I'm of the mindset that code written 20 years ago should run today. This is a breaking change on a critical piece of internet infrastructure. Google, by nature of operating the most popular web browser, has a moral obligation to the community to provide a stable environment.
There are something like 10^9 webpages out there. How many of these will become less accessible because of this change? Many well maintained webpages will have to expend resources to migrate off ftp. Less maintained webpages with FTP links will not be updated and put the burden of installing an FTP client on the user, reducing accessibility.
This is Google externalizing the cost of updating their software by forcing everyone else to update their webpages to following Google's specification.
My direct and blunt rebuttal is "I don't care". The future is worthless without the present, and more importantly, the past. The continued effective destruction of history that these big tech companies are doing with their actions is a disturbing trend.
I think it's pretty much a given since most clients change to PASV mode as a first step. And it's rare that people have a port open to them (behind all firewalls, etc)
I don't think Chrome ever supported Gopher, but if some other browser removed something that worked fine before I wouldn't like it either.
I am more worried about the very realistic that that Google would be dropping FTP from search results, seeing as how their users won't be able to open them in the Google browser.
I'm of the mindset that it should be possible to rub code written 20 years ago, but not always the same way you run code written more recently. Emulation, virtual machines etc all allow access to old work without hobbling future development with backwards compatibility concerns (see: Windows).
In this case, there are plenty of free and open source options for FTP clients. No one is going to have the world closed off to them as a result of this.
This is a very silly response.
New versions are released with security fixes, not just new features.
I don't get why there is so much resistance. It's an old, outdated technology that has to be replaced eventually. Yes, change is always painful. But this isn't like with RSS were the tech provided clear benefits we would loose.
????
Using FTP doesn't put your computer at increased risk of being hacked.
Not updating your browser _DOES_ put your computer at increased risk of being hacked.
In any case, integrating FTP in your browser absolutely increases your chances of an exploit, though. When was the last time the FTP code was tested? Or worked on at all? I suspect it will have been a very long time.
Even aside from everything else, Chrome auto-updates and it’s very difficult to disable.
IMO, browsers need to concern themselves with HTTP. Anything else gets handed off to an application that's designed for it.
It's OK if some of the more antiquated parts of the web die off. I don't want to have to support old Flash loading screens, I don't need AOL instant messenger status badges, or iFramed in MySpace comments, and I don't need to open FTP in my browser. Most users these days wouldn't even know how to use this via browser... ftp://myname:mypassword@123.123.123.123... it's just a relic.
Anyone who wants to use FTP knows how to download a program for it. We don't need it in Chrome.
Sadly a decent fraction of users might just accept something like this, even though it would be a major step backwards for internet freedom.
Given that chrome didn't exist 20 years ago, this seems like a fairly absurd position to apply to this situation. No code from 20 years ago relies on chrome having ftp support.
FTP links will still work, they just won't be displayed in a browser anymore. People are blowing this way out of proportion, as if FTP is some critical underpinning of the web or web browsers are a critical part of FTP. Web browsers have always been terrible FTP clients, anybody who uses FTP on a regular basis uses a better client. The most tangible impact of this change is to give grumpy internet commenters something to complain about.
I believe otherwise. Let the old die, so we can alleviate ourselves to create something more suitable to our times.
Really, though? Are there any examples of any web pages that are strictly available from FTP instead of HTTP? If so, why? Isn’t HTTP made specifically for the purpose of transporting hypertext? Why use another, more general protocol when you can use the one made specifically for your use case?
I’m sure there’s good stuff on FTP still. I really, really doubt people are throwing SPAs and PWAs on FTP servers though.
We should stop telling others what they SHOULD do. (Crap I just did.) Anyways, why not contact Google and offer them to pay the costs of maintaining every piece of software they ever decide to deprecate? Or just go on and build a better browser if you can.
So you would have cp/m, MS-DOS, PDP-11 RSX-11M, Coherent, Vax VMS, old broken compilers for dead languages, IBM 360 TOS (yes, tape operating system), IBM 360 DOS, Sigma 5 Real-Time-Batch monitor, autocoder, 1401 software, Palm OS?
But for the love of all that is Balloons and Seashells, why?
More to the point--did Chrome have FTP 20 years ago? No, it did not.
In my opinion, they should have never put it there in the first place.
> It's not part of the modern web, so it's just another loose end to tie up.
While I see that it is shared by many here, I do not get this sentiment. It's still out there and rarely used, yes. But similarly would not require much maintenance for the time being, remove it when users decide it's dead, not just your shiny modern partners. The codebase cannot be that burdened by support for the protocol, why even take action?
I guess what I'm trying to figure out is the reason why this even came up, is that to soothe some weird code base / maintenance metric?
The linked Google Doc says that usage in a 28-day window was at 0.01-0.03% of users (not sure which figure to use there). Sure, that sounds horrid at first but think about how many users Chrome has, that's not nothing either. Instead of implementing encrypted connections to make their software better, they jump to deprecate yet another protocol. I just find this really hard to grok. How many people used the profiling features of the developer tools in that time frame for example? Not many? That's cool, do we get to keep those anyway?
Browsers are complex, yes, it just would be nice if we could expect a bit more of a graceful deprecation process with better communication. If you don't feel like supporting it, move it to an extension/plugin to separate it and put it somewhere the community can pick it up easily.
Having worked with ssh/scp/sftp, I think sftp could be better too, and therefore more useful.
It's intimately entwined with ssh, which means that instead of just transferring files, you have to give out ssh permissions where they're not required.
Really, there should be an sftp client and server, strictly dealing with files.
- You could have "accounts" that weren't real users on the machine
- You wouldn't have to deal with users and shells and groups and permissions
- You wouldn't have users running commands on your machine
- You could have read-only sftp (sort of a repository with access controls)
- You could have write-only sftp (sort of a dropbox)
- You could make these formal, with daemons like read-only-sftpd or write-only-sftpd that were securely purposed to prevent unexpected file operations.
Read only isn't a problem, as denying writes and allowing reads is a trivial part of unix filesystems.
For a time I was using Konqueror with fish:// URLs as a SFTP browser and it worked very well and made me happy ...
I am not sure if that workflow still works but most rsync.net customers that I interact with are using sshfs on top of FUSE and mounting ssh resources as a local filesystem - so no browser involved.
Rather, there are no ads on FTP pages, so Google is going to use it's power to stop countless people to stop using FTP.