The Song Fight Bulletin Board

http://tools.google.com/chrome/intl/en-US/welcome.html

Interesting...FAST!!!

I'm definitely interested in it. Firefox 3 is merely okay but while it has a lot of neat new features over Firefox 2 it feels like a huge step backwards in terms of performance. XUL is a nice theoretical idea but in practice it just makes every app it touches ridiculously slow. So having a nice *native-widget* browser based on Gecko and a next-gen JS engine is definitely a good thing.

I hope the Mac version comes out soon since it's the Mac version of Firefox which is particularly bad (sometimes when I type fast I have to wait a good 30 seconds for it to catch up to me).

I think it's based on WebKit, not Gecko, isn't it?

http://blogoscoped.com/google-chrome/12

Oh, right, brainfart there.

Anyway, I just tried it for a few minutes, and ended up switching back to Firefox. I really hate apps which decide that their own stupid look and feel should override the rest of the system, especially since the color scheme they use is very clashy with my color scheme. Plus, there's an annoying lack of user experience customization, and their stupid custom menu widget makes a lot of things I do a lot very inconvenient. Even Safari for Windows is a better experience (it uses its own custom skin but at least it uses a fairly neutral color scheme which does an okay job of integrating, even if it's still not the most attractive thing).

Also, lack of scrollwheel speed customization, and no way to make it default to sans-serif without overriding serif, are other strikes against it.

I'm sure it'll get better in the long run but for now I'd rather put up with Firefox's slight crappiness instead of having a whole new set of crappiness to get used to.

All the stuff about process/memory management sounded nice. I think they were pretty well describing issues I've had with Firefox (memory bloat, lack of isolation of threads/processes, etc.). Hopefully Firefox will adopt similar strategies for those things.

Yeah, although that sort of change is extremely invasive and I don't think it's the sort of thing which will happen within FF3.

Ideally, Google will just get a lot of complaints about the custom chrome (no pun intended, blame them for overloading a term yet again) and decide to just use native widgets like they should have done to begin with.

I agree about the window frame design. While "clean" it is clashy and odd, but I do like getting rid of all those "default" menus that really don't have to do with browsing. You can also set a sans-serif font for serif, thus over-riding, but that's still a weird thing.

The incognito mode is cool...especially on corporate machines.

Yeah, I know you can override serif with sans-serif but that's not The Right Way To Do It, and could result in a serif-designed site getting a sans-serif font if it uses font classes instead of specific font names.

Yeah, the first thing I tried to do was change the colour scheme. I also noticed you can't set a limit on temporary files. The single bar for search/address works really well though, and the whole thing runs in almost fullscreen which is super nice. I found the new tab with thumbs of recent sites (that I didn't have to wait for) to be great. I didn't think I would like it, but I use it all the time. As for the fonts, well... I almost peed when I read how upset you guys are about the sans serif thing. I seems equivalent to holding weekly meetings about the strategy of letting Greedo shoot first. I know I'm a bit of a hillbilly, but damn... there is text, and you can read what it says.

Serif fonts aren't as easy on the eyes as sans-serif and are very hard to read at small sizes. Sans-serif fonts look ugly as headlines. Fonts and design are important.

Also, I use the browser alongside a lot of other things, since at work I use a browser as part of an environment involving many apps. I don't care how good it looks maximized, if it's a pain to deal with when I have it side-by-side with my source code editor.

And Han shot first.

Wouldn't standard widgets preclude the "tabs on top" idea? Seems like they had to basically code it all from scratch to pull that off.

I'm going to give them a shot. I use like 8 different computers these days (I include one of those as the Parallels installation on my work macbook), so having Chrome as one of my browsers isn't hard.

They'll have a chance to fix the things that real people don't like, now that real people are getting the chance to use the browser. Also, some of the behavior (specifically the "do you want to save this password" dropdown bar) looks a lot like Firefox, so I expect to see a lot more customizability coming, taking advantage of FF/Mozilla components.

Plus, their new Javascript engine would make the Javascript components of XUL-based extensions that much faster, and the process separation would help keep poorly-coded extensions from crashing or hanging the browser.

Did you see that Mozilla releases a little chart comparing (favorably, of course) their next-gen Javascript engine to V8?

Sure, the custom chrome was necessary for the top-mounted tabs, but I don't think the top-mounted tabs are even a very good UI to begin with. It's not just a matter of being unfamiliar, but a matter of them basically trying to throw out 30-odd years of incremental UI improvements in order to focus on one aspect which has gained a lot of favor in the last few years (but which could probably be done better).

My friend Jens (who works for Google but not on Chrome) has some pretty good thoughts about what's wrong with the UI.

The search and address functionality in Firefox right now are tab-specific, but they don't look that way. It's never clear whether the address will change if you type something, switch tabs, switch back, enter a search term and search on one tab, will what you typed go away if you go to another tab? All that sort of deal.

I think the TabsOnTop are great, and I disagree that 30 years of history is a reason to keep anything. In fact, that amount of time in my business (which is product managing software development) basically means the thing is old and tired and full of cruft and is due for a reimagining.

TabsOnTop makes it very clear that what you're doing right now applies to this tab, and you know that if you switch tabs you'll be starting fresh and if you come back what you've done will still be there. I think separation in this way is one of the fundamental ideas of Chrome, and I like it a lot. There are some UI bugs, and I too want to change the colors, and the animations aren't as smooth as they will be. But the primary innovations stand up well for me. Especially the "Create application shortcut" thing, which I think is definitely going to make its way into Firefox, and with its integration of Gears is going to eventually change our concept of how one uses a computer.

I have recently been thinking about what it would be like if you made your browser into your only interface with your computer. Using only web apps, storing files only online or in places accessible to the browser, etc. Chrome is a big step towards that idea, and so far I think I like it a lot.

I like Firefox too though, don't get me wrong. In my opinion, some of the immediate reactions to Chrome are a bit more knee-jerk than is warranted, including your "I gave it a shot then went back to Firefox" response, Fluffy. It's a piece of software. And it's still beta right? But that's the Internet for you. Beta isn't what it used to be, and Google knows that. If they didn't anticipate some backlash I'll eat my hat. Which is made of chocolate, of course, because you never know when you're going to promise to eat your hat if something isn't true which turns out to be.

In the end I think if you're happy with Firefox or IE then Chrome isn't going to make you switch until you encounter something you really want to do but which doesn't work well in your preferred browser-- and then you have to find out, somehow, that it works well in Chrome. Some people automatically try several browsers to find the one that does X best. For example, Chrome wouldn't play the Java applet version of Bookworm, so I'll still have Firefox open all day long.

I just read your pal Jens's blog post, and wish to note that he does not say that the tabs-on-top are bad. Also, he notes that he hasn't tried Chrome, so I don't take his comments to be directed necessarily at the specifics of the application (I have to say, I thought you were implying that he did, Fluff, and felt like that was a little disingenuous, although I'll give you the benefit of the doubt that you weren't).

That said, I agree with him almost completely. One blatantly missing piece of the Web app pie is UI and filesystem standardization. There's no consistent "where's my stuff stored and how do I get to it" scheme across the web. Each app displays the means to retrieve and save items in differently. That is one area where I'm not sure how you'd improve on the decades-old "Menu Bar" system, although there's probably a way-- but then how do you standardize on that? You need a behemoth to create a killer app that makes everybody follow suit.

So far, none of the existing behemoths have noticed the need, OR, they are desperately clinging to their out-of-browser mechanisms because that is where the money lies at the moment. We need a customizable, open-source Google Filesystem UI that everybody can adopt. Note that it's not a standardized filesystem-- it's the front end, the experience, that needs some standardization.

JB

As a software engineer I can definitely understand the impulse to want to recreate everything to make it Perfect and there are certainly things which could be improved about a tabbed interface and so on, but there are better ways they could have done it. And yes, I realize it's beta and will also be continually improved and so on, and I do expect it to get better over time. It's just that RIGHT NOW, it's not a very useful browser FOR ME, because I actually do use other apps and I like being able to see my dozens of windows easily and have a browser which plays nice along with all of them. I'd have definitely preferred it if they had started out by making a good solid browser (reinventing the scripting/process model) with an integrated UI rather than trying to reinvent two things at once. Scientific method, etc.

Also, for casual home users, web apps can certainly do a lot of stuff, BUT it's also a bit silly to have most of the processing happen on some server somewhere while everyone's CPUs just sit more or less idle. Web apps have a much harder time scaling to a large userbase than desktop apps, and if it breaks for one person it's likely broken for everyone. Also it very severely constrains what people can do with software (for example, if a company goes out of business, suddenly everyone who was relying on their software is SOL, while with traditional desktop apps you probably have installation media and so on).

I'm all in favor of cloud storage and the like - I love web-based RSS reading and using IMAP for my email storage and WebDAV for important documents, don't get me wrong. But moving to a purely web-based model is only suitable for a small subset of what computers are useful for. Sure, it's great for grandma, but I'd just love to see someone try to do a proper multitrack recording environment as a web app. Even things like photo management really need a desktop component for getting photos off the camera in a reasonable manner.

For purely network-oriented apps, Chrome is fine, BUT I'd much rather see those network-oriented apps play nicer with the system that's already there, rather than try to be in its own little virtual environment isolated from everything and where even basic drag-and-drop no longer works right. I think richer standards which allow for good user-centric interoperability between the two (which is what Jens was getting at when discussing things like file management and the like) would be very useful.

I just read your pal Jens's blog post, and wish to note that he does not say that the tabs-on-top are bad. Also, he notes that he hasn't tried Chrome, so I don't take his comments to be directed necessarily at the specifics of the application (I have to say, I thought you were implying that he did, Fluff, and felt like that was a little disingenuous, although I'll give you the benefit of the doubt that you weren't).

I didn't mean to imply that, I was just speaking in my usual generalized imprecise way which ruffles everyone's feathers. I'm surprised you aren't used to that by now.

And yeah, he didn't take an issue with the tabs-on-top, but he did take an issue with the lack of a real menu bar, which is what I was getting at with the throwing-out of 30 years of refinement. Tabs-on-top could be done well, but Google didn't do it well, is what I'm saying. There's a lot going on here and I'm not about to dissect it down to a per-pixel level or whatever.

fluffy, you might enjoy reading some of the essays of Paul Graham for a different perspective on software development.

Basically he espouses the idea that almost all good software is written by small groups (like 2-3 people) in the context of start-up companies, and his ideas about the best way to go about it match up pretty closely with how Google Chrome is proceeding. (One major philosophical point: start with the smallest possible thing that works, give it to users, and go from there.) He's had some success to back up his crazy ideas, too, so it's not too easy to just dismiss everything he says.

For example, you worry that a web-based application means that a bug affects everyone, all at once, and that this is a major problem. Graham tells stories about how that isn't a problem at all, in practice, because such bugs can be fixed, for everyone, often in a few hours or less. There were times in the development of what became Yahoo Store that he was on the telephone with a customer complaining about a bug, and he'd fix it while he was talking to them! "Try it now," he'd say, and, of course, everything just worked.


Also I love recommending Graham's book "On Lisp," which, more than any one other work, changed the way I think about computer programming. The full text is available free on his website!

I am well aware of Paul Graham's writings. They made sense in the mid-90s, and sure, some of what he says is still true when it comes to simple web apps, but simple web apps are limited in their utility. We're at the point that we've got all sorts of fun simple web apps out there which do all sorts of simple things well, but each one has its own login and storage and is in its own little world.

There are plenty of models which can leverage the best of both worlds, such as the Sparkle framework for OSX which allows apps to easily see if there's updates and then gets an update if so, and personally I'd like to see software's status quo return to where things were actually tested before they were pushed out. I'd rather have my apps work and have their defects managed than to have to worry about what fun and exciting adventure there will be in some major functionality being completely broken every time I run it (or, in the case of web apps, defects showing up WHILE I'm running it, or the server going down and losing all my work, or the storage cloud being subject to a DDoS, or whatever).

Oh, sorry to hear that. Carry on then.

fluffy wrote:Also, for casual home users, web apps can certainly do a lot of stuff, BUT it's also a bit silly to have most of the processing happen on some server somewhere while everyone's CPUs just sit more or less idle. Web apps have a much harder time scaling to a large userbase than desktop apps, and if it breaks for one person it's likely broken for everyone. Also it very severely constrains what people can do with software (for example, if a company goes out of business, suddenly everyone who was relying on their software is SOL, while with traditional desktop apps you probably have installation media and so on).

I'm gonna postulate on that idea that my computer will be underutilized in a world of only Web apps. I disagree, and here's my thinking making use of the idea of a Web-app digital audio workstation (DAW), say, a future version of Reaper, as that's a pretty indie app.

The guts of the DAW get downloaded. Let me get my new fancy dual-core Vaio laptop with the integrated 802.11g wifi and see how long it takes to download. About 10 seconds on Comcast cable, I swear to god. I haven't had a computer that could do 802.11g before this-- my desktop PC has a netgear 802.11b adapter, and my other computer is an ibook G4. When I started downloading things on this laptop I was incredulous, it was so fast. So, let's think forward a little to a hypothetical middle of 2009.

In 2009 speeds will get more consistent and the superfast fiber will get ever nearer to being reality for what we could call a "lot" of people. That 10 seconds will get shorter and shorter-- 10 seconds for 3.3MB, and I will only have to grab it once in a while when there's an update. Meanwhile, put the UI for the DAW in the browser, and swap stuff around whenever you feel like you have something new. Modularize the guts of the DAW and you can update core functionality without even downloading that 3.3MB. Put plugins on other sites that integrate into the DAW's Gears app. Small plugins could appear or disappear like extensions to Firefox.

Process separation in the browser allows me to run the DAW and the Word processor at the same time. New and improved Javascript interpreters (V8 boils it down to binary!) increase the sophistication possible in the "views" that live in the browser.

I guess the long and short of it is that I think the app-in-browser days are closer than you think, and Chrome is a big step in that direction.

Now, as for the whole UI thing and standardization in-browser, I think what that will take is an open-source project that does the following:

1. Creates the UI model for the standardized method of saving and retrieving files in a browser-based application.
2. Creates modules that implement the standard for every possible framework, including those for Javascript, ASP, PHP, and Python. Javascript to handle the front end, and the others to manage the underneath.

The thing basically needs to share as much code/theory as possible and then be ubiquitously implemented FOR developers-- that way they'll actually use it. I'm sure Google would be glad to host it.

JB

Grabbed this earlier today, and I'm getting used to it. I miss a few plugins (imagezoom, mostly), and they're obviously far from a full release, but it's a pretty solid product so far.

Bonus points for text-entry fields being resizeable. Resizable. Resize-able. wtf.

That's actually a WebKit feature, and Safari has supported it for a while as well.

Silly things:

- how to get back to my "new tab" once on a page, without the back button, or opening a new page then closing the old one.

- when I enter an address like "roymond.com" it often doesn't know quite what to do with it (the combined search/URL field is confused). What's next? Ask me to put "http://www." in front of web addresses?

I like it a lot otherwise but I guess I do things different. Needs work before I spend too much time in it.