What would you like to see in a web browser?

Yeah, I just started programming a Web Browser with the goal of making it as useful as possible, I added features I knew I could use and now I was wondering what you guys would want in a web browser. The web documents currently render using IE but I hope to switch it to render with Mozilla Gecko soon.

Current features -
- Basic Navigation Tools
- "Checkpoint" Feature

* Basically, I call it Checkpoint because: when you go to a webpage you can flag it and then return to it whenever you wish without bookmarking. For example, I am looking at the Cacoawards on Doomworld but I want to navigate to the downloads page without losing where I am, I would flag my current page, then go to download the WAD or whatever I wanted to do, and then just click the flag again to return to the exact page I was originally on. Think of it as temporary bookmarking.

Programming languages used : C# and VB.NET

One thing that would be useful is the ability to have a console quick search. an example of what I mean:

yt michael jackson

and then I would get a youtube search for MJ

so I can simply type 'yt' (youtube) into the address bar, then my search terms, and find results. Even better is make this customizable that way I can set my own search engines with their own 'identifiers.' Possible commands could be g - google, w - wikipedia, dw - doomworld (lol), a - ask.com, etc....

Perhaps you're asking a bit too much, as every one of these search engines or services works with a different parser, so either the app's programmer will have to do the parsing himself, or create a powerful enough user-configurable parser that allows you to take care of that.

What I'd like to see are the "most frequently read" toolbar menu from Firefox, support for RSS, and possibly not using a fuckton of memory and taking ages to start (but being that this is a .NET project, it will be hard to achieve, unless it's really minimalistic).

I was thinking about something very similar to your Checkpoint feature yesterday as I was methodically browsing PhotoshopDisasters.com archives.

Also the very act of hovering over the address bar should highlight a valid parse of the URI up to the point your cursor is. I hate manually dissecting an URL just to send a 'valid' link to someone over IM without session IDs, parameters or whatever the fuck I'm looking at currently. Just a matter of asking graphic contexts the width of a certain string and parse upon "/" and "?".

Patrick said:

Even better is make this customizable that way I can set my own search engines with their own 'identifiers.' Possible commands could be g - google, w - wikipedia, dw - doomworld (lol), a - ask.com, etc....

You can do this in Opera.

"imdb <title>" is something I frequently use in Firefox, having something similar for other sites would be nice.

After a long mapping session with Doombuilder, I always end up attempting to navigate webpages like you would the grid in Doombuilder. It would be interesting to see a web browser that worked like that; Left clicking and dragging moves the page around (instead of scrollbars), and mouse wheel zooms in and out. I suppose a little like the default tool when viewing a PDF in Acrobat.

Might not be useful, but it would be interesting.

Although actually, that might end up translating well to a touch-screen application.

Patrick said:

One thing that would be useful is the ability to have a console quick search. an example of what I mean:

yt michael jackson

and then I would get a youtube search for MJ

so I can simply type 'yt' (youtube) into the address bar, then my search terms, and find results. Even better is make this customizable that way I can set my own search engines with their own 'identifiers.' Possible commands could be g - google, w - wikipedia, dw - doomworld (lol), a - ask.com, etc....

Firefox already does this: right-click -> add keyword search.

destx said:

You can do this in Opera.

And Firefox. Hell, Internet Explorer probably has that feature by now.

fraggle said:

Firefox already does this: right-click -> add keyword search.

Just to clarify, you must right click on a search form.

Nomad said:

Left clicking and dragging moves the page around (instead of scrollbars)

Actually, something very similar can already be done in firefox with middle-click and then just moving the mouse pointer around a directional marker.

Maes said:

Actually, something very similar can already be done in firefox with middle-click and then just moving the mouse pointer around a directional marker.

That's a driver/OS feature, no?

Jodwin said:

That's a driver/OS feature, no?

Ehh? AFAIK it's not a global Win32 feature, it only appears on a few apps that explicitly implement it as a custom user interface extension, just like image or map panning in PSP and DB. Unless you're confusing middle-click omni-directional scrolling with wheel scrolling, which is indeed global.

I've only seen this particular "middle click" thingy only on web browsers (don't recall if the first one to use it was Opera or Mozilla though, however it became a staple later on).

Maes said:

I've only seen this particular "middle click" thingy only on web browsers (don't recall if the first one to use it was Opera or Mozilla though, however it became a staple later on).

I have it in Word, Notepad, Wordpad, IE and FF. Quite a few mouse drivers also mention it as a feature that you can turn on or off (or assign on some other button).

Jodwin said:

I have it in Word, Notepad, Wordpad, IE and FF. Quite a few mouse drivers also mention it as a feature that you can turn on or off (or assign on some other button).

It's definitively not a feature of the standard Microsoft mouse driver then. I've only ever seen it as something app-specific, just like mouse gestures. Don't you have a Logitech mouse or something?

Maes said:

Actually, something very similar can already be done in firefox with middle-click and then just moving the mouse pointer around a directional marker.

Similar, I suppose, but not quite the same.

Quality

avery1555 said:

* Basically, I call it Checkpoint because: when you go to a webpage you can flag it and then return to it whenever you wish without bookmarking. For example, I am looking at the Cacoawards on Doomworld but I want to navigate to the downloads page without losing where I am, I would flag my current page, then go to download the WAD or whatever I wanted to do, and then just click the flag again to return to the exact page I was originally on. Think of it as temporary bookmarking.

Opera can kinda do this as well...

There's a rewind button that basically goes back through different domains instead of pages.

You can also make Notes, which are more or less bookmarks but you can associate any text you want with them. So you could make a Note point to the Caco Awards and keep a list of which of the WADs you've haven't played through.

Although for the scenario you describe, I would probably just open the download link in a new tab...

The temporary bookmarking idea is pretty neat. But I can't offer much by way of help here except to suggest that you do not code it in dotnet, as it features the most ridiculous amount of overhead ever seen on a PC, and applications tend to be stupidly slow as well.

Maes, browsers tend to use tons of memory because web script parsing is a very complex process. The more protocols you support, the larger the footprint is. Many moons ago, I wrote an html 3.2 parser of my own, and even in its infancy, it consumed a considerable amount of memory because of the large number of parameters that have to be accounted for. It is not uncommon for a browser that supports just about all major protocols to consume 32MB or more of RAM when running...before even loading a web page. And most of the time, it's not program bloat either...it's all the damn code you have to have to parse a document and then display it. You also have to account for things such as image buffers, alignment details, etc. It all adds up quickly. Web browsers are probably the most complex of the common userland applications today (and that's probably why there are so few of them).

While Firefox is currently working well enough on my desktop machine, I've been hurting for a good lightweight browser for my laptop.

It should have:
- Enough Javascript support to run most websites
- Support for Flash, because some websites just plain don't work without it
- AdBlock and FlashBlock functionality
- Should take up no more than 64 MB of RAM

The rest of it I could deal with - but adblock/flashblock and RAM issues are complete dealbreakers on here.

Exactly. And any sort of managed environment (.NET, Java) isn't going to help.

However I'd just settle for the memory usage not to skyrocket with every new instance/page/tab that you open (it seems a bit too much to add a hefty 20 MB/page. Aren't all those parsers etc. singletons?)

I want a browser that punches web devs in the face if they use too much AJAX and shitty CSS that brings my browser to a crawl. It's getting to the point that I have to force my browser to identify as Internet Explorer Mobile just to see versions of websites that don't try to lag my computer with bullshit sliding windows and widgets.

But until the attitude is fixed (probably won't be), we can only hope for browsers with super-optimized js engines and css engines and whatever else we need just to see text. if they have to be programmed in assembly, so be it.

Today, the only way to use a "real" browser that's not too crippled or stripped down, and can display a majority of existing pages with some success, while being able to use Flash, Java etc. with the least resources possible is, surprise surprise, IE 6.

I guess using older versions of IE or better Netscape or Mozilla could work, if you want a lightweight browser, but there may be huge security holes. The problem is that any attempts at designing a "simple" browser are doomed because this is not 1996, and browsers are required to quite do a bit more than display and render static HTML.

In fact, they are required to be fucking Turing complete. This alone makes simplicity a contradiction in terms, and security a very relative term.

How about a COMPLETE checklist of JavaScript DOM functions that you can enable or disable instead of a tiny selection like Firefox offers. Sites that try to disable the ability to copy and paste text absolutely enrage me. Haven't they heard of fair use rights? Fuck that shit. No browser should allow that to be done.

Quasar said:

Sites that try to disable the ability to copy and paste text absolutely enrage me. Haven't they heard of fair use rights? Fuck that
shit. No browser should allow that to be done.

Heh I first stumbled upon that shit in 2000/2001 I recall. It disabled the right-click entirely, thus disallowing saving images, links or viewing the page source (although it was crude, and could be circumvented in a variety of ways). I was able to disable it by disabling Javascript for that page, on Mozilla at least, but IE also used an ActiveX control, just to make sure...heh. Remember, this was 2000, and the world wasn't still ready for the clamorous security fiascos that would soon follow.

I'd be happy with the browsers' and user controls not be fuckable with in any way, period. A few of the older scripts don't work anymore with newer browsers though (e.g. you may still get a javascript alert like "DON'T STEAL OUR SHIT, FUX0R!" but it won't prevent the context menu from popping up anymore). But yeah, selection and copy/paste can be totally fucked...and some sites go to great lengths to ensure that they are not usable without Javascript (in some they have AJAX textboxes).

Maes said:

and some sites go to great lengths to ensure that they are not usable without Javascript (in some they have AJAX textboxes).

That kind of shit should be illegal. Especially if there is no fallback, for accessibility's sake, as well as compatibility. I fear that HTML5 might be too late, and for some reason, it's not focussed on accessibility in my eyes as it is to somehow "free the world from plugin prison". Yeah, like that's gonna happen.

The real problem with HTML in and of itself is that it was never designed to do what websites do today. That's why it's had to be extended so many bloody times. And then you've got assclown companies like Microsoft who come in and try to impose their own standards (just like they did with Java before being assraped by Sun and forced to knock it the fuck off), which are not compatible with other browsers. Hardly anyone follows the W3C standard these days...some arrogant web developers even call it a crock and intentionally defy it, catering to the legions of IE-only twerps. I have always refused to work like that...I started professional web development in 1996 and have never in my life developed a site that intentionally defied the standards. HTML itself needs to be retired, and a completely new format brought out that isn't like anything that's out there right now. And XHTML isn't the answer either, as much as many XHTML zealots would like it to be.

What would you like to see in a web browser?

PORN

I'd like a web browser that doesn't suck more than 10mb to display text without any crap going on.
I remember when Netscape 4 suddenly got 10x slower than 3 and hardly had any advancements. It's like this slowdown bug was never fixed in any mozilla branch.

Patrick said:

One thing that would be useful is the ability to have a console quick search. an example of what I mean:

yt michael jackson

and then I would get a youtube search for MJ

so I can simply type 'yt' (youtube) into the address bar, then my search terms, and find results. Even better is make this customizable that way I can set my own search engines with their own 'identifiers.' Possible commands could be g - google, w - wikipedia, dw - doomworld (lol), a - ask.com, etc....

Are you using firefox? If so, you an go to goosh.org . It seems like goosh should do most of what you want there.