Hacker News Reader

Comment by huhtenberg | Hacker News Reader

> It is absolutely worth supporting powerusers.

Not just worth it, that's who their userbase actually is !

Powerusers, privacy freaks and their immediate social circle that they managed to strong-arm into using Firefox.

That's their remaining 3% of market share and Mozilla remains absolutely oblivious to it.

"Let's add whitespace everywhere. Let's remove RSS support, let's change the UI and also let's add as much useless junk nobody asked for".

Don't.

And their CEO must go. The fish rots from the head.

View on HN

FTP support too.

Which power user is using in browser ftp support?

The three people I know who use ftp for something have and always had a dedicated client for it.

Literally everyone? Not even just power users, but literally everyone.

I don't understand your position.

Are you claiming that no web pages link to download files on FTP servers?

Or are you saying that the right thing to do in that situation is to refuse to download the file?

Because both of those positions are self-evidently wrong.

Very few webpages link to FTP content. The right thing to do is to outsource that bit of functionality to an FTP client. Continuing to support a protocol that has largely been superseded and is very rarely used is self-evidently wrong.

Sorry. No. A web page links to a file, you activate the link, the link is valid, and the web browser... chokes? That is objectively a bug.

You are confused about what a URI is. Or about what a web browser is.

You need dedicated clients for plenty of protocols that can be in an href.

ftp is not a part of the World Wide Web. The World Wide Web is in HTTP. That is the “web” part of “web browser”. That they support other protocols is incidental.

No it isn't. It's a bug that web browsers had FTP support in the first place - that's the job of FTP clients, browsers should just be handing off the link to one of them. cf the 'mailto' URI scheme - I think we've all come around to agreeing that web browsers shouldn't also be email clients.

Since the only job of an a browser seems to spit out whatever http protocol says, we can conclude that it is not its job to actually render HTML and interpret Javascript because the HTTP protocol says nothing about HTML or Javascript or Web Assembly.

Email is of course out of scope. But i'd be happy to integrate file browsers (local and remote) and web browsers, for everything file transfer, markup rendering and multimedia viewing. Of course that shouldn't be monolithic, with pluggable viewer programs. Several parts could be:

- navigation (contains: bookmark, history, tab/window handling, etc)

- networking (contains: configuring proxies, credentials+cookies, traffic shapping, requests filtering/rewriting, caching+capture+replay)

- content display (html rendering + js sandboxes, userscripts and content filtering, pdf, multimedia, ..)

Can you provide a real example from the wild where mailto: linked to a file for download? Because otherwise it's hard to see how that's a counterexample to the principle I articulated, which is specifically about downloading files, a function web browsers perform routinely.

magnet: links link to a file to download. Web browsers don’t handle these. ipfs as well. And plenty more …

FTP has nothing exceptional to it beyond it being a shitty, old protocol that is much more rarely used than BitTorrent.

The Vivaldi development team firmly disagrees.

If it were still in widespread use, I might agree with you.

It is. Thanks.

But even if it weren't, the fact that a smaller number of people would run into the bug doesn't magically make it not-a-bug.

> the fact that a smaller number of people would run into the bug doesn't magically make it not-a-bug

I agree with you with respect to ftp, but I have to disagree with this. A site might link to a file with a gopher:// url or gemeni:// (these are real protocols) and the browser would of course spit out an error. That's not a bug, because that's expected behavior. No one, at present, expects a web browser to be able to access these protocols, despite the fact that both protocols are designed to access something corresponding roughly to "web pages".

The reason dropping ftp support is a bug is that there are a ton of real sites that link to ftp pages, and they do so because there is a history of ftp browsing support in web browsers and it was reasonably anticipated by these page authors that viewers would not have trouble accessing files made available in this way.

It is not widespread enough to justify continuing to work on it, and making an active decision to not include a feature anymore due to little use is not a bug.

The Chrome team published some good information about exactly how widely FTP is used: https://docs.google.com/document/d/1JUra5HnsbR_xmtQctkb2iVxR... (see heading: Usage information)

Removing insecure protocols is a good thing, especially one that is commonly used to download software of all things. If you're one of the ~0.03% (as of late 2020) that relies on downloading files through FTP for whatever reason, there's nothing stopping you from downloading a more specialized client.

I truly don't understand what your problem here is: if you install an FTP client and it registers itself as a handler for ftp links with your OS, the process of downloading a file via FTP when you click on an FTP link should be pretty seamless.

I've been online since 1995. I haven't seen an FTP link for a looong time.

And so... Firefox should cater to your use case and drop support for others'? Is that really what you're arguing?

It's none of my business what parts of the web you frequent. But I'm only a hobbyist and I see FTP links all the time. Most recently while looking for older character set information for a homegrown libiconv-type project.

Firefox 90 removed functionality that has been a part of the standard experience for a quarter-century and is still in widespread use. It is a bug.

> And so... Firefox should cater to your use case and drop support for others'? Is that really what you're arguing?

I can ask you the opposite question... I do mostly think it's sad that they're dropping FTP support, but if it's expensive for them to maintain it for the little percentage of users (I'm guessing, maybe this is an incorrect approximation) who use it...

Btw I'm just telling you my experience, please don't be rude, it's hard to find allies that way.

I'm sorry if I came across as rude. It was not my intent. Can you help me understand what specifically came across that way? I would appreciate it very much.

I would agree it's probably a small percentage of users who use it, but I don't think that's a good reason to drop support. If software only kept features used by a substantial number of users, SharePoint wouldn't allow exporting lists to Access databases, Excel wouldn't allow pivot tables, and so on. I think there's a cliché about how 80% of Word users only use 20% of its features... but it's a different 20% for everyone. As a result, adopting such a principle would lead to truly disheartening results. Will Firefox perhaps completely drop support for PIV tokens next?

I am not familiar with Firefox's architecture, I've only compiled it from source 2-3 times and only one of those times was setting up my own build environment (as opposed to a compiling package system). I've never hacked on it to any degree beyond the most trivial. So I could very well be wrong... but... I suspect it has not been very expensive to maintain, given that the protocol is simple and I believe has not changed in many years.

> And so... Firefox should cater to your use case and drop support for others'? Is that really what you're arguing?

The last sentence reads to me like "Are you for freaking real right now?!", incredulity mixed with anger.

Not hard to land in a tab with Firefox complaining that it doesn't know of an FTP handler. Google Scholar returns results to papers hosted on FTP all the time.

On the other hand, as annoying as this is, this is the one time that maintenance overhead actually makes sense to justify removal (out of all the times Mozilla has invoked that excuse), given how nonstandard FTP is and how kludgy the implementation has to be.

Reply on news.ycombinator.com