would open and I lose the previous browsing session.

This needs to be a dedicated app separate from the browser. I hope they have something in the works.

Not from MS but I think this one is great:


Does windows have anything analogous to quicklook? That's one feature I think Apple did well with.. though it's obnoxious it can no longer be extended to support non-MPEG-LA video formats...

I've experimented with using mpv controlled over json IPC as a FOSS substitute to quicklook, which works well for media but not documents...

Both QuickLook and Seer do this and both have plugins making it much more akin to the Quick Look circa Leopard and Snow Leopard. QuickLook is more actively maintained but both are very good.

[1]: https://pooi.moe/QuickLook/ [2]: https://www.microsoft.com/en-us/p/seer-pro/9pgvfjbvbzwx?acti...

Yep, I recently moved to windows 10 for my personal machine and as looking for equivalent to macOS features. QuickLook just work, and has even more options than the version from macOS

Microsoft have a weird habit of releasing niche or experimental products in a fully fledged form and then not knowing how to make them useful or popular (3D Paint, anyone?)

Did they really expect a solid ePub reader to be a significant selling point for the browser?

>Did they really expect a solid ePub reader to be a significant selling point for the browser?

Well it was the only reason I ever opened Edge, so it kind of worked?

It is for me. Also, the PDF reader, and the “Add notes” feature that allow you to annotate any web page. The support for the Surface Pen is just excellent in Edge.

I think it was a mistake integrating the browser with an Epub reader. Having said that, it was still a very nice and elegantly designed part of Edge. So I would like having it as a stand alone app, but I don't expect MS doing the right thing. They will either kill it completely or rewrite it from scratch with Electron.

Using a browser to read an Epub which is effectively just some bundled HTML seems like a very logical thing to do to me.

agreed epub is basically a zipped html, so it's sensical to implement it in a browser

That is like saying a Zip file manager should also include Epub reader since Epub is basically a zip file.

Reading a book and browsing web are fundamentally different activities and need different UX and business logic. Putting book reading capabilities to a browser makes the code base bloated in my opinion.

I have to say I disagree there. "Our program for handling HTML+CSS content should handle HTML+CSS content stored in containers" is a much more reasonable idea than "Our program for opening containers should also handle whatever content is inside those containers." It's a common thing for programs to do. VLC plays videos in zip files and RetroArch plays games in zip files, for example.

And I don't think reading a book and browsing the web are fundamentally different activities. They're near-identical. The web was created as a way to access documents consisting of text, images, and links to other points within the current document or to other documents. A book is a document consisting of text, images, and links to other points within the current document (endnotes, footnotes, "see page X") or to other documents (references/citations). The UX and features you'd want for a book reader are the same you'd want for a browser's Reader Mode, as implemented by Edge, Firefox, and Safari. I don't see any difference between reading an ePub file and reading a long article on a website.

I really wished they hadn't killed Edge. And I love EPUB in Edge for technical books - you can open several in different tabs and also search within them. The browser is the most natural vehicle for EPUB! Using a third party reader on Windows is going to be a pain.

It boggles the mind. Literally everything happens in the browser except reading hypertext documents.

Why won't they just open source the reader instead of abandoning it? I would love to have at least one 100% working epub reader on Windows...

> I would love to have at least one 100% working epub reader on Windows...

SumatraPDF is fine to read EPUBs on Windows (also supports PDF, MOBI, FB2, CHM, XPS, DjVu) - it's free, open-source software [0][1].

[0]: https://github.com/sumatrapdfreader/sumatrapdf

[1]: https://www.sumatrapdfreader.org/free-pdf-reader.html

Bizarre decision. I've done my share of epub wrangling by hand (that is with Sigil).

Epubs are basically zipped HTML. How hard would it have been to leave basic support?

So now one has to "freeze" Edge to keep epub support.

I doubt it is as easy as moving Edge to a different folder. Probably need to mess with regedit as well.

Oh, it could do that? I still use Calibre and Sigil. Then again, how can any browser not in some way support ePUB? It's just a bunch of hyperlinked XML files in a container. Rename it to a .ZIP and you're good to go.

While I'm sure that there are some adequate ePub readers for Windows desktops, I find myself doubting that there are many actual good ePub readers. At the very least when I looked from a Windows phone a year or two back there was nothing very good in the Windows App Store. Maybe Microsoft's push towards the Surface family particularly the Surface Go as a tablet means that there will be something worthwhile.

At least, that's the dream.

This is quite late, but maybe useful for later thread viewers.


I've found SumatraPDF to be the lowest overhead document browser for Windows. Surface book is actually a really nice piece of kit. I'm surprised the surface laptop came from MS.

This is really unfortunate. Killing one of the best applications? I mean it's fast, reliable, simple, easy to use, integrated...

"X browser is killing support for Y format"

Well, good. Now we can go back to when browsers handed files off to helper apps that supported the formats in question. I'm all for it.

You say this now but I’m, for one, very glad pdf.js exists.

"while adding api for direct hardware access"

strange cultural shift

After getting a Windows Tablet, I was looking hard for a good epub/pdf reader with annotations. To my surprise, Edge was by far the best reader out there. However, a few weeks later an update killed the annotations feature - I just couldn't save them anymore. None of the support-provided solutions has worked. I then started using Calibre to convert my epub files into azn3 and open them in the Windows Kindle app, which is also quite nice.

Very annoying, but I guess makes sense as chromium doesn't support epub and they'd want to keep the different edges consistent. And the old edge will be gone soon anyways..

A natural question is: is edge still going to be the default PDF reader in Windows? And will edge's pdf viewer look just like chrome's

Probably and it does in Dev and Beta builds (which is a disappointment because "Edge Classic" PDF Viewer was better and Chrome's a distinct regression in some disappointing ways).

yeah, you'd think it wouldn't kill them to have whatever you call the opposite of continuous scrolling?

They are starting wrapping up and deprecating Edge classic, as the Chromium one is now in beta

Can someone explain why one would use ePub, vs pdf? PDF seems way more popular.

I'd say PDF is all about presentation, while ePub/Mobi is closer to real books that focus on content and legibility.

EPubs are perfect for long-lengthy documents, where you can bookmark position to given line of text (not page, so I'd say it's more precise to get back in which paragraph you've finished reading), everything is searchable/copyable/highlightable, formatted into single-column, usually justified, wrapped using hyphenation, easy to change font size, colors (i.e. day/night mode) and there isn't strict paper (canvas) size. Basically, it increases accessibility and focus.

While the PDF is just like a ZIP (I think those formats share the same popularity), that consist collection of numbered crisp, anti-aliased PNGs and SVGs, sometimes with an attached metadata. I often hear, there isn't any chance to PDF document look different on other device. However, PDF is just a subset of PostScript (hence the name) and has many versions (standardized as a public document, by ISO, by other parties or just paywalled)[1]. Fortunately, quite old PDF1.5 version has majority of features used by typical document creator, also fact that there aren't many competing implementations and they seem to be up-to-date with PDF2.0 convinced users and developers to use it. That's worth noting, PDF supports reflow [0], but it is far from perfect and I guess many bookworms will refrain from using it at all.

[0]: https://en.wikipedia.org/wiki/History_of_the_Portable_Docume...

[1]: https://helpx.adobe.com/acrobat/using/reading-pdfs-reflow-ac...

Can't they develop an EPUB extension for Chrome (new Edge)?

Possibly, but they likely don't have any incentive to do that and are just going to leave it to third-party devs. Their eBooks store for mobile was the reason they added it to Edge Classic in the first place, and that store has already shut down because no one was using it.


The use of the future continuous when there is no reason not to use the simple future is both awkward and needlessly wordy, though it's more in the class of things (in a programming analogy) you'd catch with a linter than the kind that throws compiler errors.

I honestly don’t believe anyone would look at the phrase “we will no longer be supporting X” as awkward. The meaning is immediately obvious, and the construct is used frequently. Calling out a trivial style problem like this in a technical review is amateur.

It adheres to convention. "We will no longer be providing complimentary dessert" etc.

I wasn't aware of any browsers having epub support. Additionally, the practice of using browsers as your OS is bad. Native applications are the way to go. Pdf support should also be removed from all browsers.

> Pdf support should also be removed from all browsers.

Why? Instead of opening a PDF like any other webpage, or image, or video, etc. Which is instant and doesn't require any mental focus shifting, I should instead download the PDF, and use a PDF viewer instead?

For something so ubiquitous it seems kind of silly. Browsers are for browsing and consuming content, if a specific type of content has widespread usage, I see no reason not to include it.

> Which is instant and doesn't require any mental focus shifting

Opening URLs from the browser to native apps specialized for their format shouldn't be non-instant or require any mental focus shifting.

> Browsers are for browsing and consuming content

"Browser" is short of "Web Browser". They're meant primarily for HTML pages. General "browsing and consuming content" is too broad. You're including file managers and multimedia players in that. There's no reason to concentrate everything into a single program. Shifting between specialized applications should be seamless. If it's not, the window manager / desktop environment is lacking configuration / features.

> if a specific type of content has widespread usage, I see no reason not to include it.

Because you end up with a bloated mess. The modern web browser is an OS inside an OS.

> "Browser" is short of "Web Browser". They're meant primarily for HTML pages.

I don't see how you get from the first of these statements to the second. Web browsers browse the web—an interlinked set of hypertext documents. A "hypertext document" being anything that embeds URLs you can tell the browser to navigate to, usually by clicking. A PDF is a "hypertext document."

> There's no reason to concentrate everything into a single program.

An ePub is also a hypertext document, though. And one that is made out of HTML and CSS files! Would you suggest that a web browser shouldn't be able to open an MHTML archive? Because an ePub is almost exactly the same format, just with a different base CSS style + media selector. (There are some restrictions on scriptability, but from what I recall those are ePub UA restrictions, not restrictions in the standard.)

> Shifting between specialized applications should be seamless.

What if you could shift between specialized applications... inside other applications? Remember OLE? That's how web browsers have traditionally displayed most formats. The whole need for special plugins for QuickTime, Flash, ActiveX, Java, etc. was just to allow those formats to be embedded in HTML pages. But if you're just opening a URL in the web browser and handing the viewport over to a COM component to render—browsers (and many other types of applications) do that just fine, without any need for "concentrating everything into a single program."

One problem is browsers have incomplete PDF implementations. PDFs with 3d models is big in the engineering industry as dumb as it is and Chrome completely fails there and results in numerous less engineering user complaints :/

PDFium and PDF.js have only implemented a small subset of the spec. Other features they don’t support include certain image compression formats or colour management. It’s really only suitable for the simplest of PDFs...

PDF is a really complex format that Google and Mozilla don’t want to spend a lot of time maintaining.

PDFium is an an acceptable rendering engine for simpler PDF files, but doesn’t handle complex files very well (resulting in crashes or rendering inaccuracies). Same with PDF.js. Unless the browsers step up, these issues will continue to get worse.

An ePub is an archive of HTML files plus a bunch of metadata. It's almost literally a website in a zip.

In-browser PDF support is what finally cured the internet's allergy to PDF links.

EPUB is based on XML if I recall, so needing a third party app to view them seems burdensome.

More like zipped html and some metadata, but yeah, basically it's that simple.

PDF support should be removed from all applications, including PDF viewers. Horrid format...

What would be the alternative? Postscript?

TIF? No thanks

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact