Evan Czaplicki's talk did make a great observation about how easily we compare small open source projects with multi-million-dollar corporate projects with a straight face.
Though we treat small projects far worse because we don't even have the expectation that a large corporate project like Swift gives us individual attention, yet we entitle ourselves to the attention of a free open source volunteer project and then make a loud stink when we don't get it.
> make a kick-ass browser that people love, strive for even 0.5-1% of browser market share, and then sell your default search preference.
Well this seems like the hard part. I think it is actually not easy to get 1% market share, because if you are good enough to hit 1% you are likely to overshoot 1% by a lot, and if you aren’t good enough to make it past 1% you probably won’t make it to 1%. And obviously creating a browser that is good enough to overshoot 1% by a lot is very difficult.
The companies who actually have 1% of a market (for a sustained period of time) are like really weird edge cases.
It's only 1% on average. Instead you name a product that has 50%-90% market share among certain demographics, a narrow sliver of the general audience. You solve their particular problem well enough. It's just not everyone's problem.
For selling the search preference it works even better because it instantly adds a bit of targeting information.
Opera has been floating around 2.5% for a long time, but to be fair that is not so much 2.5% across the board as a higher percent in a specific demographic.
Safari can demand such a large price due to the value of their traffic. Size and demographic factors of mobile Safari users really dominates here.
Firefox would probably be worth less than 1/10th of the $400m stated here if Google wasn't constantly looking for ways to 'support' them. It is an open secret that Google actively makes decisions affecting Firefox based on how they believe regulators will view those decisions re: Chrome monopolistic behavior. I have heard direct first-hand accounts of meetings specifically about how to make sure Firefox remains somewhat relevant in the market to counter any regulators that may come knocking. The cash injection is one major factor to make sure they stick around as an 'alternative'
> make a kick-ass browser that people love, strive for even 0.5-1% of browser market share, and then sell your default search preference
Isn't this what Opera (3% marketshare) and Vivaldi (unknown% since they use Chrome's user-agent) are doing? And yet to make a browser that's good enough to capture this "small" chunk of market share, Opera needs >600 employees, and Vivaldi >50.
What are we, as a civilization, losing when we let these surveillance companies dictate how technology is used? Increasingly every aspect of our digital lives is being controlled by the need to sell ads.
Ah, the joy of getting search results instead of an error when you mistype a domain name. Or better yet, being forced to type http:// when you want to navigate to a custom hostname. This combination address bar and search input must be stupidest UI element ever made. Thankfully at least Vivaldi allows turning this behavior off.
Personally I love it. Yes, it has the edge cases you describe but they are easily managed. More often I have the opposite problem when I want to search for a tech that has a dot in its name.
I worked on the first implementation of this at Netscape many years ago. The intent was to move away from URL as the primary location specifiers and more towards a semantic web interface. Obviously this didn’t evolve beyond that first step of directing to search, and the semantic web effectively never happened. But the original syntax was if it was a single contiguous string without whitepace it was always resolved as a URL, unless prefixed with a ?. (Thus no need to struggle with hosts being searched etch.
The intent was there would be other operators and ways of specifying some sort of resolving agent. It was a pretty interesting idea but Microsoft destroyed internet technology competition for the next 10 years and the rest was lost as the next time someone picked up basic web technology it was a search engine.
That was sensible. Actually, I do search from the address bar, but I have it set up such that there's no default search engine. I have to prefix my query with "g" for Google, "gi" for Google images, "w" for Wikipedia, "y" for Yandex, and so on.
On the semantic web thing, in the end, it turned out URLs are useful for everyone — at the very least because a URL is something you can share with other people and be sure they'd reliably end up seeing the exact resource you want to direct them to. Same can not be said about anything that involves web searching. Search results move around all the time for a million different reasons.
It was a younger simpler time, with more dreams and less dollars involved. I don’t think it should have evolved from what I did in retrospect. Search + URLs is entirely sufficient and I don’t think anything else would have ever emerged even without Microsoft.
another brain-dead 'feature' of all this is automatic URL parsing of search queries with a leading/trailing slash or colon.
it's a victim of "all the ways for all the people", and it can't do either very well.
The optimal middle ground for me was the separation between search query bar and URL bar.
And why did we combine the two? Well, it was to reduce screen clutter and combine features so that we could make room for....
making the URL bar longer, generally far longer than any sensibly human-readable URL. Brilliant!
(the REAL reason was because Chrome wanted to shovel people into Google Search as promptly as possible, and that started a Stupid Browser Trend that firefox got suckered into following.)
The one I’ve seen lately is when I type a correct domain, but the autocomplete appends “ login” so I get taken to a search result with links that are not what I wanted!
Who was the one with the original idea to make the url bar double as a search bar (I mean person, not company - i assume the first company to do it was google?)
The first time I saw it was in Google Chrome way back when Firefox still had more traction. Back then both Firefox and Internet Explorer had separate url and search fields.
Pretty sure it was Nicholas Jitkoff[1] and his team. He had done Quicksilver[2] a few years back which popularized this single input interface for osx desktop.
started with ISP selling your mistyped dns queries to google or whichever search engine they could sign a deal.
from there there was a race to capture as much mistyped domains until google had the end-game idea to pay Firefox to send non-standards domains to google directly, not even wait for it to fail dns resolution.
when later google launched Chrome, initially as a cost saving for the firefox deal, they dialed the dns to search hack so aggressively that they broke the internet (search for chrome type ahead incident).
all search engines (called portals at some point in time) goal was to be the dns replacement. AOL had "keywords" which they advertised heavily in their tv channels like TNT ("we will be back with Terminator. see more at Terminator keyword on AOL")
> All of this really drove home that the browser URL bar doubling as a search input has to be the single largest money-printing UI element ever made.
This is mistaken; Google has been paying to be the default in Firefox since long before Chrome popularized the idea of merging the URL bar and search bar (which, thankfully, you can still turn off in Firefox; I'd prefer not to render my URL bar useless for searching my history).
By putting the search into the URL, Chrome is able to track every page you go to (since the URL is perhaps a search term) and how long you stay there, and what page you go to from each page you visit.
With search suggestions on (the default), they get that even for FF users.
Not only that, but every GET argument on every page is sent to Google as well, since they are perhaps search terms - so as you type your content is being sent; You can imagine the value even BEFORE getting to the fact that it generates more searches.
Which is a shame, since I used to use the URL bar to search my browsing history and the Search bar to search the net, and the whole UI was so much improved, even ignoring all the times I accidentally am sent to a search results page I didn't want.
This doesn't sound correct. Everything you type or paste into the box goes to Google but when you click on links the browser knows that's a URL and doesn't search for it.
On Chrome, when clicking a link, the URL is copied into the search/url bar, sent to Google, and only then does the browser follow through to the URL. You can see it in Wireshark.
I am not sure about FF and have no time to test it now.
A setting to completely disable support for media elements per-origin would be nice. Not this "we try to determine whether a video is eligible for autoplaying" bullshit. I want an "I wasn't asking" approach of the browser literally treating <video> and <audio> as unknown tags when this setting is off.
Native support for Flash via Ruffle would also be nice.
If it's a mobile browser, I really, REALLY want a setting to just completely annihilate all the PWA stuff. No, I don't want to add this random news website to my home screen, thank you very much.
More broadly, I want most of the "progress" of the web platform undone. Sure, new CSS features, like flexbox and grid, are nice. But all those new JS APIs that (try to) turn a hypertext document viewer into a (terrible) operating system? No thanks. I want my clear boundary between the "document" and the "application" back, hence the Flash thing.
Something less annoying to opt into web features would really be an improvement. Perhaps some "upgrade icon" next to that https lock icon? This already is the place were we take away the permission when we accidentally allowed some we did not want, and the "upgrade" would be somewhat parseable as all three of "upgrade to local(ish) installation", "upgrade to more site features" (push) and "allow more data upload" (location). All implicitly connected to the hostname.
Mouse gestures, opera had these back in the day. The right-left click to go back, ability to close a tab from anywhere on the page with just a single hand, etc.
Maybe it's no longer possible due to all the click hijacking introduced by web2.
Though we treat small projects far worse because we don't even have the expectation that a large corporate project like Swift gives us individual attention, yet we entitle ourselves to the attention of a free open source volunteer project and then make a loud stink when we don't get it.