That seems really ass-backward, and not just because I use the search feature a lot.
If I'm reading that right, they're deprecating support for discoverable browser-independant markup for searches; and replacing it with the requirement that each site actively develop (and maintain!) a software plugin for every browser their users might want to use.
The whole point of a "user agent" was to go out and do things for me on the web; and the idealistic goal was that each person could choose an agent suited for them, which then had tools to programmatically discover and interact with the web in a common manner (reducing engineering load on the webdevs).
And I don't want to try and use a separate search tool (with new flashing graphics and ads!!!!) for every site I go to... I want a single search tool, like FF offers right now. (Aside: not to mention chrome's "auto-discovery of opensearch when you tab after typing a domain" is actually MORE useful than FF's manual mode!).
Taking a step back and removing support for a declarative api seems to me like the really wrong direction for an open web. Instead of sites supporting a single declarative browser-independent markup; they now have to deal with a long tail of (2-3 + who knows how many) browsers; and users with a niche browser have to spend effort convincing every site to support their browser.
Why not try to improve the opensearch markup instead?
Followup - I'd actually been using chrome a bit more heavily; and was wanting to use FF more, just to support open standards. One of the main things I was missing was being able to type in google docs domain, tab, and type a document. I was planning to research how to make something similar work in FF, and now I know how, and that they're removing it :(
There is ancient feature of regular bookmarks to serve as "search keywords" [1]: just give it a some keyword and use `%s` in place of URL you want to substitute with value encoded as URI fragment (IIRC, or `%S` to be used more verbatim). In Firefox it is directly in "New bookmark" form; in Chrome it is dug somewhere in "Search engines" corner of settings.
For example setting keyword `t` for uri `data:text/plain,%S` and entering `t foo` into location bar will navigate you to `data:text/plain,foo`, i.e. "make document". If Google Docs have GET endpoint for creating documents, it should work. For searching you can apparenly use `https://docs.google.com/document/?q=%s`.
The parent comment was referring to how Chrome automatically adds search engines. You only have to "dig into" the settings menu if you want to change the keyword or add a custom search.
Even if you're adding it manually, I think the list of search engines is a more intuitive place to put such a feature than "bookmark keywords".
Compared to similar features in Chrome or DDG's bangs, Firefox's bookmark keywords seem less discoverable to me.
The exact same trick works with custom search engines in Chrome as well.
They're equally powerful. Automatically adding search engines means users don't have to do so by hand, but you can still manually create a "search engine" too (e.g. "https://xkcd.com/%s/" with the keyword "xkcd").
In Firefox I keep a folder with bookmarks that have keywords, but I would prefer the UI in Firefox's search engine settings (the bookmark manager doesn't have a keywords column).
The problem is few will discover Firefox's bookmark keywords unless they're told about the feature, and manually create such bookmarks, while Chrome automatically creates keywords for search engines and prompts users to try them out.
I completely agree that it's a pity such nice feature isn't known better among wide audience and yes, Firefox bookmarks management ("Library") UI leaves much to be desired. [2]
Just one reminder: in Firefox there is "Add a Keyword for this search..." command on any (form) input field that triggers keyword bookmark creation wizard [1], so what Chrome does automagically by visiting page with form (or using the form once?) you can do quite easily in Firefox as well, but you must find the input field, shift+f10 or click few times and pick keyword.
Also, using same keyword for different URL will (at this moment) silently "transfer" the keyword to new URL, with no warning about
[1] yet again, this wizard obscures resulting bookmarked URL with relevant `%s` part, so regular user cannot find out how this thing works. (I'm sad how hard recent browsers tend to hide whole concept of URL from users, in general. I understand it, but it's sad.)
[2] I had to `select moz_keywords.keyword, moz_places.url, moz_places.title from moz_keywords inner join moz_places on moz_places.id == moz_keywords.place_id order by keyword;` last time I wanted to see all my keywords. (And I'm trying to keep them in a single folder as well.)
> I completely agree that it's a pity such nice feature isn't known better among wide audience and yes, Firefox bookmarks management ("Library") UI leaves much to be desired. [2]
My main worry is that I mainly use tags to sort my bookmarks before using folders, and even after many years they (tags) aren't showing up on mobile. I'm afraid they'll eventually pull tags support out and become way less useful.
Tags are IMO a better sorting system than folders.
I've ended up using neither tags nor "topic" folders (besides few folders in toolbar): for retrieval I rely on titles (names) and URLs alone. Every time I bookmark something I evaluate its title and URL, look for missing terms my future self could use, and then either reword the title or add raw "tags" in the end. Extraordinary bookmarks I supply with several * proportionally its extraordinarity (and make sure names contains no such sequence).
It serves me well: using native bookmarks search accelerator [1] in url bar * , keywords and 'rating system' allows me to for example quickly pull "best personal blogs of people writing about javascript":
* *** javascript guru
I understand your worries, IIRC there used to be bookmark "description" field that was just recently removed (I've used it maybe twice) and if tags are globally used like this description was, your concern could be quite relevant. If need arises, perhaps some sqlite-fu could transfer tags to titles, at least.
I have noticed that menu entry for years, and tried using it, but I had no idea how (and I guess it wasn't big enough of an issue for me to look it up). Thank you for explaining.
I for one am rather curious about many recent Mozilla moves. They seem to be away from decentralized and distributed models, towards centralized ones.
You ask why not improve opensearch markup, just like I have asked why not improve RSS. I am quite convinced that the RSS model has a lot of untapped potential for a distributed internet, but it also thwarts central control, surveillance, tracking etc.
This title is a little misleading. The original title of the post is "Search Engine add-ons to be removed from addons.mozilla.org", which implies that a special class of addons will be removed from the online store and merged with regular extensions. Firefox has never supported OpenSearch in the same way Chrome has, where you can type in "yout" + tab and search YouTube immediately.
Opensearch engines could be added to firefox from this page: https://addons.mozilla.org/en-US/firefox/search-tools/ .This in practice was just an extension but a different schema. They probably just want to consolidate and maintain only one schema for all extensions.
Now, it is still truly surprising to me that it's easier to use third-party search in Chrome than in Firefox, given all of Mozilla's talk about the Google monopoly. In Chrome, adding a search engine is as simple as going to chrome://settings/searchEngines and adding a string like "bing.com?q=%s" while in Firefox the only way I can find out so far is packaging an extension and loading it up.
You can do this in firefox, too, but it's quite hidden and doesn't have a nice UI. Bookmarks can have keywords associated with them in the bookmarks editor, so you can add a "Bing" bookmark with URL "https://bing.com/?q=%s" and Keyword "bing". It'll work, but the "bing" prefix won't be highlighted as it is in Chrome when using it to do a search.
The "Add a keyword for this search" context menu option automates creation of this bookmark, but you can also do it manually if firefox doesn't recognize a search field or something like that
It is quite "highlighted": after typing exact keyword it is the first suggestion below URL bar and after typing space and some query it reflects this by displaying "domain.tld: query" there, seams good enough to me.
The title is a quote from the second sentence in the article. The original title of the post is about the immediate change, but Mozilla is still planning to deprecate OpenSearch.
Firefox definitely has supported OpenSearch. When you arrive at a page that has an OpenSearch description document, there is a button in the address bar triple-dot menu to add the site's search engine to firefox. On mobile, you could long press on an input field to do the same.
Even though I'll miss it myself as a power user I think this is actually fine so long as there's a common standard for integrating third-party engines.
Opensearch came about when search was in its nasency and people envisioned all sorts of complex needs for it. 90%+ of opensearch schemas just end up being something like "searchengine.org/?q=%s" in practice.
The main thing we'll lose out in the simpler pattern is autocomplete, which you could specify via "<Url type="application/x-suggestions+json"
What the article doesn't make clear and I wish it did was what Firefox's future of supporting third party search engines looks like. I just came back to firefox after 8 years away, and it seems from some of the UI changes like the "This time search with" feature that they do want to encourage a diversity of search engines overall.
a. From the Mycroft project page
b. From the search engine's own page by right-clicking and selecting "add a keyword for this search".
In any case, it seems [1] the announcement is mostly about the addons.mozilla.org listing than the removal of OpenSearch or other capabilities in Firefox. The wording is confusing and the link doesn't completely clear it up, suggesting it may indeed be on the way out (but not necessarily on the same timeline as the AMO removal announced here).
You used to be able to make an XML file with the search query and base64 encoded icon and put it in the search engines folder in your profile. It wasn't the easiest way for non technical users, but once you knew how it wasn't bad. It made it easy to backup, together with bookmarks.html.
As of at least Firefox 69.0.3, you can right click most query fields and choose "Add a keyword for this search". It works about the same in e.g. Opera and Vivaldi.
It's unclear if the "intent to deprecate OpenSearch" also means removing this functionality.
I'm.. at a loss. Please, if any Mozillers are on here, can you explain? I want Mozilla to be successful as a proponent of the open web. I've said before I'd gladly pay for it. With real, hard, proper money.
My biggest concern is the unhealthy dependence on Google. I want a successful Mozilla precisely because I value the free web, not the contorted surveillance monster that Google, Facebook et al seem intent on.
So in the absence of convincing rationale, I assume announcements like this to be nefarious intervention at Google's behest. Return on that $300M "contribution".
Please, Mozilla. Don't give in to the dark side. There is a palpable increase in privacy awareness - you can be at the nexus of growing it, and grow yourself accordingly. Or you can side with the Emperor. Who will crush you once you've served your purpose.
OK, maybe that's dubiously allegorical. Here's the brutal commercial interpretation. The _only_ way for you to reverse your declining market share is to have a differentiator. People need a reason to use Firefox. You can choose that path. Google isn't "supporting" you - it's adopted the "embrace, extend, extinguish" mantra so beloved of 90s/00s era Microsoft. To be clear: you're well into stage 3.
Seems to me you have two choices. Find a differentiator and adopt a growth business plan for independence. Or remain at best Google's sop against anti-competitive accusations, and be fully extinguished when you're not needed any more.
Look back at the initial success when they were up against Internet Explorer 6. They offered users what Microsoft wouldn't. They didn't hamstring themselves with IE6 compatibility, and instead broke the rules of what a browser was supposed to be by only considering what was best for the user, and spurning harmful standards and technologies.
Where Microsoft hid extensions away in the registry, and their extensions had degenerated to the point where most of them were malicious, Firefox did the revolutionary thing of offering direct access to them by users. There's a lot wrong with Chrome, like their draconian and invasive updater that sometimes damages people's computers and can not be easily turned off.[1] Or how Chrome's extensions, while safer, are also crippled and prone to abuse by their forced updates.
Mozilla could easily leverage Google's weaknesses against them, as they did with Microsoft. Mozilla is not a for-profit platform, they can afford to make the things Google can not. They can have a powerful community driven extension architecture that empowers users to do things Google never would. So of course instead of improving and securing their old extension architecture, or replacing it with something just as powerful, they copied Chrome again.
It may be hard to believe, but including a built in popup blocker was revolutionary. It was called "irresponsible" by advertisers, and forced everyone else to change. Could you ever imagine Mozilla undermining the big players like that today? Even though bold moves like that are what got them their success in the first place. It wasn't branding or PR or any of that crap!
I think the only good thing they've done is containers. And containers are such a good thing for users that I am convinced that many in Mozilla were opposed to it and fought against it, and that it will be deprecated with a flimsy UX metrics justification. And toadies on HN will try to justify it because "why are you thinking about this so much? This is fine, stop complaining, jeez."
One of the smartest things Mozilla used to do was not try to implement bad technologies. They didn't try to re-create the quirks of the Trident engine; they didn't let Micosoft dictate the standards, so Javascript could not eject the CD tray, and you could not change the color of the scroll bars! They didn't waste effort trying to copy Microsoft's Windows-first proprietary ActiveX model. A much better compromise was the community created IETab extension, which just ran an instance of Internet Explorer in one of the tabs for when you needed it.[3] Could you imagine the Mozilla of today saying that if you wanted to watch DRM'd Netflix, you should just use Chrome? Is Firefox even technically capable of having something like Chrome tab anymore? Think of the resources that would be freed up if they stopped trying to have feature parity with Microsoft/Apple/Google.
They just need to have the courage to say "that technology is evil, we won't support it. Maybe someone will make an extension to support it". Instead, essential technologies like RSS are deprecated and left to the community to implement, while DRM and proprietary and problematic standards are given first class support.
Mozilla has made all of the wrong moves in the last 10 years. Their priorities are completely misguided. And I am certain that if they just centered the user over the big platforms, they would experience more popularity than a billion dollars in marketing and PR.
It may be hard to believe, but including a built in popup blocker was revolutionary. It was called "irresponsible" by advertisers, and forced everyone else to change. Could you ever imagine Mozilla undermining the big players like that today?
That is not at all the same as when they first did it. When they did it, popups (that is, ads that would open a new browser window) were a staple of online advertizing, and they disrupted it.
Google is also doing the same thing, including punishing websites that use abusive interstitials and popups[1].
It would not be all that disruptive at all, in fact it is actually helping Google and the big platforms by forcing ads to be more palatable.
I doubt that any of Mozilla's underwriters have a problem with this.
Something that would be disruptive would be if Firefox bundled uBlock by default, which Google is trying to prevent users of Chrome from using[2]. If Google wanted to, they have the means to remove and blacklist the extension from most users computers immediately, thanks to their updating mechanism. They could even push changes to Chrome that break the extension or any like it, immediately. Frankly I think the only thing that acts as a counter balance is Firefox, but they making themselves less relevant by the day.
Thank you. It's a relief to see someone else finally saying the same things I have ever since Australis was teased. We need more clear voices saying this.
Reading back through their mailing lists from 12 years ago is so telling.
So many responses to users who were complaining that they wanted features from Internet Explorer were met with "then use Internet Explorer, we but we refuse to embrace their vision of the web".
And like I said in prev post, extensions like IE Tab would let users appropriate proprietary technology without embracing it. And the extension architecture was powerful enough to do stuff like that. They should have tried doing something like that with Chrome, while keeping their technology.
In 2009, I would watch Hulu on Linux using instructions I found on a forum, which used a python scraping script and fed it to my Firefox VLC plugin. It was the first solution that the community came to for watching Hulu. That was the culture back then.
So, I can see why the big platforms hated it so much. Such a powerful extension architecture and the culture around it had big potential to undermine their platforms.
I really miss the old Firefox. I miss how I could quickly cut through mountains of pages with hierarchical vertical tabs, and export or mirror groups of urls in bulk. And it was faster, because I had a lot more power to restrict what content was loaded.
There were so many powerful extensions that gave me direct access to the content I wanted, which are impossible to use now.
Yes, it had security problems, but those are solvable. Just on my own, I cobbled together some fairly good sandboxing systems. I wish they had just worked to improve XUL.
It's worth noting that Tree Style Tabs still exists and works really well. You can multi-select tabs and move them around or bookmark them, too.
Firefox is dramatically faster now, and you still have full abilities to restrict content loading with WebExtensions like uBlock Origin and NoScript. What functionality relating to restricting content loading are you missing?
Excuse me, but post-e10s Firefox is not anywhere near as fast as old Firefox. It's a slow mess that freezes constantly, and the switch from XUL to WebExtensions (and the changes to the rendering engine resultant from that) have made it agonizing to even load basic HTML without CSS, ECMAScript, or XML content.
I switched to Pale Moon for performance and philosophy, and because it has worked since then exactly as I need it and want it to, and has been uniformly faster (now I can confirm on all three machines I use - the x64s at work and x86 at home), and it has given me no security concerns, I've made a 100% switch.
> "It's worth noting that Tree Style Tabs still exists and works really well."
YMMV. I use it but I find it buggy and slow. Once or twice a week it seems to crash completely and although it comes back up by itself, it takes a minute or so with large numbers of tabs. In the past, although not recently, I've also had trouble with it flattening my tab hierarchy when firefox is restarted, and having tabs at the wrong indentation level or no longer displayed under their parents.
Also, you can't have the bookmarks sidebar and the TST sidebar open at the same time anymore, which is a real bummer.
I tried it recently and it was buggier and slower than in 2011. It's pretty heartbreaking for me, I don't want to do it.
I heard that it wasn't going to be possible to do tree style tabs, but Mozilla hacked it enough so that it would work, and convinced the original author to re-implement it. It feels like a hack.
Maybe, but I don't think so. It tends to happen when I'm interacting with it, often dragging subtrees around. It might be coincidental, but it sure seems to me like there are lots of unstable edgecases.
> Could you ever imagine Mozilla undermining the big players like that today?
Enhanced tracking protection[1] was turned on by default for all users recently. This is a feature that hits at Google and Facebook's bottom line.. so I guess yeah, I can imagine Mozilla undermining the big players today. Not to mention other privacy initiatives like Containers/Facebook Container.
> "Not to mention other privacy initiatives like Containers/Facebook Container.
Except the Multi-Account Containers extension has a trash UX[0], and there is no Google Container extension from Mozilla. Why does Mozilla give us a first party Facebook Container extension but force people to rely on a community provided Google Container extension?
Am I supposed to give Mozilla the benefit of the doubt and assume that the lack of Google Container extension has nothing to do with Mozilla getting more funding from Google than they do Facebook? And that Mozilla's failure to throw a few engineers at fixing up Multi-Account Containers has nothing to do with such extensions earning them no good will from Google/etc?
[0] I realize calling something trash might be considered inflammatory, so let me defend that remark. The number of steps needed to say "always open hacker news in a container called 'hacker news'" is absurd:
1) Click the Multi-Account Containers button.
2) Click the "Create new container" button.
3) Enter a name, select from a (senselessly limited and visually similar) number of colors and icons, and press Ok.
4) Click the Multi-Account Containers button a second time, then click your new container to switch to it.
5) In the new container tab, navigate to news.ycombinator.com
6) Click the Multi-Account Containers button a third time.
7) Check the "Always open in [container]" box
8) Open a new regular tab.
9) Navigate to news.ycombinator.com a second time.
10) Check "Remember my decision for this site"
11) Click "Open in [container] Container"
The redundancy in steps 7 through 11 is particularly frustrating.
Maybe I'm reading this wrong, but it seems like Mozilla plans to remove custom search engines from Firefox, except through WebExtensions. So you won't be able to go to a site and easily install a search engine for it. The site author has to build a web extension and submit it to the firefox addon site.
Yeah, this is crazy. Chrome's implementation of this (which allows you to type the beginning of a website's domain and then plus tab to use that site's search functionality) is already one of the things I miss most when I use Firefox. They should be improving this, not removing it.
I switched from Chrome to Firefox literally this week, and I am missing this so badly. Can't understand why they wouldn't copy this behavior in all these years. I hope they put it on their agenda now.
I was preparing to create my top 25 OpenSearch plugins manually (less popular plugins at addons.mozilla.org often lack autocomplete and have other quality issues). Glad I read this before wasting 2 hours of my life.
Now I'm just switching back to Chrome. Does Mozilla actually expect site owners to create search extensions for ~5% Firefox users (of which in turn ~5% will actively install such an extension)? Firefox performance has come a long way, but Chrome is still way more polished. Apparently, Chrome is at the same time more poweruser-friendly, which really irritates me.
No need, firefox supports custom search engines, but it a) doesn't store them automatically, you have to manually add them and b) doesn't have as nice a UI for them as chrome does. But it works. Why it isn't implemented better is beyond me, too. See https://news.ycombinator.com/item?id=21292283 for instructions.
The article is about search engine addons (which approximately nobody is using), I don't see how this would relate to search keywords? I don't think those are going away any time soon.
As we're approaching the end of the decade, for posteriority we should put up a site with 2000's and 1990's web features that we took for granted (such as RSS, also used by OpenSearch result listings) but don't have anymore in 2010's web of shit.
Maybe not completely dead yet, but each time any webpage gets a redesign these days, they usually kill RSS feeds. It's more or less a niche "old web" tech now.
The fact that it doesn't have "open" in its name doesn't actually make it any less open or interoperable. This is about standardizing on having a single and modern way to define search engines, nothing more and nothing less.
That seems like a very narrow and reductionist view. Sure you still can create search plugins for both FF and Chrome. But that isn't the (entire) point. OpenSearch is a well-known, proven, battle-tested open standard... exactly the kind of thing we need the Open Web to be built around. And doing anything that impedes whatever momentum OpenSearch has, seems like a loss to me in a very general sense.
I lump this in with the kinds of decisions like removing RSS support from browsers, browsers (hello Chrome) not supporting MathML[1], and a zillion other "death by a thousand cuts" things that impede the Open Web.
[1]: Yes, I know that there is an initiative underway to re-add MathML support to Chrome. I'll remain somewhat skeptical until it is actually shipping and works.
There are two separate gatekeepers in AMO and the CWS. On the other hand, OpenSearch is used for the "Add search for this field" option which just needs some metadata defined on the web page.
If they were just removing OpenSearch in xpi on AMO, that'd be fine, but the post says they intend to remove OpenSearch as a whole.
> Both Firefox and Chrome support a common format for defining search engines from add-ons
...which have nothing to do with the Open Web, mentioned in the parent reply. Both require participating in, and consequently adhering to the terms of, stores operated by third parties.
Simple enough and it seems OpenSearch will still work embedded on websites for a while. But it's unclear how they plan to get website owners to migrate.
> But it's unclear how they plan to get website owners to migrate.
And it's unclear why they should have to. OpenSearch is a standard that all modern browsers currently support. Why should a site author have to support OpenSearch for every browser and build/maintain a WebExtension just for Firefox?
Looks like Mozilla may also be planning to kill the search bar. With this change, how can anything be added to the search bar itself?
As usual, Mozilla just follows Google lead rather than innovating browser features anymore.
How does mandated signed addons get you "offers users more controls for opting into changes"? They offer me less controls, as someone else (AMO) controls what I may opt into.
Well, that blows. I know a good number of the search engines I use have shortcuts in DuckDuckGo, but I felt pretty clever when I discovered I could configure Firefox to search Wikipedia/Amazon/IMDB directly rather than going through DDG intermediary. It also avoided clogging my address bar suggestions with:
This is nuts. Mozilla should at least allow the user to set a custom search engine using an about:config option, like the old keyword.URL [1] option, which was removed after Firefox 23. I use DuckDuckGo with several customizations as POST parameters and it would be very cumbersome if I have to write an extension to use DDG the way I want it.
Mozilla is basically the fig leaf covering Google's hegemony, they get all their money from Google, they use Google to track Firefox users ( https://news.ycombinator.com/item?id=14753546 ).
I'm pretty sure it's intentional at this point. You don't go shooting yourself in the foot, reload and do it again consistently for years without intent.
I wonder if it's possible to write one extension using WebExtensions, which does a better job supporting OpenSearch? I know that WebExtensions is limited, but all one would need to do is read the OpenSearch data, store it somewhere for later recall, and open up some path in the ui to search on a site that we stored the info for. And, of course, launch a tab with the results in it.
I hope this is as easy as someone writing an extension that restores the functionality. Maybe this can spur some innovation for Firefox users via extensions.
Extensions can add address bar suggestions, but not a new search bar. It's unclear if Mozilla intends to remove that as well, I haven't seen any webextensions that add a search provider to it.
There are several extensions that use userscripts, like Greasemonkey and Tampermoney, so loading the search manifest files shouldn't be a problem.
Well, this is a great reason for Firefox users to remove the Google search engine from their browser and use DuckDuckGo. Better results and the bang syntax!
If I'm reading that right, they're deprecating support for discoverable browser-independant markup for searches; and replacing it with the requirement that each site actively develop (and maintain!) a software plugin for every browser their users might want to use.
The whole point of a "user agent" was to go out and do things for me on the web; and the idealistic goal was that each person could choose an agent suited for them, which then had tools to programmatically discover and interact with the web in a common manner (reducing engineering load on the webdevs).
And I don't want to try and use a separate search tool (with new flashing graphics and ads!!!!) for every site I go to... I want a single search tool, like FF offers right now. (Aside: not to mention chrome's "auto-discovery of opensearch when you tab after typing a domain" is actually MORE useful than FF's manual mode!).
Taking a step back and removing support for a declarative api seems to me like the really wrong direction for an open web. Instead of sites supporting a single declarative browser-independent markup; they now have to deal with a long tail of (2-3 + who knows how many) browsers; and users with a niche browser have to spend effort convincing every site to support their browser.
Why not try to improve the opensearch markup instead?